Académique Documents
Professionnel Documents
Culture Documents
GyulaSallai
Cognitive
Infocommunications
(CogInfoCom)
Cognitive
Infocommunications
(CogInfoCom)
123
Adam Csapo
Institute for Computer Science
and Control of the Hungarian
Academy of Sciences
Budapest, Hungary
Pter Baranyi
Szchenyi Istvn University
Gyor, Hungary
Institute for Computer Science
and Control of the Hungarian
Academy of Sciences
Budapest, Hungary
Gyula Sallai
Budapest University of Technology
and Economics
Hungary
Future Internet Research
Coordination Centre
University of Debrecen
Hungary
ISBN 978-3-319-19607-7
DOI 10.1007/978-3-319-19608-4
Preface
In the past decades, the role and value of information has significantly increased in
nearly all aspects of daily life. Information besides matter and energy has become
a primary conceptual component of human thinking, as evidenced by the growing
appearance of the word information in a widening range of contexts. Alongside
the factories, power plants and relevant industries which provide, store, process
and transport matter and energy, a system of infocommunication services providing
analogous functionalities with respect to information now pervades our everyday
experience. As a result, access to infocommunications has become a basic need, in
much the same way as the ability to access electric power through outlets in walls is
taken for granted. In fact, most services pertaining to matter, energy and information
are co-dependent at various levels, with each service relying on others for its own
sustained efficiency.
At the same time, it is important to realize that while services relevant to matter
and energy primarily stem from and serve physical needs, infocommunications
is motivated by and serves high-level mental activities, which are supervenient
on human knowledge and intelligence. Given that human mental capabilities are
generally more flexible to adaptation than material-energetic capabilities governed
purely by physical-biological reality, new patterns, modes and structures are
constantly appearing in the ways we interact with information. Coupled with recent
advances in technology, this inventiveness in means of interaction is culminating in a
process of human entanglement with information and communication technologies
(ICT). This process has been identified and described by many authors, and can
be observed from a number of different perspectives relevant at different scales
of time, including the key measures used to characterize the performance of new
technologies, the multiple (physical) levels at which interaction can occur, and
the increasing relevance of what we refer to as human-oriented fields in the
development of new technologies.
With respect to the ways in which the performance of new technology is characterized, it can be observedat least in the case of personal informatics devices
and from the perspective of end usersthat measures of hardware performance
are losing importance in comparison with high-level (soft) functional capabilities.
v
vi
Preface
While 20 years ago the specifications of personal computers would have included
details on power consumption, processor clock frequency, cache size and other
low-level hardware details, later the capacity of random access memory, as well
as networking and graphical capabilities gained relative importance. Today as end
users of personal informatics devices, we rarely take note of such specifications,
despite their still being accessible. Instead, we take for granted that a new device
will include capabilities for the most common communication protocols, and will
include a camera, for instance, that produces images of decent quality for most
practical purposes. Even our focus on battery life is much more oriented towards
capability than towards raw measures of hardware performance: if the battery
life of a device can last for a full working day, most practical requirements for
functionality will be satisfied. As this tendency continues, the specifications of
personal informatics devices most important to us will soon beand already are
to some extentthose that focus on high-level capabilities built into and around the
operating systems running on them, such as the richness of associated application
stores, the various capabilities of built-in personal assistants such as Siri, Google
Now and Cortana (including capabilities for generating and understanding speech,
text, gestures and other emotional behaviours), or capabilities for handling multiple
or multi-user profiles for family use. In the end, our personal informatics devices
can be expected to evolve into tools that merely provide an intelligent window
into a ubiquitously and collaboratively accessible augmented world of personalized
information servicesa fact that yields significant motivations, and important
support for key notions behind research areas such as speechability, ethology-based
CogInfoCom and mathability, as discussed extensively in the book. The capabilities
of future devices will be judged based on the extent to which they are motivated by
and fulfill the goals of such branches of CogInfoCom, and once this is the case, it can
be said that these are the research areas which supply the most relevant parameters
requiring specification.
With respect to the multiple levels at which interaction can occur (i.e. in terms of
the second perspective behind entanglement listed earlier), it can be observed that
the phenomenon of entanglement is gaining strength at levels of communication
ranging from the low level of cell-electronics connections (as in, e.g., braincomputer interfaces and some areas of bionics) through connections expressed
at the representational, structural or command level (involving, e.g., multimodal
interaction), all the way to high-level collective relationships among populations of
humans and ICT (involving e.g. high-level services based on data aggregation and
analysis). These three levels and their impacts are discussed extensively in the book,
in relation to research areas such as CogInfoCom channels and socio-cognitive ICT.
The point is also made that as the three levels of interaction gain strength, they can
be expected to influence each other in as yet unforeseen ways.
Finally, with respect to the increasing relevance of human-oriented fields in
new technologies (i.e. in terms of the third perspective behind entanglement listed
earlier), it can be observed that as our interaction with information evolves, both
the theory and practice of infocommunications are increasingly relying on results
from fields of scientific inquiry motivated by the goal of better understanding how
Preface
vii
humans think, feel and act. In particular, all fields dealing with high-level aspects
of human cognition, including even the humanities, can ultimately be expected to
make important contributions. It is true that this is a slower process, characterized
by a larger time constant than the processes of change behind functional capabilities
and levels of interaction; nevertheless, it is important to note that the effects of such
human-oriented fields can already be observed in the plethora of new research
directions emerging within ICT. For example, the Future Internet concept that is
actively supported by the EU has several branches that are strongly relevant to this
phenomenon, such as the Internet of Things (IoT) and 3D Internet. In IoT, the goal
is to deploy as many sensors and actuators as possible into the surrounding physical
network, enabling the measurement and actuation of previously inaccessible parts of
physical-biological-mental reality. Irrespective of how the data collected in this way
is represented, it will serve as a kind of informational, or virtual image of everything
in the world that is important to us. Through the concept of 3D Internet, researchers
are seeking ways to attribute 3D geometry to this growing collection of data
including both data that is naturally and directly amenable to 3D representation
and abstract data sets requiring virtual visualizationin order to support our
capabilities for interpretation and manipulation. The increasingly popular domain of
Big Data also represents an initiative towards processing IoT data, albeit primarily in
more automated, less human-interpretable ways. A common effect of these research
areas is that a new infrastructure is emerging, which can be seen as a collection
of artificial sensory organs that extend across the globe and are capable of
extracting different kinds of information in a way that integrates humans into the
picture. It can be expected that this human integration will eventually bring about
new forms of entanglement, which extend through periods of time that are much
longer than the relatively episodic interactions imagined today. New generations
are already growing up in a kind of co-evolution with ICT, as evidenced by
the growing number of psychological effects that have been documented with
respect to changing cognitive-social-technological relationships. In the book, the
term cognitive entity is introduced to characterize such entangled relationships
in terms of cognitive capabilities. Further, the term generation CE (generation
of cognitive entities) is introduced by analogy with the generations X, Y and Z,
to reflect the fact that members of todays generation are growing up in this new
world starting from the first phases of their personal and social development, and
that through their entangled co-evolution with ICT, they can be seen to embody
an entirely new set of cognitive capabilities. The general conclusion is that if
researchers would like to record, explain and understand the diverse phenomena
surrounding human-ICT entanglement and newly emerging cognitive entities, as
well as build new technologies based on those phenomena, then having recourse to
results from relevant human-oriented research fields will be essential, in much the
same way as the development of personal informatics devices today involves joint
efforts from information and electrical engineers as well as experts of psychology
and ergonomics.
The goal of the emerging field of cognitive infocommunications (CogInfoCom) is to support the investigation of human-ICT entanglement related
viii
Preface
phenomena as well as how such phenomena lead to new cognitive entities and
the appearance of the generation CE. The field places special emphasis on the
qualification and quantification of cognitive capabilities, and aims to adopt a multiand interdisciplinary approach in doing so. The distinction between multi- and interdisciplinarity is an important one: while the formulation of the scope and goals of
the field encourages the involvement of a large variety of pertaining disciplines, it
also aims towards an integration of the perspectives and methodologies which they
represent. Thus, besides aiming to create a common language, it is also the goal
of CogInfoCom to establish a common way of thinking about relevant problems,
and a common approach towards designing new technologies. This is not without
challenges. The engineering sciencesand by extension, infocommunications
generally emphasize the synthesis of novel systems so as to provide functionalities
that are guaranteed to be stable and reliable in at least some respect. Crucially, even
the modeling of component structures and phenomenaserving as a prerequisite to
engineering designproceeds with these ultimate goals in mind. Considerations
in engineering often do not extend to the requirement of holistic completeness;
instead, specific interpretations are focused on that are relevant to specific design
goals. In contrast, the primary focus of most branches of the cognitive sciences
is analytical in nature, with the goal of recording, understanding and explaining
cognitive phenomena in a way that is phenomenologically complete. However, such
analyses do not necessarily lead to functional reproducibility. As a result, it is often
the case that a set of experiments and models separately deemed precise from the
perspective of an engineering field are seen as inadequate from the perspective of
a cognitive science-related field, and vice versa. If a personal computer from our
century were to have suddenly appeared functionally complete 100 years ago, surely
its operation would have been described from a different set of perspectives than
those descriptions that are accepted as blueprints for their production today. The
former set of descriptions would also likely have been inadequate for the design of
a fully functional computer. From the perspective of CogInfoCom, the key point is
that while investigations in the cognitive sciences focus on a biological structure that
has been highly differentiated through the complex processes of evolution, and that
as a result cannot easily be copied or reproduced using todays engineering
concepts, it is nevertheless possible to obtain new knowledge by attempting to
extend human cognitive capabilities through the synthesized design of technologies
in infocommunications. This latter goal in turn necessitates a broadening in scope
of the human-oriented methodologies used in the cognitive sciences to address
(artificial) cognitive phenomena relevant to artificial systems.
It can often be observed that well-established research fields become more
specialized, or perhaps show increasing signs of segmentation, while specific
segments meet with previously distinct research fields to form completely new
synergies. This process can be likened to the roots of a tree, which meet at various
points to create the trunk, which in turn gives distinction to different branches, which
then meet with the foliages of neighbouring trees. In this process, some researchers
choose to maintain focus and achieve completeness in their own specialized
research, while others instead are in constant search of new possibilities for synergy
Preface
ix
with other fields. Although seemingly there may be a contradiction between the
two approaches, in fact both of them together are necessary for the emergence
of valuable new research directions. So it is the case with CogInfoCom: one can
find in it the influence of well-established research fields (e.g. augmented cognition,
human-computer interaction, virtual reality and others), while at the same time new
directions and specializations are also appearing due to the unique viewpoint of the
field, as discussed in several chapters of this book. It is important to emphasize
that these new specializations born under CogInfoCom are often difficult, if not
impossible, to categorize into already existing fields. Interesting and novel changes
in viewpoint are often possible by considering various phenomena from tighter, or
broader perspectives. All of these tendencies can be observed in parallel at the IEEE
CogInfoCom conferences series. In some cases, sessions with well-specified topics
have included presentations of researchers coming from widely different areas, such
as information, medical or cognitive sciences, which has led to valuable exchanges
with respect to a focused topic. In other cases, sessions named after completely
new research directions appeared, and have included presentations which aimed
to establish novel perspectives. This heterogeneity can also be observed in the
fields represented by researchers who have co-authored papers at conferences and
special issues on CogInfoCom; for example, joint papers were published together
by researchers with backgrounds in computer science, systems and control theory,
mathematics, ethology and cognitive psychology.
As in the case of all newly established conference series, the CogInfoCom conferences have been attended by both supporters and sceptics of the motivations and
ideas behind the field. At the 3rd IEEE International Conference on CogInfoCom,
it happened that a professor dealing with human-computer interaction voiced his
criticism, saying that There is nothing new in CogInfoCom: : : it is just another
name for what I am working on. Later at the same event, another scientist from
the field of computational linguistics made almost the same remark, expressing
his view that CogInfoCom was just another label given to the already existing
field he was dealing with. As the same criticism was later raised with respect to
cognitive informatics and neurocomputing as well, it is worth considering what
these criticisms would entail, if accepted as true:
CogInfoCom D HCI
CogInfoCom D computational linguistics
CogInfoCom D cognitive informatics
CogInfoCom D neurocomputing
Were these equalities to be accepted as true, clearly they would also suggest
relationships of equivalence among their right-hand sidessomething that has never
been claimed by any serious researcher acquainted with the scope and goals of these
fields:
Preface
Preface
xi
Solvang, Peter Varlaki, Gabor Vattay, Klara Vicsi, Alessandro Vinciarelli, Carl
Vogel and Yeung Yam, who have significantly contributed to the success of the
conference series by helping to organize sessions and tracks that were often
pioneering in their subject matter, and also by supporting through their work the
appearance of several special issues on CogInfoCom in international scientific
journals. We thank the entire CogInfoCom community for helping to create a
friendly and scientifically stimulating atmosphere, as well as for bringing their
valuable perspectives into the fruitful discussions held during and oftenthrough email and Skypebetween the CogInfoCom conferences. Last but not least, special
thanks is due to Anna Szemereki for her Herculean efforts in the organization of
every last detail of the CogInfoCom conferences year after year.
Budapest, Hungary
April 2015
Pter Baranyi
Adam Csapo
Gyula Sallai
Contents
Part I
Foundations of CogInfoCom
3
4
5
5
7
8
13
13
14
14
15
16
16
18
21
21
23
23
25
26
27
28
31
32
xiii
xiv
Contents
Part II
39
40
41
43
44
45
45
47
50
51
52
54
57
57
59
60
62
63
66
69
71
73
74
75
76
77
78
79
79
80
81
82
91
92
92
93
95
97
98
Contents
Part III
7
xv
CogInfoCom Channels
105
105
107
109
110
111
114
116
119
120
120
121
122
123
123
125
126
127
129
129
130
131
134
134
135
136
138
139
141
141
144
144
145
146
147
xvi
Contents
157
157
159
160
162
162
164
165
166
166
173
173
175
178
179
181
181
181
182
183
183
184
184
184
xvii
xix
Part I
Foundations of CogInfoCom
Chapter 1
The term cognitive sciences is used here as a comprehensive term for those fields that deal
with the capabilities of biological systems, includingamong othersthe fields of psychology,
neuroscience, cognitive modeling, cognitive ergonomics and human factors, linguistics, biology,
anthropology, some branches of artificial intelligence, etc.
effects on human cognition, and how they can be applied to sharing information
in ways that appeals to human cognitive capabilities. As a result, any research
focusing on interactive media, or interaction technologies in general will have strong
relevance to the field. However, another important factor behind cognitive media is
the growing prevalence of artificial sensory capabilities implemented in media: in
a way analogous to the human nervous system, the Internet as an interconnection
of globally distributed devices and nodes, together with the media applications
based on it can be seen as an artificial nervous system and an artificial set of
sensory modalities. Data available on social network platforms such as Facebook
and Twitter are increasingly used to predict and understand physical, mental and
social processes taking place in the world. According to this view, cognitive media
targets not only the cognitive effects that media has on humans, but also the
cognitive capabilities of media itself. As this viewpoint gains increasing viability, it
can be expected that at a certain point the term cognitive will no longer be reserved
exclusively for the description of human-oriented phenomena: although this was the
original meaning of the term (from a time when modern media, not to mention ICT
was still a dream of the future), today the ICT infrastructure surrounding us has
growing capabilities for sensing, representing and interpreting a growing scope of
phenomena.
Similarly, the terms cognitive informatics and cognitive communication/cognitive communications are omnipresent in delineations of research
focus. A definition of cognitive informatics can be found on the official web page of
a number of research organizations (e.g., the Pacific Northwest Laboratory funded
by the U.S. Department of Energy). Today there are several research institutes
dedicated to cognitive informatics, such as the Institute for Cognitive Informatics
(ICON) in Sheffield, UK, which has an annual budget of over one million GBP.
An IEEE International Conference on Cognitive Informatics has been held every
year since 2002, and several symposia with leading researchers invited as keynote
lecturers have been held in the past few years. Cognitive communication(s), when
used without the trailing s, can refer to the study of ways in which humans
anticipate context as a contributor to the choice of communication style, and
perceive the consequences of communication choices (in short, it is the study of
human mental models of communication). On the other hand, with the trailing
s, cognitive infocommunications is used to refer to devices and networks which
can assign resources and functionalities in dynamic and intelligent ways (as in
e.g., cognitive radio or cognitive networks). It is possible that in the future, these
two directions will reach common formulations for certain problemssee e.g.
Mitola and Maguire (1999), which specifically mentions user needs and application
scenarios as possible contributors to the adaptivity of cognitive radio.
In a way analogous to the evolution of media informatics, media communications
and infocommunications, examples of results which can be categorized as belonging
to cognitive media informatics, cognitive media communications and cognitive
infocommunications are gradually emerging, even ifas of yetthese fields are not
always clearly defined. The relations between the fields discussed here are shown in
Fig. 1.1 (further significant details on this figure are provided later in Chap. 3).
Fig. 1.1 Relationships between fields relevant to the convergence process behind CogInfoCom.
The figure highlights the fact that the emergence of CogInfoCom is a result of the way in which
the fields of informatics, media and communications have evolved. From a historical perspective,
CogInfoCom can also be regarded as an extension of cognitive communications (an in-depth
discussion on this aspect is provided in Chap. 3 of this book)
On the other hand, the electronic media production industry had yet to become
involved in the convergence process.
3. Infocommunications. The electronic media content producing industry, as well
as the appropriate institutions for regulation joined the convergence process
to produce the technological and social-economic revolution that is todays
infocommunications industry.
4. Cognitive infocommunications. A natural fourth step is the integration of
cognitive sciences into the convergence process behind infocommunications. On
the one hand, this involves an expanding content space, in which new information
types are gathered, inferred and reasoned upon in novel ways. On the other hand,
it is essential that this content be used in effective ways to extend the cognitive
capabilities of both users and ICT, allowing for the formation of new cognitive
entities.3
As a result of the synergy between the cognitive sciences and infocommunications, the content space that is accessed and manipulated by both users and ICT is
expected to grow in size and richness of interpretation. If the potential behind this
expanding content space is to be harnessed, it can be expected that the respective
unique capabilities of humans and ICT will be applied through new, long-term
interaction patterns, leading to the emergence of new cognitive entities. A significant
challenge in CogInfoCom is how to power these cognitive entities with the kinds
of information and functionality that are suitable to their roles and requirements.
For a fully detailed explanation of what is meant by the term cognitive entity, the reader is
referred to Sect. 2.3.2.
Fig. 1.2 Five phases of Gartners hype cycle in terms of technological expectations with respect
to time
4. Slope of enlightenment. Potential benefits of the technology become increasingly widely understood as an increasing number of successful use cases become
apparent. Second- and third-generation implementations of the product appear,
new pilots are increasingly funded, while more conservative actors remain
cautious.
5. Plateau of productivity. Mainstream adoption begins to take off. The broad
market applicability of the technology becomes clear, and its adoption yields
clear advantages.
One of the messages of the hype cycle model is that almost all new technologies
are initially surrounded by a kind of hype that is, in a way true to the definition of
the word, founded more on a combination of illusions, expectations and a desire
to provide something new than on a true realization of benefits, drawbacks and
associated tradeoffs. This is what leads to the trough of disillusionment, which
in turn often (or in the case of successful technologies, always) brings about a
reassessment with respect to potential manufacturing approaches and use cases.
The end result of this reassessment combined with a few instances of successful
applications is a better informed proliferation of the technology.
Besides making possible these general conclusions, however, the model also
allows for the tracing of technological advances through time, both in nominal and
relative terms. On the one hand the location of an individual technology on the
hype cycle curve allows for assessments to be made with respect to the timeliness
and the expected success of that technology in the near future. On the other hand,
10
11
12
that time). This suggests that the maturation processes of such capabilities are
governed by larger time constants and are expected to remain in the main focus
of CogInfoCom-related research areas for years to come.
In summary, the evolution of hype curves from the past 45 year period strongly
suggests that the primary technological components underlying CogInfoCom have
reached the phase of maturation, while the integration of these components as
building blocks into more complex capabilities has begun both in subtle ways as
well as at a very explicit conceptual level. This integration process can be expected
to gain further traction in the coming years as the technologies based on which it is
grounded remain at the stable end of maturation, and as new technological triggers
continue to appear based on those technologies (as demonstrated by the hype curve
of 2014, many new concepts are already appearing). Such salient processes of
technological integration are a strong motivating factor behind CogInfoCom.
Chapter 2
In this chapter, a description of the scope and goals of CogInfoCom is provided. This
is followed by an overview of novel conceptssuch as those of mode and type of
communication, as well as the more general notion of cognitive capabilitywhich
have emerged through the field. Further, a set of assumptions, primarily founded on
the existence and consequences of the merging process between humans and ICT,
are described in terms of their relevance to CogInfoCom research.
As we will see, in many cases this separation between natural and artificial is no longer
meaningful.
13
14
Gza Nmeth, Mihoko Niitsuma and Gyula Sallai at the 1st International Workshop
on CogInfoCom, held in Tokyo, Japan in 2010. The finalized definition is as follows:
Definition 2.1. Cognitive infocommunications (CogInfoCom) investigates the
link between the research areas of infocommunications and the cognitive sciences,
as well as the various engineering applications which have emerged as a synergic
combination of these sciences. The primary goal of CogInfoCom is to provide a
systematic view of how cognitive processes can co-evolve with infocommunications
devices so that the capabilities of the human brain may not only be extended through
these devices, irrespective of geographical distance, but may also interact with the
capabilities of any artificially cognitive system. This merging and extension of
cognitive capabilities is targeted towards engineering applications in which artificial
and/or natural cognitive systems are enabled to work together more effectively.
15
16
17
3. Finally, a third level of entanglement can be seen to occur at the collective level of
multi-user interactions. Applications in this layer can have relevance to collective
behaviors in two ways: by making use of collective behaviors in order to support
individual users interaction with a system; or alternatively, by supporting the
prediction or a posteriori analysis of collective events based on an analysis of past
behaviors (both individual and collective). Such applications often rely on the
mining and analysis of vast amounts of heterogeneous data sourcesincluding
e.g. activity on social communication platforms.
From a general perspective, Hodder defines entanglement as the sum of four
types of relationships in which humans depend on things (HT), things depend on
other things (TT), things depend on humans (TH) and humans depend on humans
(HH) (Hodder 2014, 2012). All four of these co-dependence relationships can be
equally observed in the particular case of human-ICT entanglement. However, it is
important to note that all of them can, and should be interpreted at multiple temporal
scales, ranging from episodic interactions that are point-like in time to decadeslong periods of co-existence. An important quality of human-ICT entanglement
is that co-existence with ICT is becoming an inseparable feature of the everyday
experience of children growing up today. This is a new phenomenon that is opening
new avenues of research which extend far beyond human-computer interactions;
such research must take into consideration the not only the capability to achieve
certain functions in comfortable and effective ways, but also the psychological and
mental effects of long-term use, and how ICT can evolve together with humans in
order to become a natural, ecologically valid part of the everyday human experience.
For example, the psychological effects of ICT can be grasped by considering the
effects of ubiquitous e-mail access and connectednessas suggested by popular
notions such as e-mail apnea,2 phantom vibrations or internet indispensability
(Stone 2011; Drouin et al. 2012; Platzer and Petrovic 2011); as well as the general
feeling one experiences when a device or network malfunctions (the feeling that
part of oneself is malfunctioning rather than an external technology) as described
in Pang (2013). It is also worth considering how the Internet is physically closer
to users than before: while a decade ago, shutting down the computer for the night
meant that access to the Internet was finished for the day, today nothing is easier
than checking our e-mail or the weather report one more time on any of the number
of mobile devices surrounding us.
Such long-term co-existence in turn raises awareness on the usefulness of
applications with long-term goals. Apart from enabling users to perform a specific
task in a specific domain, an equally viable goal might be to support, through
ICT, the development of a capability to perform an increasingly complex set of
tasks (this is the case, for example, in speechability and mathability, as will be
described in Sects. 5.2 and 6.6). It should be noted that long-term interactions are
also capable of generating functionalities that are not planned for a priori (see
also Sect. 2.3.3). For example, as users continuously interact with, and increasingly
2
A temporary absence or suspension of breathing, or shallow breathing, while doing email
(Stone 2011).
18
become entangled with systems of ICT components, new use case requirements,
and new possibilities for previously unknown functionalities are discovered. The
accommodation of such functionalities, in turn, can create new patterns of usage
many of which may then have a recurrent effect on usage patterns, and potentially
even on the social organization of the community of users involved in the merging
process. This mutual influence between usage patterns and use-case requirements
is creating an open-ended evolution of functionality. From a different perspective,
long-term co-evolution between humans and ICT is also enabling the offline
collection of vast amounts of data which can later be instrumental in developing new
applications (for example, if those applications rely on machine learning techniques
requiring large amounts of data). As a result, the creation of new functionalities is
facilitated in more than one way through the implicit knowledge that is generated
from past interactions. Based on the above, we introduce the term tangleface to
be used instead of the word interface when characterizing ubiquitous human-ICT
relationships. Additionally, we use the term tangleaction instead of interaction
when such relationships persist through extended periods of time. These notions
will be used often and elaborated from a variety of perspectives in later parts of this
book.
19
The bottom line is not that one would be philosophically inclined to specify a
boundary between entities, but that it is also necessary to specify such boundaries
from the functional perspective of engineering design. On the one hand, in a domain
where difficult problems of synthesis can be effectively tackled only by breaking
them down into smaller components and gluing those components together through
some form of communication once they are complete, the functional boundaries
at which this is done can make or break the tractability and sustainability of an
implementation. On the other hand, once it is accepted that the boundaries between
artificial and natural are not as clean as they were a few decades ago, unprecedented
possibilities emerge for the development of new functionalitieseven cognitive
capabilities. Such capabilities can be seen as implemented in the dependencies
between components in much the same way as lower-level functionalities are
created as a result of several different components working appropriately in
mutually specified contexts. This hierarchical dependence among capabilities can
be seen as leading to a hierarchical organization of cognitive capabilities.
The embodied perspective of cognition that is currently favored in the cognitive
sciences adopts the view that human cognitive capabilities and human intelligence
are emergent properties which cannot be separated from the physical, biological,
neurophysiological and higher-level bases of our mental existence (Deacon 2013).
It is also clear that the social and technological context of our interactions with other
humans and ICT further influences the kinds of mental and physical work that we
are able to perform (Hollan et al. 2000; Deacon 2013). Further, analogous emergent
properties can be identified in the functionalities of ICT devices and networks.
Although the view that computers are merely symbol processing systems has been
implicitly accepted for decades by thinkers and technologists of all backgrounds, an
emergentist view of computing is now gaining acceptance. Though fundamentally
different from living systems it can be argued that all computational systems
(apart from purely theoretical constructs such as the Turing Machine) have some
form of embodiment, and that furthermore, computation in general has physical
underpinnings and physical ramifications (Heder 2014). As highlighted by several
authors, even lexical knowledge such as knowing the derivative of the sine function
or knowing the capital of a country is strongly embodied in the sense that without
direct experience in working and manipulating functions, or without being able to
travel and experience through our bodies what really constitutes a city, our notions
would be entirely different (Picard 2003a; Heder 2014).
The extension of such notions to human-ICT entanglement and emergent
cognitive capabilities can be seen as a natural development. It can be argued that
not only are new functionalities and cognitive capabilities formed through humanICT entanglement, but that they can also be seen as higher-order in the sense that
they are dependent on lower-level foundations. However, this point of view also
makes it possible for such higher-order capabilities to be combined into newer
ones that are located at still higher levels of hierarchy. This process is illustrated
in Fig. 2.1. From a practical point of view, the figure shows that all cognitive
capabilities, however trivial in a human-ICT context, can be analyzed in terms of
other capabilities on which they depend. This observation can serve as an important
20
human sensory
modalities
artificial sensory
modalities
time
Fig. 2.1 New, higher-level artificial capabilities and emergent cognitive entities are created
through time as new use cases are generated through a broadening of artificial sensory modalities
and the increasing possibility for users to co-evolve with them through extended periods of time
The analogy with the Internet of Things is clear, when IoT is regarded as a cyberization of the
physical world together with humans, as will be detailed Chaps. 3 and 4.
21
the daily experience of this generation, and become irrevocably intertwined with its
expectations and thinking processes, both at a conscious and subconscious level.
Expressed more directly, one can say that ICT cannot be taken away from this
generation without bringing about significant cognitive effects and psychological
discomfort.
22
point of view of the user.4;5 For instance, a driver using Google Maps or Waze for
direction information does not need to know whether and how many other drivers
are queried for traffic congestion information before the system recommends a
specific route. Similarly, to a user interacting with several social networks at the
same time, the route taken by a link before it is shared with the user is unimportant;
only its original source is important besides that fact that it eventually reaches
the user. Many of the technological services we use on a daily basis we do so
unconsciously, while taking for granted that they will be available through the
flexible coordination of available devices and software components.
In this way, the emergence of CogInfoCom is implicitly based upon, and
also further supports a transition from operation to high-level functionality. This
transition is characterized by a decreasing need for explicit specification of use cases
and operational details, and an increasing prevalence of high-level and adaptive,
transparent functionalities that are seen and treated as tools rather than procedures
of operation.
Of course, from the perspective of an application designer, implementation details will always be
important.
Chapter 3
This chapter discusses the foundations of CogInfoCom from the historical perspective of the diffusion of information and communications technology (ICT)a
phenomenon resulting from the evolution of digital electronics and the convergence of telecommunications and information technology. A holistic overview of
this digital convergence process is provided, with special focus on changes in
technological background, changes in value chains, social-technological phases (in
particular, the cognitive phase) as well as newly emerging applications. Through an
understanding of this convergence process, the position of CogInfoCom within the
Digital Ecosystem is clarified.
23
24
The next step of the convergence is also outlined by the combination of infocommunications,
cognitive science, quantum mechanics, nanotechnology and biotechnology
25
networks and services, and to find the best regulatory regime to deploy this process
(Henten et al. 2003; Melody 1997). Electronic communications provide the bearing
digital infrastructure for the digitalized content services and applications, whereby
the convergence process has been naturally expanded to all information processing
and content management functions; and the integration of telecommunications with
information technology and electronic media has been realized, the infocommunication sector was born (European Commission 1997, 2010; WEF 2007; Sallai 2007;
Liu 2013). Content involved voice, data and visual information, integration of media
streams and explosion of multimedia are to be highlighted. However content space
cannot be limited to these media contents. All other senses, including touching,
smelling, as well as emotions, sensory information, three-dimensional gestures, and
any other cognitive information can be also processed, transmitted and displayed.
Recently the content space is expanding by cognitive and sensory contents, billions
of smart devices are to be interconnected (CogInfoCom 2010, 2013; Nishinaga
2010; Smith 2012; Vermesan and Friess 2013; European Commission 2013a).
Cognitive infocommunications by definition aims to manage the expanded content
space, and to provide opportunities for creating enhanced, richer applications
(Baranyi and Csapo 2010, 2012). At the same time the identification capacity
of the todays Internet is running out, Internet architectures are reconsidered for
better managing scalability, mobility and quality requirements, application creation,
for integrating cognitive capabilities and security issues, for handling billions of
connected devices (Internet of Things) and big volume of data generated (Big Data)
as well as for exploiting the opportunities derived from the intensive technological
development. This convergence process transforms business and bank spheres,
administration, production, agriculture, transport, health, education and knowledge
systems, etc., our everyday life. The future networked knowledge society is going
to be established on Internet base, but the limitations of the todays Internet must be
eliminated. Recognizing the challenges of the todays Internet and the opportunities
for a more advanced Internet, European Commission supported intensively the
research activity on the Future Internet (FI) and initiated the organisation of the
Future Internet Assemblies (FIAs). Since 2008 eleven FIAs were held and four
books were issued on the research results (Tselentis et al. 2010; Domingue et al.
2011; Alvarez et al. 2012; Galis and Gavras 2013).
26
27
sectors had their own specific technology; the penetration of digital technology
has been started by implementing the sector-specific communication functions. The
sector-specific technologies are more and more characterized by the intensive use of
the digital technology and the whole separate network, services and terminals have
been reintegrated on digital base (separated internal digitization). The digitization of
telephony started with the use of digital transmission, followed by the introduction
of digital control and switching, which led to an integrated digital telephone network
(OECD 1992; Freeman 1994; Telpolicy 1994; Melody 1997; ITU 1999). Modern
mobile telephony is already fully based on digital technology. Fixed and mobile
telephony convergence (FMC) can be considered a specific case in the frame of the
voices value chain. Digital solutions have also penetrated into AV communications.
28
Fig. 3.3 Phase2: value chain of electronic communications with horizontal convergences
integrated network, the Internet protocol suite (TCP/IPv4) was proved to be most
efficient connectivity technology. Additionally, the combination of voice, data and
audio-visual services offered new IP-based multimedia service opportunities. At
the users, various integrative terminals appeared. The value chains of the voice,
data and AV communications have been merged; horizontal convergence and some
integration of the services, networks and terminals can be identified, and instead of
the separated vertical value chains a single value chain with three horizontal layers
can be shaped (Fig. 3.3) (Henten et al. 2003; Sallai 2007). A unified telecommunication sector has emerged, which is formally called electronic communications.
In deploying these horizontal convergences, uniform regulation was introduced for
electronic communications in the European Union (Ryan et al. 2003).
29
to all information and media technology functions. The same digital message
form is used in the computer sector for passing messages within and between
the computers, together with the growing operational use of computers within
telecommunications, resulting in a synthesis between the telecommunications sector
and the computer-based information sector (info-telecom/info-com convergence).
The IP-based connectivity technology provided a platform on which higher functions can be built. The IP-based solutions in both sectors generated an integrated
structure for processing, storing, accessing and distributing information (ITU 1999;
Valtonen 2001). The electronic media and content production and management
have also been involved into the convergence process (media convergence) by
using the opportunities delivered by IP-based telecommunications and information
technology (ICT), which is demonstrated by the integration of media streams
(multimedia) and the explosion of social media, and the immense spread of the
Internet provided by telecommunications and Internet service providers (ISPs)
(Domingue et al. 2011). Electronic content services and applications based on web
technologies and delivered by electronic communication networks and services
have emerged, e.g. e-business, e-commerce, e-government, e-health, e-learning,
on-line administrations, voting, population census or games, etc. Sophisticated
Internet applications have also been conceived and developed for managing a smart
home or office, an intelligent transportation or energy system, or intelligent digital
cities (smart cities). In general they can be called e-content or infocommunication
applications, the terms information societys services, content services and eservices are also used (European Commission 1997, 2010; Henten et al. 2003; Sallai
2007; Liu 2013). The third step of digital convergence process is the key step of the
process; it is specially called infocommunications convergence. The value chain of
infocommunications contains three additional layers (Fig. 3.4) (Sallai 2012a,b):
the layer of content space, symbolising the jointly-managed information sources
and the customers payable demands;
the layer of e-content or infocommunication applications, including from simple
content services to the wide variety of secure and multi-content Internet services;
the layer of the common IT infrastructure for applications, the middleware
layer, including common message handling, content management functions
(e.g. directory assistance, editing, indexing), browsers, portals, search engines,
security systems etc.
The layers of IT infrastructure, e-communication services and networks together
can be considered the infocommunication infrastructure. Similar layer models of
infocommunications embracing the functions provided by the Internet technology
have been shown and discussed to allocate the players of infocommunications to
layers (Fransman 2002; Krafft 2003, 2010).
Today, the term infocommunications as expanded telecommunications in
the above mentioned meaning has generally been used by telecommunications
30
Content space
E-content IT infrastructure
31
32
33
Infocommunications (Infocom) is considered in the Sect. 3.2 the natural expansion of telecommunications with information processing and content handling
functions including all types of electronic communications (fixed and mobile
telephony, datacommunications, videocommunications, broadcasting, etc.) on a
digital technology base, mainly through Internet technology.2
Information and Communications Technology, usually abbreviated as ICT,
has been in use from the second half of the 1990s (Stevenson 1997) and is used as an
extended synonym for information technology (IT) to emphasize the role of unified
(tele)communications, the integration of telecommunications with computers, as
along with the necessary software, middleware, storage and audio-visual systems
that enable users to create, access, store, transmit and manipulate information. In
other words, ICT consists of IT as well as telecommunications, broadcast media and
all types of audio and video processing and transmission (EITO 2014; ITU 2009).
At present, the term ICT is generally used and usually refers to the integration of
information and telecommunication technology sectors involving their convergence
with the media technology sector based on common digital technology. The term
Information Society Technologies (IST) was generally used to the integration of
telecommunications, IT and media sectors in EU research, technological development and demonstration framework programmes (FP5 and FP6) between 1998
and 2007 (Arend 2002). However, since 2007 in FP7 and Horizon 2020 the term
ICT has been used. The EU Horizon 2020 programme (20142020) preferably
supports the ICT research and innovation, in particularly the development of
(European Commission 2013a):
next generation computing, advanced computing systems and technologies;
infrastructures, technologies and services for the Future Internet;
content technologies and information management, including ICT for digital
content and creativity;
advanced interfaces, robotics and smart spaces;
nano-electronics and photonics technologies and embedded systems.
The member countries of the Organisation for Economic Cooperation and
Development (OECD) agreed in 1998 to define the ICT sector as a combination
of manufacturing and services industries that capture, transmit and display data
and information electronically. In 2011, the OECD set a number of standards for
measuring and comparing the information society across countries (OECD 2011),
including definitions of ICT as well as Media and Content products (goods and
services). Accordingly, ICT products must primarily be intended to fulfill or enable
the function of information processing and communication by electronic means,
including transmission and display. Content published in mass communication
The terms info-com(s), info-communications (with a hyphen) are used to express the integration
of the IT and (tele)communication sectors (Krafft 2010; Valtonen 2001).
34
media such as printed, audio-visual and online contents and related services are not
considered as ICT products, but are referred to as the Media and Content sector.3
Over the past several years, the abbreviations TIM, as the Telecom IT
Media sector or TIME, as the Telecom IT/Internet Media and Electronics/Entertainment/Edutainment sector, are frequently used to express the full
integration of these sectors and to enhance the significance of content respectively.
The integrated sector was aptly called IST (Information Society Technology) (Arend
2002), and Digital Technology sector seems to be also appropriate. Generally we
use the term TIM, where M denotes Media and Content sector in wide sense,
including cognitive and sensory contents, too.
The alternative terms Digital Ecosystem (DES), Internet Ecosystem, Digital/Internet Age and Digital World have also emerged to embrace all those sectors
that are already or on the verge of being based on digital technologies (WEF 2007;
IDATE 2009). However these terms generally have a broader meaning, the terms
involve the participants, the customers too. The Digital Ecosystem is defined by the
World Economic Forum as the space formed by the convergence of the Telecommunications, IT and Media and Content sectors, and consists of users, companies,
government and civil society, in addition to the infrastructure that enables digital
interactions (WEF 2007). The evolution of digital technology reached and integrates
the customers by their terminals, devices, sensors, comprises cognitive and sensory
content and the applications may involve the active collaboration and context of
customers (human-centric community applications). Smart phones are part of the
personal life; one device with more and more sensory capability is for private and
business. If it is so, the term Digital Ecosystem organically incorporates cognitive
infocommunications, Internet of Things, Big Data, 3D Internet and community
applications.
On the usage of the term ICT, some additional remarks are in order here: (1) Originally,
only information and communications technology (with communications in the plural) was
considered correct since ICT refers to communications (in the sense of a technology of sending
and receiving information), not communication (the act of sending or receiving information by
speaking, writing, phoning, emailing, etc.). Nevertheless, recently, the single form information
and communication technology is becoming increasingly common, and it is also used by the
International Telecommunication Union (ITU 2009). In order to express these dual forms, we use
communication(s) (Giles 2009); (2) The International Telecommunication Union (ITU) and the
European Telecommunications Standards Institute (ETSI)according to their own definition
deal with not only telecommunications, but also ICT issues; however, their relevant activity, their
recommendations and standards resp. are focused on network-centric issues (ETSI 2011; ITU
2009). Some characteristic citations are from their web-sites: ITU is committed to connecting all
the worlds people. : : : We allocate global radio spectrum and satellite orbits, develop the technical
standards that ensure networks and technologies seamlessly interconnect, and strive to improve
access to ICTs to underserved communities worldwide. ETSI produces globally-applicable standards for ICT, including fixed, mobile, radio, converged, broadcast and internet technologies. : : :
ETSI is the recognized regional standards body dealing with telecommunications, broadcasting and
other electronic communications networks and services; (3) The term infocommunications is also
used in politics in a wider sense as a shorter form of information and communications technology
or simply to interpret the abbreviation ICT.
35
36
cyan, magenta and white). That means that, for example, media informatics is a part
of ICT but not part of infocommunications.
Figure 3.7 shows the transformed prism representing the convergence with
cognitive sciences, and the relationship of the cognitive infocommunications, cognitive communications, cognitive informatics and cognitive media. In this historical
perspective cognitive infocommunications can be considered as an expansion of
cognitive communications.
Part II
This part provides an overview of those research areas that are either in a
synergic relationship with CogInfoCom, or have emerged as a result of the unique
perspectives and research efforts represented by the field. A wide range of research
areas are treated here, and it is important to bear in mind that all of them are multifaceted and constitute individual research domains in their own right. At the same
time, all of them incorporate at least some aspect that makes them relevant to the
use or support of cognitive capabilities in infocommunications. For this reason, a
discussion on how they may contribute to CogInfoCom is well justified.
In Chap. 4, several key points of synergy are discussed from the perspective
of existing research fields relevant to the merging process between humans and
ICT. In particular, relationships with the fields of affective computing, augmented
cognition, body area networks, brain-computer interfaces, cognitive informatics,
cognitive networks and Future Internet, HCI and multimodal interaction, sensory
substitution, social signal processing, virtual and augmented avatars and virtual
and augmented reality are considered. It is important to emphasize that while
all of these fields have their own motivations and unique set of methodologies,
primarily those aspects are highlighted which represent opportunities for synergy
with CogInfoCom.
In Chap. 5, an overview is provided of three key research areas which have
emerged through the influence of CogInfoCom. In particular, the areas of CogInfoCom channels, speechability and socio-cognitive ICT are considered. Research
efforts and results relevant to these areas are presented from a perspective that
focuses on a set of generic techniques referred to as iconic, context-based and
pattern-based. The purpose of this discussion style is to highlight possibilities for
the unified treatment of challenges relevant to the entanglement between humans
and ICT. It should be noted that a detailed treatment of CogInfoCom channels is
provided later in Part III of the book.
Finally, in Chap. 6, several new initiatives are described which have been
proposed at the various scientific fora on CogInfoCom to guide future research in the
field. Specifically, areas relevant to augmented virtual reality mediated neuroscience
research, ergonomics of augmented virtual collaboration, ethologically informed
38
Chapter 4
In this chapter, several key points of synergy are discussed from the perspective
of existing research fields relevant to the merging process between humans and
ICT. It is important to emphasize that while all of these fields have their own
motivations and unique set of methodologies, they also incorporate some aspect,
or some future potential that makes them relevant to the use and support of
cognitive capabilities in infocommunications. In this chapter, we aim to focus
primarily on such aspects. However, partly due to the fact that the information
concept underlying CogInfoComas discussed earlier in Sect. 2.3.3focuses on
functionally relevant by-products of interaction rather than exclusively on the
transfer of explicit knowledge, it will not always be possible to draw a clear line
between what is relevant and what is not. Nevertheless, those aspects that are already
clearly relevant to CogInfoCom are presented in some detail. Modes of usage which
focus on long-term co-evolution rather than episodic interactions are of particular
interest.
A structural visualization of the relationships treated in this chapter is provided
in Fig. 4.1. As shown in the figure, all of these fields, including CogInfoCom
itself, are simultaneously characterized by both overlaps and unique distinguishing
features. For example, an important notion that sets apart the goals of CogInfoCom
from any of the research fields discussed in this chapter is that it aims to reach
an understanding of how extended periods of co-evolution can trigger novel
functionalities. This aspect of long-term co-evolution is rarely acknowledged, not to
mention planned for, in other fields. The concepts of tangleface and tangleaction
as analogies to the words interface and interactionwere introduced earlier
in Sect. 2.3.1 precisely for this reason: to clearly distinguish interfaces designed
for episodic/stateless interactions from tanglefaces designed for, or evolved through
extended periods of stateful entanglement. Nevertheless, it goes without saying that
a host of challenges, both in terms of analysis and design, can be better addressed if
past results from synergically related fields are considered.
39
40
Fig. 4.1 Relationships between fields with synergic contribution to CogInfoCom discussed in this
chapter. While the figure shows that CogInfoCom has common motivations with these fields,
and may also benefit from finding suitable adaptations of their respective methodologies, it also
highlights the fact that the borders between the synergically relevant fields themselves have begun
to disappear. For example, as discussed in the chapter, sensory substitution is increasingly regarded
as a form of sensorimotor extension, which in turn increases its relevance to augmented cognition;
at the same time, the motivations of augmented cognition are strongly salient when laid aside and
considered together with those of human-computer interaction; similarly, body area networks have
strong relevance to the concepts of Future Internet and Internet of Things, and so on. Today, it is
not uncommon for papers focusing primarily on one of these fields to be equally well suited to,
and most often accepted to conferences dealing with another: this fact alone supports the message
conveyed by the figure
41
42
AugCog primarily addresses cognitive aspects such as attention, memory, cognitive biases and learning capabilities using cognitive state gauges based on
psychophysiological and neurophysiological measures derived from sources such as
EEG, pupil dilation, mouse pressure, heart rate and many others (St. John et al. 2004;
Stanney et al. 2009). By creating a closed loop system between the user and the
device, measurements on cognitive state can be directly incorporated into both shortterm control and long-term adaptation strategies, allowing for the compensation of
cognitive limitations (Fuchs et al. 2007; Hale et al. 2008).
A research area of growing importance within augmented cognition is that of
cognitive biases (Kahneman 2011; Gilovich et al. 2002). Some authors suggest
that the use of a corrective interface between incoming data and the user can
help eliminate cognitive biasesthat is, systematic errors in diverse aspects of
human thinking and decision makingwhile at the same time supporting new
ways of thinking about problems (Greitzer and Griffith 2006; Griffith and Greitzer
2007; Staal et al. 2008). As pointed out recently by Professor Peter Foldesi and
his colleagues, this can be achieved in several ways, e.g. by modifying input data,
modifying underlying processing and optimization algorithms, or transforming the
way in which output data are presented (Foldesi and Botzheim 2012; Danyadi et al.
2012). Such possibilities are well reflected in the area of human augmentation
detailed in the Gartner hype curves reproduced in Fig. 1.3.
Based on the above, AugCog can be seen as a research area that provides ways to
tighten the coupling between users and systems by improving sensing capabilities
of cognitive states and manipulating operation and feedback information in appropriate ways. The field shows strong parallels with human-computer interactionas
highlighted in its definition quoted abovebut it can also be seen as providing a
complementary perspective in the sense that its name speaks about the augmentation
of (human) cognition as opposed to the augmentation of artificial capabilities (i.e. to
render devices more suited to seamless interaction). When ideas and approaches
from AugCog are applied to the modulation of functionality and information flow in
infocommunication systems, the links between the field and CogInfoCom are clear.
Especially interesting applications are possible when results in AugCog are applied
to the sensing of cognitive states in scenarios with multiple participants and multiple
devices, as suggested, for example, in (Skinner et al. 2013). Such applications are
eventually expected to lead to new kinds of augmented sensors that are capable
of understanding distributed phenomena based on multimodal activities in ICT
networks (possible examples include the filtering of relevant data from vast records
of social network activity on e.g. Facebook or Twitter; or from sensor measurements
in spatial memory systems (Niitsuma and Hashimoto 2009; Niitsuma et al. 2007) or
other distributed physical environments).1
These aspects are strongly relevant to socio-cognitive ICT, a field that was inspired by CogInfoCom as described in Sect. 5.3.
43
44
45
http://www.pnl.gov/coginformatics.
46
interpretation from the cognitive sciences, and merge those findings with modern
network technologies. This aspect of networked technologies described through
the cognitive sciences further strengthens the notion that not only human-oriented,
but also technology-oriented research fields (whether focusing on individual or on
networked technologies) can lay a strong and direct claim to the connotations of the
term cognitive (such a development can be in fact be expected in a wide range of
fields relevant to informatics in the near future).
One research area that seems particularly relevant to the emergence of cognitive
networks is Future Internet, which deals with the theoretical and practical aspects
of creating cognitive networks through which users can interact not only with other
people and computers, but also with everyday objects and abstract computational
capabilities (Tselentis et al. 2010). Two major directions in Future Internet research
are the Internet of Things and 3D Internet.
The Internet of Things (IoT) focuses on the integration and virtualization of the
physical-biological world (hence both physical objects and humans) together with
their capabilities into a comprehensive network through billions of interconnected
sensor technologies.3;4 In a sense, IoT creates a strong physical analogy to the
human nervous system: just as the latter provides humans with a sensor network,
IoT implements a global, physical sensor network [examples of this analogy abound
in the literature, through less direct terms such as the central nervous system
as in the case of the Central Nervous System of the Earth project carried out
by HPas well as in direct architectural considerations (Ning and Wang 2011)].
Inasmuch as Internet of Things enhances the capabilities of humans for effective
communication, it is expected to provide key insights into the field of CogInfoCom
as well. The possibility of treating the human sensory system and the global
sensory system implemented by IoT in a unified framework is also strongly relevant
to CogInfoCom. Although today the engineering and cognitive science based
perspectives through which the two areas are addressed are still markedly distinct,
it is conceivable that similar terminologies and methodologies for investigation will
be developed in the not too distant future. An overview of architectural designs and
applications in IoT can be found in Uckelmann et al. (2011), while (Perera et al.
2014) provides an in-depth survey from an application-oriented point of view.
The concept of 3D Internet (3DI), which is a more recent development, focuses
on the growing expectation of users for high-quality 3D imagery and immersive
online experience (Alpcan et al. 2007; Kapahnke et al. 2010). 3D Internet is seen
as a natural part of the Future Internet concept, because with the appearance of
virtualized interfaces to everyday objects, users will still expect to be able to handle
them in the same ways (importantly, in relation to the same functionalities) as
they do in the physical world. However, once this is established, it also becomes
This means that rather than seeing IoT as a network of objects, it can regarded as humans and
objects that matter to them.
Through the human aspect, other domains such as affective computing and body area networks
when used for infocommunication purposesalso become particularly relevant.
47
clear that 3DI can also be useful for handling 3D representations of content that
is not in its natural form amenable to direct physical representation. Further, once
modifiable representations of both (virtualized) physical objects and abstract data
are created, the ability to navigate (i.e. move along trajectories and remember
those trajectories) between objects and data sets in a way that seems natural also
becomes important. Taken together, these perspectives clearly show that 3D Internet
is about much more than visualization and interaction in 3D: it is a mapping between
the digital world and (physical) 3D metaphors based on highly evolved human
capabilities for representation, manipulation and navigation in 3D. Any technology
that achieves this is an important contribution to the field. As a case in point, spatial
memory systemsproposed by Professor Niitsuma (Niitsuma and Hashimoto 2009;
Niitsuma et al. 2007)which map physical locations and gestures to digital content
and automated functionality are strongly relevant to 3D Internet.
Both the Internet of Things and the 3D Internet are expected to pervade our
everyday lives in the near future. A consequence of both of these research directions
is that users are expected to be able to communicate with both physical and virtual
things (i.e. everyday objects and objectified abstractions) through the Internet,
and also to collaborate with them in ways that depend on both the (artificially
cognitive) capabilities of the objects and on the context (i.e., users will need access
to different components of the network depending on situational context). These
criteria introduce a host of challenges. First, there is the question of augmented
collaboration (i.e., the objects, as viewed from the Internet, can be a combination
of physical and virtual componentssee also Sect. 4.11 for more details on the
relationship between virtual reality and CogInfoCom). Second, there is the question
of scalability (i.e., due to the rapid expansion of content space as an increasing
number of objects try to communicate with the user in the Internet of Things). In
these regards, CogInfoCom has common interests with Future Internet in terms of
selecting the information types that are relevant to the given context, and the modes
of communication that are necessitated by those information types. For further
perspectives on the relevance of Future Internet research to CogInfoCom, readers
are referred to Chap. 12.
48
49
50
51
between artificial systems and the user may not always be directly perceptible by the
available sensory and cognitive subsystems (hence the need for sensor-bridging). In
other cases, the transformation of information from one sensory modality to another
(sensor-bridging) and from one representation to another (representation-bridging)
may provide more effective interaction between the user and the system.
It is clear that the terminology used in CogInfoCom is in many respects a
reflection on the suggestion of Bach-y-Rita and his colleagues. The distinction
between intra-cognitive and inter-cognitive forms of cognitive infocommunications
reflects their first criterion regarding the new terminology (i.e., regarding the
source of the information). The distinction between sensor-sharing and sensorbridging reflects the proposition to distinguish between different types of human
information display. The concepts of representation-sharing and representationbridging are somewhat related to the third point, namely to the question of whether
the transferred information is substituted or augmented.
52
53
54
55
CogInfoCom has common interests with virtual and augmented reality research
along these two aspects. One of the primary goals of CogInfoCom in this respect
is to be able to augment the sensory capabilities of both the human user and the
artificially cognitive system so that they can communicate with each other at a
higher level, thus alleviating the difficulties of human-virtual machine interaction
and the conflicting goals of situation awareness and unencumberment.
Chapter 5
The decision to present results from CogInfoCom fora serves the goal of tractability, and in no
way intends to suggest that important and relevant results have not been published elsewhere.
57
58
In this case, high-level refers to the conceptual level that can be readily interpreted by a cognitive
entity, as opposed to lower-level concepts, which constitute higher-level ones but are not usually
interpreted directly.
5.2 Speechability
59
5.2 Speechability
There is strong empirical support for the view that humans evaluate interactions with
ICT and human-human social interactions based on analogous criteria (Nass and
Yen 2010). As a result, the various aspects of how humans communicate in everyday
social interactions cannot be neglected when augmenting the social capabilities of
cognitive networks.
The modality of speech is a central component of social interactions in everyday
life. For several decades before the emergence of CogInfoCom, researchers have
aspired not only to transmit speech between geographically distant locations, but
also to enable artificially cognitive systems to understand and communicate through
verbal utterances in the most natural (i.e., human) way possible. The reductionist
approach applied to this latter problem has quickly resulted in the appearance
of increasingly specialized sub-branches focusing on a wide range of verbal
phenomena. This fragmentation of a research paradigm that was once fueled by
a single goal is not surprising, given that speech itself cannot be fully characterized
by any single dimension. Rather, it involves co-dependent interactions among such
aspects as word usage, prosody, facial expressions, hand gestures, body postures
and movements, as well as higher-level constraints on the dynamics of dialogue
imposed by social norms and cultural specificities. A further source of heterogeneity
is brought about by the fact that the modality of speech is used for more than a single
purpose: its goal is not merely to support the sharing of information, but also to help
create trust and more generally strengthen those kinds of social/behavioral patterns
that are conducive to the maintenance of society. Removing any of these aspects
from human speech, as is clear from extensive research on virtual communication
agents, renders interaction unnatural and cumbersome.
Speechability aims to reverse the tendency towards fragmentation described
above by attempting to link cognitive linguistics with verbal and non-verbal social
communicative signals through human-ICT tangleactions (Campbell 2012; Benus
et al. 2014b). The latter qualification refers to the emergent effects of long-term coevolution in terms of a broadening range of natural and artificial modalities applied
together towards long-term goals, and consequently the increasing availability
of raw data that is collected for e.g. machine learning or other post-processing
purposes. Such long-term entanglement will lead to capabilities whose scope
extends far beyond the generation and understanding of speech, so as to encompass
application areas such as speech rehabilitation, tutoring for the learning of foreign
languages, or communicational training.
Approaches applied in speechability research are also rooted in the observation
that speech is an embodied phenomenon, i.e. it is interpreted through reference
to physical interaction in specific social contexts. Thus, a unified approach is
suggested that acknowledges the fact that humans and machines have different
embodimentsalbeit ones that are converging as novel cognitive entities. If this
point of view is adopted, then even without long-term tangleactions, speech-related
capabilities and phenomena can be mapped onto artificial ICT capabilities in ways
60
Fig. 5.1 Adaptation of central notions in Fig. 2.1 to the domain of speechability. New, higher-level
artificial capabilities and emergent cognitive entities are created through time as new use cases are
generated through a broadening of artificial sensory modalities and the increasing possibility for
users to co-evolve with them through extended periods of time
that suit underlying differences in embodiment rather than denying them. In this
way, human capabilities are supported rather than copied, and the criterion for
technology to be human-like becomes relatively less important. This explains the
term speechability, as distinct from speech ability (i.e., the ability to speak as
humans do), which reflects the complex, embodied nature of speech phenomena in
cognitive entities. Crucially, while speechability includes speech ability (including
artificial capabilities for speech generation and recognition), it also encompasses a
broader range of tanglefaced applications.
Based on these observations, the central notions of Fig. 2.1 can also be applied
to speechability, as illustrated in Fig. 5.1. The CogInfoCom conference series has
hosted several tracks and sessions on speechability, initiated by Professors Benus,
Campbell and Esposito from Columbia University, Trinity College Dublin, and the
University of Naples II, respectively. Several application examples relevant to the
research area are described in Sect. 5.4 of this chapter.
61
Many recently developed applications can be seen as directed towards augmenting the social capabilities of cognitive network based on the analysis, manipulation
and management of information flow. For example, content and metadata-based
analysis of user activity is used to gain a better understanding of spatially distributed,
crowd-generated phenomenaincluding social-political tendencies, the spread of
ideas and epidemics, etc. Similarly, high-level structural and organizational management of cognitive networks is applied to the optimization of information flow
in critical situations, such as in workflow management and disaster recovery. The
effective treatment of such critical situations necessitates a hierarchical allocation
of both human and technological resources that is crucially enforced by technology,
in much the same way as face-to-face human interaction patterns are governed and
to some extent enforced by social conventions.
Applications such as these can be categorized as operating at the highest level of
entanglement between humans and ICT, or among heterogeneous cognitive entities
in more complex cases (see also Sect. 2.3.1). In this case, collective behaviors
are used to support either individual user interactions, or the prediction/analysis
of collective behaviors and events. Due to the clear relevance of both social
and cognitive capabilities to such applications, the term socio-cognitive ICT
proposed by Professor Hassan Charaf and his research group at the Budapest
University of Technology and Economicshas been increasingly used to describe
them (Szegletes et al. 2014; Fekete and Csorba 2014; Kovesdan et al. 2014).
Interestingly, this choice of description is not without parallels from the past,
even from a technological perspective. Hemingway and Gough described design
challenges relevant to ICT as a conflict between different goals and underlying
values in the fields of software engineering, human-computer interaction and
information systemsthe latter of which is generally concerned with the longerterm impacts of information and communication systems on organizations and
society (Hemingway and Gough 1998). In this interpretation, the term sociocognitive is used to describe all aspects encompassing the ways in which social
phenomena affect, and the ways in which they are affected by ICT infrastructure.
Sharples et al. describe socio-cognitive engineering as a general methodology that
aims to analyze the complex interactions between people and computer-based
technology and then transform this analysis into usable, useful and elegant sociotechnical systems involving technology and social context. In a way somewhat
similar to earlier works, the paper also views this approach as an integration of
fields such as software, task, knowledge and organizational engineering (Sharples
et al. 2002). In general, the important difference between such earlier uses of the
term socio-cognitive and its use in socio-cognitive ICT is that socio-cognitive
ICT focuses on all aspects of network management and networked experience as
opposed to the design-centric perspective adopted in earlier works.
62
63
and message-like representations can also be simplified into iconic ones (if the
structure of events in the world is such that the meaning of the message can still
be unambiguously reconstructed).
Finally, pattern-based techniques often employ both iconic and contextual
approaches, but are also crucially different from the two in that they also take into
consideration characteristic patterns in cognition and behavior that, while often are
extrinsic to the technology that is developed, nevertheless have important effects on
the success of the application.
In this section, example applications presented or published at various CogInfoCom fora are surveyed in terms of the three classes of icon-based, context-based
and pattern-based approaches, as well as in terms of application area. This will help
elucidate some additional properties of these techniques, and how they can be used
together to develop powerful applications.
64
65
with respect to discourse markers, Abuczki demonstrated that when more than a
single lexical meaning is possible, duration with speech and the accompanying
presence of certain manual gestures can serve to disambiguate between them
(Abuczki 2014).
66
67
very simple correlations between structure and meaning, which can later form the
basis of the definition of iconic relationships, then the resulting approach is icon
based. If instead an algorithm is applied to uncovering a set of highly intertwined
relationships between temporally distributed instances icons, potentially at different
levels of granularity, then the resulting approach is context based. As the level of
complexity at which cognitive phenomena are analyzed increases, the latter case
can be expected to become more common; nevertheless, this cannot be regarded as
a general rule.
5.4.2.1 Context in Verbal Signals
One aspect of human dialogue which exemplifies well the importance of context is
the notion of entrainment, which can be explained as the propensity of conversational partners to behave like one another, or as a natural inclination of people
who interact with each other to develop similar, matching, or synchronized forms of
behavior (Benus et al. 2014a,b, 2012). Entrainment is a high-dimensional concept
that can be analyzed from the perspective of temporal changes in any of a number
of lower-level features, including filled pauses, conversational fillers, patterns of
turn-taking, linguistic styleand can even body postures and gestures. In a detailed
investigation, Benus et al. have shown that in many cases, the result of supreme
court case hearingsboth in terms of polarity of decision and in terms of level
of agreement between judgescan be predicted based on features of entrainment
between lawyers and judges (Benus et al. 2014a). While several of the measures
used are local in terms of the length of speech they characterize, many of the same
measurements, aimed at detecting changes in time contribute to the final decision.
Past research suggests that time is essential: humans react highly positively to
communication patterns which change to match their own communication style
through time (and even more positively than if they match to begin with) (Nass
and Yen 2010).
In the area of social signal processing, a recent work by Cheng et al. provides
further examples in which machine learning based aggregations of low-level icons
of voice quality are used to support inference on the popularity of videolecture
content (as demonstrated in the paper, the approach can be combined with data
from other modalities for increased accuracy) (Cheng et al. 2014).
Contextual approaches can be used to effect in other complex dimensions as
well, such as in emotional analysis. In a recent contribution, Galanis et al. applied
a support vector machine based learning approach on a set of over 1500 acoustic
icons defined at various levels of temporal granularity. These features were learned
in conjunction with a small number of contextual parameters such as the gender
and role of the speaker (Galanis et al. 2013). Contextuality, however, in the sense of
the word specified earlier, pervades the entire decision process applied in the study:
a number of low-level representations are combined using machine learning into a
single decision. In another example, Origlia et al. introduce context into a syllablebased emotion tracking algorithm by adding features from neighboring syllables,
and demonstrate that this approach improves system performance (Origlia et al.
2014).
68
Generally speaking, the Bayesian framework of learning can also be very supportive of contextual solutions when the goal is to introduce contextual information
into the decision making process by influencing maximum a posteriori estimates.
For example, Dropuljic and his colleagues applied Gaussian mixture model based
a priori knowledge to emotional classification, both with respect to the distribution
of the training set, and with respect to individual valence-arousal categories within
the training set (Dropuljic et al. 2013). In both cases, the a priori knowledge used to
calibrate categorization results was based on an aggregate characterization of a large
number of independent measurements based on structural units (i.e., the training set
of the estimator) that were defined in the same way as the texts under analysis (in
any other case, they could not have been used to effect within the training set).
69
For a possible unification of the many ways in which the concept of modality is interpreted, see
Sect. 8.1.3.
70
71
Chapter 6
In this chapter, several new research initiatives are described which have been
proposed at various scientific fora on CogInfoCom since the first international
workshop on the field in 2010. These initiatives are discussed in a separate chapter
because they are still relatively young, and continued research can still be expected
to lead to significant developments in their scope and goals. Nevertheless, it is
important to mention them for the reason that all of them reflect the freshness of
perspective and interdisciplinary outlook that is promoted by CogInfoCom.
The chapter is structured as follows. In Sect. 6.1, a recently developed virtual
collaboration platform called VirCA is introduced that in many ways reflects the
conceptual underpinnings of CogInfoCom approaches. As a result of its strong
relevance to the field, the platform has been used in many studies, both as a
development tool and as a tool for investigation, and is extensively referred back
to in later sections of the chapter. Section 6.2 focuses on new perspectives created
by the VirCA platform and the CogInfoCom approach in general in the augmented
virtual reality mediated study of issues in neuroscience. Section 6.3 gives an
overview of how the VirCA platform has been used in ergonomics-related studies of
augmented virtual collaboration. Section 6.4 introduces the emerging research area
of ethologically-based CogInfoCom (EtoCom), in which design goals targeting the
communicational behaviors of devices are supported by concepts from ethology.
Section 6.5 provides insight into to the directions in industrial engineering that can
be supported by CogInfoCom research as evidenced by contributions appearing
at the CogInfoCom conference series. Finally, Sect. 6.6 provides an overview of
the new research initiative of mathability, which focuses on the modeling of
collaborative human-ICT capabilities towards mathematics.
73
74
http://www.sztaki.mta.hu.
75
Fig. 6.1 Structure of the VirCA framework. The VirCA backend is composed of two server
applications: the RTM nameserver and the VirCA system editor. Each component (VirCA clients,
CDs and RCs) appears as a set of CORBA objects in the naming service. The System Editor is
responsible for handling data flow (RTM Data Port) and Remote Procedure Call (RTM Service
Port) connections according to the application scenario that is defined by users through a webbased front-end
76
Fig. 6.2 Two scenarios which demonstrate some of the interaction capabilities afforded by the
VirCA platform. The top subfigure shows a virtual representation of a PR2 household robot (the
original robot was developed for research purposes by Willow Garage). The bottom subfigure
shows a use-case in which multiple usersrepresented through virtual avatar headsare capable
of collaboratively manipulating and functionally controlling the virtual scenario
77
that merges together physical and virtual realities, thus providing an important
testbed environment for new concepts in cognitive networks and, by association,
CogInfoCom research.
The emphasis that VirCA lays on the flexible creation and merging together
of augmented/virtual capabilities is also very much in line with the point of view
adopted by CogInfoCom, and given the opportunity for users to increasingly apply
full-body interaction approaches in collaborative ways, it supports the kind of
experimental data collection that can lead to important new insights in a variety of
research areas. By significantly reducing the time and cost requirements associated
with the deployment of multi-disciplinary testbed environments, VirCA encourages
researchers to focus more on their research questions and less on the technical
details of implementation. This will be demonstrated through several examples in
the following sections.
78
79
80
selected. The key point here is not that unrealistic interactions are a problem, but
rather that it is important to address users expectations encapsulated in implicit
rules.
At the same time, it was noted that in some instances the errors listed above presented themselves not as a typical form of rule-based errors, but as a manifestation
of the curiosity of users and a lack of generalized rules, or more precisely, metaphors
of interaction in virtual environments. For example, many users were curious from
the outset to see whether they could reach through the walls of the scene, and
tried to do so as soon as the experiment began. This curiosity was independent
of the augmented/virtual reality platform used, and was much more relevant to the
uncertainty associated with the use of such platforms in general. Such errors might
aptly be referred to as lack-of-rule-based errors.
81
Fig. 6.3 Schematic view of closed-loop interaction between CogInfoCom entities. The meeting
point at which information is exchanged is referred to as the tangleface
82
From the perspective of entangled interactions between cognitive entities, additional levels of complexity are introduced into this picture due to the complex nature
of the entities themselves. Stability, for example, can be interpreted from a celllevelstructural, hormonalparametric or emotionalbehavioral point of view,
depending on the parts the entity most affected by communication and whether those
parts are natural, artificial or a combination of the two. As a result, the following
notions of CogInfoCom tanglefaces and engines can be introduced:
Definition 6.1. CogInfoCom tanglefaces (entangled interfaces) are the meeting
point through which natural and artificial sensors and actuators influence each other
over extended periods of time.
Definition 6.2. CogInfoCom engines are the modules which participate in determining the dynamic properties of communication between CogInfoCom entities.
From the broad perspective of entangled interaction, the dynamic properties
created through CogInfoCom engines can be derived from a variety of different
considerations. By way of example, it is possible to differentiate between the
following kinds of entangled interactions:
If the control performance of interaction is driven by psychological aspects of
mental processes (e.g., perception, memory, mental models, response times), the
entangled interaction can be referred to as psychology based.
If the control performance of interaction is driven by aspects of human capabilities and/or considerations of safety, efficiency and comfort, the entangled
interaction can be referred to as ergonomics based. It is possible to further
distinguish between interactions that are physical ergonomics based, cognitive
ergonomics based, cognitive ergonomics based, social ergonomics based, or
organizational ergonomics based.
If the control performance of interaction is driven by aspects of human and
animal behavior, the entangled interaction can be referred to as ethology-based.
Based on this interpretation, the role and purpose of ethology-based CogInfoCom
(EtoCom) becomes clearer. At the same time, it becomes evident that a wide
range of factors can be considered when developing applications supporting the
communication between cognitive entities.
83
Fig. 6.4 Schematic view of the relationship between the abstract EtoMotor and other components
providing links between cognitive entities in ethologically driven applications
the interface between a human user and an artificially cognitive system can be driven
by ethology based concepts relevant to behavioral expressions of emotion (Persa
et al. 2012; Kifor et al. 2011). Both applications make use of the same ethologically
based CogInfoCom engine referred to as the EtoMotor (a schematic view of the
relationships between relevant components is shown in Fig. 6.4). For this reason, the
EtoMotor is presented first, while details on the specific applications are provided
subsequently.
(6.1)
where x denotes the short-term, while xI denotes the intermediate level emotional
state vector formulated in terms of the basic emotions listed earlier (happiness,
despair, fear and anger). The model was then transformed into a state-space
representation with a single, aggregate state and was further complemented with
a long-term emotional state as follows:
xA k C 1 D AA xA k C BA uk
yk D C.ExA k C d/
(6.2)
84
Fig. 6.5 Block diagram of the EtoMotor proposed in Szabo et al. (2012)
where
x
xI
B
BA D
0
A 0
AA D
NM
ED II
xA D
and I is the identity matrix. Finally, the model was complemented with a probabilistic extension based on a set of variance values capable of modifying the
mean weights originally stored in matrix B. Based on its final form, the model is
capable of:
1. Associating with each salient external event a level of emotional arousal and
valence that is well suited to the current emotional state (this is the emotional
impact of the input stimulus);
2. Combining this emotional impact together with an aggregate of its own internal
states (short-term, mid-term and long-term) to update its state, and to produce
an output signal that is amenable to emotional interpretation (i.e., both the
internal state and output signal are interpreted as a combination of the four basic
emotions).
An example of how the emotional impact of external eventsinterpreted in the
context of interactions between a dog and its ownercan be specified depending on
the state at any given time is shown in Table 6.1 (for more on recent investigations
on modeling human-dog interactions and using them in ICT, readers are referred to
Sect. 6.4.2.3).
85
Table 6.1 Example of how specific environmental events can be mapped onto a dogs basic
emotional states in terms of arousal, valence and variance
Context
Greets Leaving
O
O
Happiness 3"1 3#1
Despair
3#1 3"1
Fear
2#2
Anger
1"3
Touched
by O
2"1
2#2
2#2
1#2
O init.
int.
2"2
2#1
O rej.
int.
2#1
2"1
O not
as exp.
1#2
1"2
1"2
1"3
Ambi.
sit.
2#1
1"2
1"2
1"2
Play
w. h.
2"1
2#1
2#1
2#1
2"2
1"2
The following abbreviations are used in the table: O (owner), init. int. (initiates interaction),
rej. int. (rejects interaction), not as exp. (does not react as expected), ambi. sit. (ambivalent
social situation), play w. h. (plays with human), nov. stim. (novel environmental stimulus),
unfam. env. (unfamiliar environment)
86
Fig. 6.6 Visual appearance of the Emotional Display Agent used both in VirCA and in the
EtoPhone application. The bottom part of the figure lists the degrees of freedom through which
the emotional display is capable of modifying its behavior
87
Fig. 6.7 Screenshots of the advanced backend for configuring the EtoPhone application. For each
emotional category that is specified, a different set of emotional sub-ranges can be configured
(top left subfigure). Different behaviors can then be associated with each of those sub-ranges (top
right subfigure). The behaviors that are available are configured separately (bottom two subfigures).
Device-specific settings (such as what events are bound to what emotional states) are configured
separately after the application is deployed (Fig. 6.8)
to evaluate attachment behaviors in various dog breeds, and showed that differences
in individual results could be categorized along the scale of insecure to secure
attachment in the terminology used by Ainsworth (Topal et al. 1998). In this section,
a recent adaptation is described for the evaluation of physical-virtual agents through
their behavior in a VirCA-oriented experimental scenario (Vincze et al. 2012).
88
Fig. 6.8 Screenshots of the EtoPhone application running on an iPhone device. The first two
subfigures show the settings screen that allows users to bind input events with emotional states.
The remaining subfigures from top left to right, and top to bottom are static images of the EDA in
the happy, despairing, fearful and angry state
89
room, and the stranger leaving; a condition which was satisfied in the later works to
begin with). The phases were the following:
1. The owner (O) and the dog (D) were in the room, with O passive for the most
part, only beginning to play with D towards the end of the phase.
2. The stranger (S) entered the room, and started to play with D.
3. O unobtrusively leaves the room, leaving D and S alone. S tries to play with D.
4. O returns to the room, and S eventually leaves.
5. O leaves as well, leaving D alone.
6. S returns and tries to engage D.
7. O returns and S leaves. O tries to engage D.
Each phase lasted for 2 min (slight deviations from Ainsworths original study
were allowed in this aspect as well, as in the original test phases had variable lengths
ranging from 1 to 3 min).
Some examples of the kinds of fuzzy rules introduced into the system include
(Kovcs et al. 2009):
If Owner is not in the room, Then Dog increasingly misses Owner;
If Owner-to-Dog distance is small, And Stranger-to-Dog distance is large Then
Dog anxiety level decreases;
If Owner not in room And Dog misses Owner very much, Then Dog going to
door very likely;
If Dog is anxious Then willingness to explore the room is low;
If the room is unknown Then willingness to explore the room is low.
Although fuzzy techniques are well suited for capturing functionality through
linguistic rules, in many cases their applicability is hindered by either a lack or
over-abundance in datapoints. In such cases, fuzzy rule interpolation can be used
90
Fig. 6.9 Fuzzy automaton based architecture for interpolated rule-based behaviors
as a technique to either compensate for, or increase the sparsity of the rule base. In
the current application, the fact that the exponential input space could have easily
led to intractability prompted the authors to apply a fuzzy rule interpolation based
method. The architecture of the applied model is shown in Fig. 6.9.
91
92
93
94
6.6 Mathability
95
Data sources (generic sensors), which help maintain the validity of the representation, e.g. in terms of configuration of manipulators, position of workpieces,
grasping forces, welding currents, and positions of human workers within the
robots workspace;
Orchestration modules (high-level process controllers), which are responsible
for macro-level process control, e.g. with respect to production scheduling and
cell-level PLCs;
Task-level controllers, which are responsible for specific tasks such as path
planning for autonomous forklifts, SLAM functionalities, etc.;
Actuator-level controllers, which include all manner of components capable
of modifying physical states in the environment. Elementary examples include
servo motion controllers, temperature controllers, etc.
It is important to note that depending on the concrete application, not every
level will necessarily be present. In some cases, low-level control cannot be realized
according this pattern, due to particular design considerations, e.g., when the applied
communication technology does not allow for sufficiently fast sampling, or a robot
controller does not offer low-level interfaces. The system designer has to choose the
abstraction level that is appropriate from the technical, theoretical and economical
point of view.
VirCA implements this idea of networked sensing, decision and actuation at a
proof-of-concept maturity level. The VR engine and the corresponding databases
represent the portion of reality that is relevant with respect to the given process. Each
connecting sensor module pushes the gathered information into the information
pool while the control modules work with the high-level, source independent
representation that is accessible in VirCA.
6.6 Mathability
Mathability was initiated by Professor Attila Gilanyi, and defined at the CogInfoCom 2013 conference as a research direction that investigates artificial and natural
[as well as combined] cognitive capabilities relevant to mathematics ranging from
low-level arithmetic operations to high-level symbolic reasoning (Baranyi and
Gilanyi 2013; Borus and Gilanyi 2013; Torok et al. 2013). Importantly, in much
the same way that the focus of speechability (described earlier in Sect. 5.2) extends
further than speech capabilities, mathability focuses not only on human and artificial
mathematical capabilities, but also on the mathematical capabilities of humans and
ICT together with the heterogeneous cognitive entities they give rise to. Thus, one
of the key questions behind mathability is whether mathematical capability can
be understoodand abstractions of it createdso as to facilitate the design and
strengthening of mathematical capabilities in emergent cognitive entities.
An important motivation behind mathability lies in the observation that in
the past decades, even the notion of what qualifies as a proper solution to a
96
mathematical problem has changed. While a few decades ago, only analytically
closed formulae would have been accepted, today it is not uncommon for so-called
granular (i.e. numerical, sub-symbolic) formulations to be seen as equally useful
and acceptable. Although initially contested by many, by today this tendency is
seen as natural as increasingly complex problems need to be addressed in both the
engineering and social sciences (as the importance of this paradigm shift cannot be
stressed enough, further discussions can be found in Sect. 6.6.1). However, despite
these changes, the human capacity to think in numerical terms remains limited, and
analytically tractable deductive methodologies are preferred. The challenge, then, is
how to bridge between these two worlds of analytical and numerical approaches in
a way that is suitable for the problems at hand to be tackled together by cognitive
entities (Baranyi and Gilanyi 2013; Torok et al. 2013). Ideally, ICT devices to be
able to guide users through solutions to mathematical problems, letting users know
how they are able to help along the way (the outlines as well as some characteristics
of such a process are shown in Fig. 6.11). If such processes were possible, humans
involved would be able to make analytical decisions on what deduction route to
further pursue whenever necessary, while the ICT components involved would be
focused on applying the numerical tools that are most suited to the given context.
A simple example in which the solution sets to functional equations are found
through human intervention is provided in Sect. 6.6.2.
Fig. 6.11 As entangled processes within and between human-artificial cognitive entities are
applied to problems in mathematics, the relative strengths of the entities involved are expected
to be relied upon at different times and for different subtasks. An important consequence is that
the problem spaces that can be handled in such a way are qualitatively different from problems
involving only analytical or only numerical approaches; i.e. instead of merely strengthening
the analytical or numerical capabilities of the entities involved, such processes create a new
dimensionindicated in the figure as a rectangular areathat is created through the combination
of such capabilities
6.6 Mathability
97
98
6.6 Mathability
99
fi .pi x C qi y/ D 0
.x; y 2 X/
(6.3)
iD0
.x; y 2 X/:
It is obvious that the so-called Pexider equation, as well as the square-norm equation
(also known as the Jordanvon Neumann, or parallelogram equation) and the Jensen
equation are also special cases:
f .x C y/ D g.x/ C h.y/
.x; y 2 X/:
f .x C y/ C f .x y/ D 2f .x/ C 2f .y/
f
xCy
2
D
f .x/ C f .y/
2
.x; y 2 X/I
.x; y 2 X/
.x; y 2 X/
100
.x; y 2 X/:
(6.4)
.x; y 2 X/:
(6.5)
.x; y 2 X/:
(6.6)
.x 2 R/
(6.7)
with real constants a0 ; : : : ; an satisfy Eq. (6.5). Similarly, with a c 2 R, the classical
monomial
f .x/ D cxn
.x 2 R/
(6.8)
satisfies (6.6).
It is also well-known, however, that both classes of functional equations
[i.e. those of Eqs. (6.5) and (6.6)] have solutions that are different from (6.7)
and (6.8): hence, the concepts of polynomial and monomial functions are a
generalization of the concepts of polynomials and monomials. Further, it can
be shown based on Eq. (6.4) that the polynomial equations (6.5) and the monomial
equations (6.6) are in turn special cases of the class of equations given in (6.3). This
means that systems of equations of the type introduced in Eq. (6.3) potentially have
polynomial and/or monomial solutions.
A necessary and sufficient condition for all solutions to a functional equation
of form (6.3) to be available as a sum of monomials can be obtained from two of
Szkelyhidis theorems, which are re-iterated here:
Theorem 6.1 (L. Szkelyhidi). Let X and Y be linear spaces over the field of the
rationals, p0 ; : : : ; pnC1 , q0 ; : : : ; qnC1 be rational numbers and let
.i/
Mk W X ! Y
6.6 Mathability
101
n
X
.i/
Mk .x/
.x 2 X; i D 0; : : : ; n C 1/
(6.9)
kD0
fi .pi x C qi y/ D 0
.x; y 2 X/
iD0
.i/
j kj
.i/
pi qi Mk .x/ D 0
(6.10)
iD0
.i; j D 0; : : : ; n C 1; i 6D j/:
(6.11)
fi .pi x C qi y/ D 0
.x; y 2 X/
iD0
n
X
.i/
Mk .x/
.x 2 X; i D 0; : : : ; n C 1/
(6.12)
kD0
where
.i/
Mk W X ! Y;
.i D 0; : : : ; n C 1; k D 0; : : : ; n/
j kj
.i/
pi qi Mk .x/ D 0
(6.13)
iD0
102
The interpretation of the two theorems is that if the condition given in Eq. (6.11)
is satisfied, then all solutions to the functional equation can be expressed as sums
of monomial terms. However, if the condition is not satisfied, then the exclusivity
of the solutions expressed as sums of monomial terms is not guaranteed. This has
consequences when different equations in a system of equations have different sets
of solutions and appropriate intersections have to found, as demonstrated below.
.x; y 2 X/
.x 2 X/:
(6.14)
(6.15)
Here the coefficients in the first equation satisfy (6.11), while those of the second
one do not. If the program received as input the second equation alone, it would
yield only monomial solutions of degree 0 (i.e. constant solutions). However, just
by looking at the system of equations, even a human reader could easily tell that the
second equation imposes no real restriction on the general solution: any function
that satisfies the first equation will also satisfy the second one. Thus, the developed
program will signal to the user that there is an ambiguity, and will allow the user to
choose the solution that is to his or her liking.
Part III
CogInfoCom Channels
104
III
CogInfoCom Channels
such a model is rendered difficult due to the fact that the function which links all
possible combinations of generation parameter values to perceptual qualities is both
difficult to compute and also practically impossible to invert. One possible solution
to this challenge is to apply a tuning model that allows users to interactively explore
the parametric space used to generate CogInfoCom messages. The chapter proposes
the spiral discovery method (SDM)a tuning model that fulfills these requirements
and also empirically aims to support flexibility and interpretability.
Finally, in Chap. 11, the possible evolution of CogInfoCom channels is considered based on various dynamical perspectives of biological and human communication. As this subject area has more relevance to the long-term co-existence of
humans and ICT than to episodic interactions, its details are relatively less well
understood. Nevertheless, the chapter aims to highlight some of the more important
challenges to be addressed by future research, and introduces some concepts which
may help guide future work.
Chapter 7
105
106
Type of communicationwhether the same sensory modality and representation is used on the two ends of communication (hence the terms sensorsharing, sensor-bridging, representation-sharing and representation-bridging, as
described earlier in Sect. 2.2.2
Directness of communicationwhether or not the cognitive entity knows that
it is communicating, and whether the act of communication requires its active
involvement in terms of the dedication of cognitive resources
Voluntariness of communicationwhether communication is initiated because
the cognitive entity wants to accomplish a task, or because another entity, such
as a user, application, device or network makes a query
Temporal reference of communicationwhether the message that is communicated pertains to the past, present, or future.
Some of these dimensions are currently much more relevant to human experience
than to any technological consideration, while others can be equally relevant to
humans and artificial cognitive entities. For example, the level of directness and
voluntariness of communication would be of particular interest to users, for whom
too many unsolicited queries could easily be perceived as a nuisance. At the
same time, the type of communication can be relevant irrespective of the kind of
cognitive entity under consideration. For example, representation-sharing could be
preferred over representation-bridging in networked applications where low power
consumption is important.
While the list of factors provided above can by no means be considered
exhaustive, it can be reasonably expected, or has otherwise been made clear in past
research, that all of them contribute to user experience and usability in significant
ways. Further, it can be expected that the subtleness of variation within these factors
of communicative context can add to users appreciation of the cognitive capabilities
with which they are communicating.
On the other hand, it is important to observe that in many cases, the above
descriptions of relevant factors require further clarificationeither because of the
ambiguity of the terms used, or as a result of their multi-faceted qualities. For
example, the following questions arise naturally:
What is meant by level of cognitive capability. What determines whether two sets
of cognitive capabilities are equivalent to each other? Although in Chap. 2 it was
highlighted that this may be dependent on the context and the application, the
added value of distinguishing between such cases is not always apparent [a few
attempts have been made to show either that intra-cognitive communication can
be useful in crowdsensing applications, or that a partitioning of users based on
cognitive capability can be useful (Balata et al. 2014; Blum et al. 2014)].
What is meant by sensory modality? If a user is reading through touch
(as in Braille writing), does communication occur through the visual or tactile
modality? If a user receives text information in once instance, and graphical
information (as in a bar chart or pie chart) in another, are both instances of
107
108
109
Perhaps this explains why some researchers have not allowed the ethymological structure of their
terminologies to influence their interpretation.
110
it has been suggested that auditory icons and earcons might be used in conjunction
to create more effective interfaces (Gaver 1997; Hearst 1997; Mustonen 2008). As
an extension to this idea, a compositional relationship, such that abstract earcons are
composed of auditory icons at lower levels of granularity, has been suggested in e.g.
Hermann and Ritter (1999), Csapo and Baranyi (2012d), and .
A second level of unification has emerged between conceptually driven and
data-driven communication. An example of the latter in the auditory domain is
sonification, in which data is used to create reproducible patterns of auditory signals.
Hermann defines sonification as the data-dependent generation of sound, if the
transformation is objective, systematic, reproducible and can intentionally be used
with different data (Hermann 2008; Hermann et al. 2011). It is important to note in
this regard that the key distinction between sonification and auditory earcons/icons
is made in terms of whether the signal is data or concept-generated, and not in
terms of any kind of strictly defined perceptual consideration; as noted in Hermann
(2008), a sonification can include iconic and message-like elements if at least part
of the data is discrete and/or conceptual.
In the framework of CogInfoCom channels, these two levels of unification are
reflected in the modality-independent concepts of CogInfoCom streams, CogInfoCom icons and CogInfoCom messages, as well as in a hierarchical relationship
between the three. Additionally, questions of how data or information is mapped
onto these communicative elements are treated as a separate question from the
decision to use one or the other. More specifically, the conceptual associations
underlying CogInfoCom channel based communication are chosen without regard
to whether those associations will be encapsulated in icons or abstract messages.
This distinction arguably helps to more clearly delineate the tasks that need to be
performed when designing and implementing new communication patterns between
cognitive entities. The remainder of this chapter focuses on the structural aspects
of CogInfoCom channels, while the mapping of data/information is considered in
Chap. 9.
111
112
113
114
115
Remark 7.5. Although many researchers have pointed out that message-like representations often require training, two points can be made to show that this is by no
means a general rule:
It has been demonstrated, for example in the domain of auditory interfaces,
that messages structured appropriately and taught to users through suitable
training methods could often be just as easily retained as update icons (for an
overview of such studies, see Csapo and Wersenyi 2014). Further, when messages
carry information that is based on individual interpretations of its constituent
icons, understanding the intended meaning can translate to understanding the
constituent icons.
Even if complex training suites were required to teach users the meaning of
a message, the software and technology related challenges associated with the
processing of messages in artificial cognitive entities will not necessarily be as
significant.
From the point of view of CogInfoCom, it makes sense to regard CogInfoCom messages as being comprised of more elementary icons, because this
view encourages the joint use of the two classes of signals without any kind of
restriction. A degenerate CogInfoCom message with no icons does not contradict
these definitions, and neither does the use of separate CogInfoCom icons which do
not perceptually and/or conceptually integrate into a single message.
In the same way as was the case with CogInfoCom icons, due to the sensory
system independent and technology-agnostic point of view adopted in the definition,
a virtually infinite range of signals can be cited as examples of CogInfoCom
messages as long as they create abstract associations to some kind of relevant
information. Some examples relevant to human perception include:
Eyecons: visual messages used to convey information that has an abstract
relationship with the message structure. Although this term has not been defined
outside of the scope of CogInfoCom, it arises as a generalization of earcons,
smicons and other message-like elements to the sensory modality of vision.
Earcons: non-verbal audio messages used in the user-computer interface to
provide information to the user about some computer object, operation, or
interaction (Blattner et al. 1989). Contrary to the original approach of Blattner
et al., who treated both representational earcons and abstract earcons as earcons,
today the term is used exclusively to refer to message-like communication.
Smicons: scents used to convey information that has only an abstract relationship with the data it represents (Kaye 2004).
Hapticons: Enriquez and MacLeans terms of haptic icons and hapticons are
separated into two different levels in CogInfoCom. As described earlier, the term
haptic icon is used when referring to haptic signals that are iconic, and the term
hapticon is used when referring to haptic messages that convey information that
has an abstract relationship with the hapticon itself. By adopting this view, it can
be seen that Brewsters tactons can be considered as specific kinds of hapticons
for the tactile sense (Brewster and Brown 2004).
116
117
Chapter 8
Based on the preliminaries of the previous chapter, this chapter provides a holistic
view of CogInfoCom channels that unifies syntactic and semantic aspects. CogInfoCom channels are defined as a structured unification of icons and messages
with perceptual interpretations of meaning. This definition uncovers previously
overlooked features of the concept of modality. It is demonstrated in the chapter
that the parametric structure behind CogInfoCom channels can be represented in a
tensor-algebraic form that is canonical and interpretable. It is argued that the use
of this representation can be advantageous in that it allows researchers working on
similar problems to share and reproduce each others results. Some of the tools for
working with CogInfoCom channels developed in later chapters are also based on
this representation.
The chapter is structured as follows. In Sect. 8.1, the concepts of icons and messages are extended based on their applicability towards communicating semantic
meaning. The concepts developed in the section are used to formulate a definition
of CogInfoCom channels, as well as to provide a definition of modality which
takes into consideration both the technology-agnostic perspective of CogInfoCom,
as well as the fact that communication is a multi-lateral rather than a uni-lateral
issue. Section 8.2 introduces a tensor algebra based canonical and interpretable
representation of CogInfoCom channels. Finally, the results of the chapter are
briefly summarized in Sect. 8.3.
119
120
121
Fig. 8.1 Hierarchical relationship between icons, compound icons, messages and messagegenerated concepts in CogInfoCom
122
123
Based on this definition, it is clear that the proposed concept of modality reflects
a specific relationship between communicating entities. It reflects a relationship
because, through the input and output device, it takes into consideration the sensory
systems of both entities, as well as the encoding of messages that are transferred
between them. It also reflects a specific relationship because the context of the
specific instance of communication, as reflected by the noise level in the channel, is
also taken into consideration.
As a result, if different parameter values (or even different parameter types)
are used to generate two sets of CogInfoCom messages, then for all practical
purposes, they can be regarded as belonging to separate modalities. This point
of view acknowledges that an application can be multimodal even if it only uses
e.g. audio feedback, given that the feedback messages can nevertheless be multifaceted, requiring high levels of attention or processing power.
Similarly, two applications can be regarded as using different modalities if they
use different output or input devices, or if the noise conditions of communication are
different. This point of view acknowledges that various measures of effectiveness
and efficiency (including test suites, empirical usability tests, etc.) can no longer be
automatically seen as valid if any of these devices are changed for different ones,
or if the environment is different. These points highlight the fact that no amount
of testing will produce results that are generalizable to environments with different
parameters.
124
rough they are, based on their level of tonality or atonality, etc. The results of these
comparisons will generally be different depending on the perceptual dimension that
is considered. For example, if there are three earcons (A, B and C), then earcon A
may be louder than earcon B, which in turn may be louder than earcon C; while
earcon B may be perceptually more rough than earcon C, which may in turn be
more rough than earcon A. It may also be the case that the three earcons cannot
be reliably ordered along a given perceptual dimension, as demonstrated in Barrass
(1998). This observation is general and carries on to other sensory systems.
A significant challenge in the design of CogInfoCom channels is posed by the
fact that the generation parameters used to synthesize CogInfoCom messages have
no directly specified, general relationship with perceptual qualities. In other words,
there will generally be no analytical formula available to specify this relationship.
This will be true even if, for example, the parameter types and synthesis method are
fixed, due to differences in other aspects of the modality that is used. Nevertheless, a
link between generation and perception will always exist (even if it is hidden to the
application designer), that is, we can at least assume that for any given CogInfoCom
modality, there exists a relationship that maps perceptual orderings onto orderings of
parameter sets. To describe these orderings and relationships, we use the following
terminology:
Let the parameters used to generate individual CogInfoCom icons in CogInfoCom channels be represented by an H-dimensional generation vector of real
values, f 2 RH .
Let the perceptual parameters that describe a given CogInfoCom icon be
represented by an N-dimensional perceptual gradation vector of natural values
p 2 NN .
Let the relationship between the perceptual vector and the generation vector be
represented by a parameter-generating function F W NN ! RH , which takes
as its input a perceptual gradation vector, and returns a generation vector that
can be used to synthesize a CogInfoCom icon in the CogInfoCom message that
represents the perceptual qualities contained in the perceptual gradation vector.
The primary challenge in designing CogInfoCom channels is how to specify
and represent the parameter-generating function. If the function could somehow
be specified, that would allow cognitive entities to communicate by encoding the
desired perceptual gradation into a set of parameters which can be in turn used
to generate appropriate messages. Based on the definition of modality, the same
parameter-generating function could be used as long as the input and output devices,
as well as the environment of communication does not change.
In order to address such challenges, however, it is necessary that the parametergenerating function in question should have a formal representation. Such a representation would ideally allow for design choices to be shared among researchers and
developers, and for existing channels to be manipulated so as to suit the modality
that is used for communication.
125
(8.1)
such that Pn is the number of unique discretization points along the nth dimension,
and n D 1::N. A discretized function of this kind can always be represented by an
.N C 1/-dimensional tensor F D.G/ of size P1 : : : PN H
D F.gp1 ;p2 ;:::;pN /
FpD.G/
1 ;::;pN
(8.2)
such that F.gp1 ;p2 ;:::;pN / is the output vector of F corresponding to the input vector in
D.G/
point gp1 ;p2 ;::pN on grid G, and Fp1 ;::;pN is the subtensor of tensor F D.G/ obtained by
fixing the index of the first i dimensions (i D 1::N) to the pi th gradation along the ith
dimension of discrete grid G. F D.G/ is referred to as a discretized bounded-input
POMF.
126
In the following section, it will become clear that the parameter-generating functions associated with CogInfoCom channels can be represented using the concept
of discretized bounded-input partially ordered multivariate functions (DBIPOMFs).
Later, in Sect. 10.2, it will also be shown that such functions can be transformed
into a canonical representation that allows for the interpretation of important tensor
algebraic concepts, such as rank and condition number. More importantly, it will
also allow for the comparison and manipulation of existing solutions using tensor
algebraic operators.
(8.3)
where F.gp1 ;p2 ;:::;pN / is the output vector of F corresponding to the generation
vector of H dimensions in point gp1 ;p2 ;::pN on grid G, and the generation vectors
obtained as the output of F are represented along the .N C 1/th dimension of
tensor F . Each dimension in the tensor is indexed by the gradation specified along
the corresponding dimension of grid G, save for the last one, which is indexed by
the position of a specific parameter in the generation vector.
The H-dimensional generation vector can be used as input to synthesize a
CogInfoCom icon within the CogInfoCom message that carries the perceptual
values contained in the perceptual gradation vector p.
127
Fig. 8.2 Structure of a CogInfoCom message in a CogInfoCom channel, based on the perceptual
gradation vector and the generation vector. The synthesis method accepts a set of generation
parameters based upon which the structural and temporal properties of CogInfoCom icons can
be derived (synth refers to the synthesis algorithm used for CogInfoCom icon generation). The
succession of CogInfoCom icons through time creates CogInfoCom messages
128
Chapter 9
Based on the results of the previous two chapters, this chapter focuses on techniques
that can be used to map semantic meaning onto CogInfoCom channels. This
mapping task is formulated in terms of finding appropriate perceptual concepts
and a corresponding set of generation parameter typesfor the establishment of
useful semantic associations. In this way, the chapter develops a set of tools that can
support the development of readily interpretable CogInfoCom channels.
The chapter is structured as follows. Section 9.1 introduces a denotational
framework that is applied to the formal treatment of concept mapping. Parts of
the framework have been used in the past [i.e., under the name Object-AttributeRelation (OAR) model (Wang 2008; Yao 2009)], while others are provided as
extensions based on definitions in previous chapters. In particular, formal definitions
are provided of CogInfoCom message generated perceptual concepts, perceptual
evaluation functions, and interactive concepts. Based on the framework, Sect. 9.2
introduces various forms of direct and analogical mapping between CogInfoCom
messages and CogInfoCom message generated perceptual concepts. Future perspectives for automated mapping are discussed in Sect. 9.3.
129
130
(9.1)
(9.2)
(9.3)
where
O O ; A A ; Rc O A; Ri C0 C; Ro C C0
(9.4)
such that O and A are the power sets of O and A, respectively; and Rc , Ri and Ro
are a set of internal, input and output relations, respectively. A concept is said to be
defined over a particular object o, or a set of objects o, if that object or set of objects
is included in the objects of the concept. Formally:
DO.c D .O; A; Rc ; Ri ; Ro /; o/ , 8o 2 o W o 2 O
In case an attribute has a scalar numerical value associated with it, that value is
returned by the fval W A ! V 1; 1 function. This number is referred to as
the value of the attribute. The number of possible values that can be assumed by
an attribute (or in an infinite case, the cardinality of the set from which it can take its
values) is referred to as the (spatial) resolution of the attribute, and is denoted by
sres.a/ for attribute a. The frequency at which the value of an attribute is updated is
referred to as the (temporal) resolution of the attribute, and is denoted by tres.a/
for attribute a.
131
Definition 9.3. The intension of a concept is defined as the intersect of the sets of
attributes of all objects within the concept:
a .c / D a .O; A; Rc ; Ri ; Ro / D
#O
\
.Aoj /
(9.5)
jD1
(9.6)
that is, Aoj refers to all attributes within the internal relations associated with
object oj . Whenever the intension of a concept is not empty and contains one or
more attributes, we refer to those attributes as the defining attributes of the concept
(this is in recognition of the fact that those attributes can be used to characterize all
objects in the concept). The cardinality of the set of defining attributes of a concept
is referred to as the dimensionality of the concept, and is denoted by dim.c/ for a
concept c.
Definition 9.4. The set of instant attributes of a concept c is defined as the union
of all attributes of all objects within the concept:
A .c / D A .O; A; R ; R ; R / D
c
#O
[
Aoj
(9.7)
jD1
(9.8)
that is, Aoj refers to all attributes within the internal relations associated with
object oj .
Based on these definitions, a number of relational operators can be defined
in order to formally model conceptual knowledge-based systems. For details,
the reader is referred to Wang (2008), in which a set of operators is defined
to encapsulate human reasoning in a formal concept algebra based denotational
language and to manipulate knowledge in artificial cognitive entities.
132
Definition 9.5. CogInfoCom message generated perceptual concepts, or perceptual concepts (Cmg ) of a set of CogInfoCom messages X are a set of abstract
concepts that belong to the context defined over those messages as objects, and have
at least one defining attribute:
o
n
Cmg .X/ D c
mg D .X; A; R/; 9a W a 2 a .cmg /
(9.9)
(9.10)
Remark 9.1. As described earlier in Sect. 8.1.1, perceptual concepts can be identified in terms of the specific sensory system to which they pertain. Hence, in the
case of earcon generated concepts (where X is defined as the set of all earcons),
examples would include smoothness, roughness, consonance and dissonance (given
that earcons can be described using such concepts). Similarly, concepts such as
sweetness, perfume, and chocolate could be relevant to the olfactory domain (given
that smicons can be described using such concepts). All objects (i.e., earcons and
smicons) that can be characterized by these concepts can belong to the relevant
formal concept (whether they actually do depends on whether they are considered
at all in a particular implementation).
Remark 9.2. The notion of a perceptual being defined over a message is analogous
to the earlier notion of a concept being defined over a particular object, or set of
objects (cf. Definition 9.2).
In many cases it is advantageous to broaden this definition by ensuring the lowest
possible cardinality for the set of defining attributes (i.e. finding concepts with the
lowest possible dimensionality). In this case, the set of objects (i.e., CogInfoCom
messages) within the concept is augmented until no further reduction in the number
of defining attributes is possible:
Definition 9.6. A CogInfoCom message generated perceptual concept is fully
specified if no augmentation of the set of messages within the concept can lead
to further reduction in its dimensionality:
8
9
D .X; A; R/; 9a W a 2 a .c
mg /; >
>
>
>
>
>
>
>
0
>
6
9x;
9c
.X
[
x/
W
<
=
mg
Cmg .X/ D cmg
>
>
>
dim.c0mg .X [ x// > 0;
>
>
>
>
>
>
:
;
a .c0mg .X [ x// a .c /
(9.11)
133
There are two reasons why it is useful to consider fully specified CogInfoCom
message generated perceptual concepts:
1. Given that the remaining defining attributes characterize all messages within the
concept, and also that this group of messages is as broad as possible, any of the
messages can in theory be used within a CogInfoCom channel associated with
one of the defining attributes.
2. By reducing the cardinality of the intension to a minimal number, the cardinality
of the set of instant attributes of the CogInfoCom concept will simultaneously be
increased to the broadest set of attributes possible. This will have consequences
in high-level direct mapping, as described later in Sect. 9.2.
Naturally, in practical implementations a number of considerations can lead
to approximate applications of this definition. For instance, the computational
requirements of an application might cause an application designer to settle for
an acceptably small cardinality of defining attributes rather than considering all
possible messages with which the existing concept can be augmented.
Definition 9.7. A perceptual evaluation function is a function that associates a
real number between 0 and 1 with a CogInfoCom message x 2 X and a CogInfoCom
message generated perceptual concept defined over it:
feval W X Cmg .X/ ! 0; 1
(9.12)
such that the output value represents the degree to which the given CogInfoCom
message generated perceptual concept represents the CogInfoCom message. The
output value of a perceptual evaluation function is referred to as the perceptual
value of the message with respect to the given concept.
Remark 9.3. It is important to note that the output of function feval is a characterization of the perceptual relationship between a message and a concept defined
over it, as opposed to the fval function defined earlier, which provides an objective
characterization of attribute values.
Remark 9.4. The output of function feval is equivalent to one dimension of the
perceptual gradation vector defined earlier in Sect. 8.2.1. This means that if it were
possible to invert feval , one could obtain the CogInfoCom message (or set of CogInfoCom messages) which generate the specified perceptual value. Unfortunately,
such an inversion is not possible in general, as argued in the following remark.
Remark 9.5. The output of feval will in general be dependent not only on the given
message and perceptual concept, but also on the cognitive entity performing the
evaluation, as well as on other circumstances reflected in the CogInfoCom modality,
as described in Sect. 8.1.3. This fact shows that even if feval could be specified in
analytic terms, doing so would be impractical for realistic applications. As a result,
134
the inversion of feval as mentioned in the previous remarkis also not possible in
general. This suggests that while the feval function might be useful for empirically
gauging the class of CogInfoCom messages that are suitable to produce certain kinds
of perceptual effects (a task of primary importance!), it will not in general be suitable
when searching for specific CogInfoCom messages with specific perceptual values.
<
c D .O; A; Rc ; Ri ; Ro /; 8x 2 X
: 9a 2 a .c/ W
fval .a/
fval .a/
9
>
=
>
/ feval .x; cmg / ;
(9.13)
135
Norman refers to any artificial device that is capable of storing, displaying or modifying
information as a cognitive artifact. Norman writes about different levels of directness with respect
to such devices, to qualify the relationship which exists between a users actions and the changes
these actions create in the stored information (Norman 1991).
136
137
necessarily parameters that are related to the generation of the transmitted concept,
and also do not generally yield perceptual evaluations that would in any way
resemble the original concept.
One example of structural mapping would be the use of graphical triangle
representations to convey information on the height, weight and age of a group
of individuals. A triangle representing a certain person could use a color coding of
age, and the height and skewness of the triangle might reflect the height and weight
of the individual, respectively. Clearly there is no readily available semantic link
between age and color or weight and skewness, however, the dimensionality of the
two representations is the same and can therefore be brought into correspondence.
Another example of structural mapping can be taken from the domain of
vibration-based interaction with mobile devices. As the number of possible types
of vibration is limited by hardware constraints, a bijective relationship between
vibration type and information type is generally impossible to achieve. Hence, while
some information can be conveyed through spatial resolution (e.g., more complex
vibrations represent a more complex information types), generally users are forced
to rely on the temporal resolution of the vibrations. A vibration perceived on an
hourly basis, for example, would be interpreted differently than a vibration that
occurs once a day.
Structural mapping can also be useful in augmenting the sensing capabilities
of artificially cognitive entities. The structural parameters associated with various
activities can for instance allow a mobile device to determine whether the user is
walking, running, reading a book or making coffee, irrespective of semantically
grounded a priori knowledge.
Definition 9.12. Mapping between a concept c1 and two or more CogInfoCom
messages msg1 ; : : : ; msgn is corroborative stimulation (co-stimulation) if association between the concept and messages is created through temporal simultaneity.
Remark 9.10. Corroborative stimulation can rely both on natural associations (e.g.,
this is the case when the contact sounds of a remote probe are fed back to the
teleoperator along with a video stream to provide augmented information on what
is happening in the remote environment), or on associations that are formed through
training (e.g., by simultaneously providing contact sounds and an abstract graphical
representation, in which case the contact sounds would support the user in learning
the meaning of the abstract graphical representation).
Definition 9.13. Mapping between a concept c1 and a CogInfoCom message msg
is scenario-based if the data flow within CogInfoCom messages is specified based
on the structure of some (physical or virtual) interaction.
In the physical case, conceptual mapping can be formally modeled using an
interactive concept ciact of c1 , such that a defining attribute a1 of ciact is mapped
onto a defining attribute a2 of a perceptual concept c2 that is defined over msg:
fval .a.ciact .cmsg1 /// D ) feval .msg2 ; c2 / D
(9.14)
138
139
Chapter 10
This chapter investigates the need for designers and users to be able to customize
CogInfoCom channels. It is argued that the availability of tools for this purpose
is important due to the specificities of the CogInfoCom modality that is used
especially with respect to the input device and the noise level characteristic of the
transfer medium. However, the task of creating such a model is rendered difficult
due to the fact that the function which links all possible combinations of generation
parameter values to perceptual qualities (referred to as feval in Chap. 9) is both
difficult to compute and also practically impossible to invert. One possible solution
to this challenge is to apply a tuning model that allows users to interactively
explore the parametric space used to generate CogInfoCom messages. The chapter
introduces the spiral discovery method (SDM)a tuning model that fulfills these
requirements and also empirically aims to support flexibility and interpretability.
1
Since the feval function introduced in Chap. 9 is not analytically available and also not invertible,
tuning will by necessity resemble a searching process rather than a computation of parameter
values from a quantification of a desired perceptual quality.
141
142
From an engineering perspective, this point can be appreciated if one considers how difficult it
used to be to manually tune a radar locator. Although today this is a task that can be performed
through automation, a few decades ago it was a task for humans to tune the elevation, azimuth and
carrier frequency of the radar locator. This required the use of just three controls; nevertheless, it
was a hugely difficult task that demanded much practice and attention.
It is also important to note that in general it is difficult to specify in the first place what perceptually
optimal CogInfoCom messages would be like.
143
From the users perspective, the direction of the hyper-spiral is a transparent parameter, as it will
coincide with the direction of the principal component of a set of control points.
Note that the interactive scheme described here is somewhat relevant to the paradigm of
Interactive Evolutionary Computation, as described by Takagi (2001), in the sense that the users
subjective evaluations are used to guide an iterative search process.
144
path of
discovery
Fig. 10.1 The schematic in this figure shows how SDM, the proposed interpretable tuning model
works. Instead of using the original parameter spacedenoted in this case by p1 ; p2 and p3 (it
should be noted that the number of parameters in general can be much more than 3)it is always
possible to describe a rough approximation of the original parameter space using the parameters of
a discovery spiral. These parameters can be set transparently, without the users intervention (for
example, the direction of the spiral can correspond to the direction of the principal component of
a set of control points already specified by the user; parameters r, d and can be set so that the
hyperspiral covers a representative portion of the parameter space, as required by the application,
etc.). The user, in turn, is only required to manipulate two parameters: a distance parameter in
terms of discrete steps along the hyperspiral (denoted by s in the figure), and a velocity parameter
which specifies the size of the steps (denoted by v in the figure). An important idea behind SDM
is that the users perceptual sensitivity will generally have some tolerance, thus it is sufficient to
discover a structured subspace of the parameter space rather than its entirety
mD1
lD1
F.x; y/ D S wm .xm / ul yl
(10.1)
145
where:
S 2 RI1 :::IN H is a core tensor of finite dimensions
Each wm .xm / D wm;1 .xm /; wm;2 .xm /; : : : ; wm;Im .xm / is a vector of univariate
weighting functions defined over a partially ordered set
Each ul yl D ul;1 yl ; ul;2 yl ; : : : ; ul;Il yl is a vector of univariate weighting
functions defined over an ordered set
xm denotes the input in the mth partially ordered input dimension
yl denotes the input in the lth ordered input dimension
POMFs which can be written in this form are referred to as partially ordered
multivariate tensor product (TP) functions (POTPFs).
Property 10.1. Tensor product (TP) functions are multivariate functions which can
be represented by a core tensor S and a set of univariate weighting functions wn .xn /
(Baranyi 2004, 2014; Baranyi et al. 2006). The concept of POTP functions is similar
to this class of functions, but it allows certain input dimensions to be partially
ordered. Hence, we can say that the original concept of TP functions is a special
case of POTP functions, in which all input dimensions can be ordered.
Definition 10.2. Discretized partially ordered TP functions are partially ordered
multivariate functions (POMFs) that are discretized along an N-dimensional grid,
G D fgp1 ;p2 ;:::;pN 2 Ra1 ;b1 a2 ;b2 ;:::;aM ;bM Nc1 ;d1 c2 ;d2 ;:::;cL ;dL gPpnnD1 , and can be
expressed in the following tensor product form:
M
mD1
lD1
D.G/
Ul
F D.G/ D S WD.G/
m
(10.2)
where
S 2 RI1 :::IN H is a core tensor of finite dimensions
D.G/
Wm D wm;pm .xm / is a matrix in which the pm th row represents the pm th vector
of univariate coefficient functions along the mth continuous input dimension
Each coefficient function in the pm th row of the matrix is evaluated at the pm th
value of the mth discretized dimension (xm )
D.G/
Ul
D wl;pl yl is a matrix in which the pl th row represents the pl th vector of
univariate coefficient functions along the lth discrete input dimension
Each coefficient function in the pl th row of the matrix is evaluated at the pl th
value of the lth discretized dimension (yl )
146
D.G/
M
L
NC1
D S Wm Ul NC1 ULC1 D S Xn
mD1
lD1
nD1
(10.3)
in which:
.m/
.m/
.l/
.LC1/
3. ULC1 D .u1
; : : : ; uILC1 / is a orthonormed matrix of size .H ILC1 /
4. S is a real tensor of size I1 : : : IN H, the subtensors Sin D of which have
the following properties:
all-orthogonality: any pair of the subtensors of S are orthogonal, i.e. for all
possible values of n; and subject to :
< Sin D ; Sin D >D 0
(10.4)
ordering: All of the subtensors of S along any given dimension n are ordered
according to their Frobenius norm, i.e. 8n D 1::N C 1:
jjSin D1 jj jjSin D2 jj : : : jjSin DIn jj 0
(10.5)
Proof. The HOSVD of any N-dimensional tensor with real values was introduced
by de Lathauwer et al. (2000). The fact that discretized multivariate functions can
be stored in such tensors, as demonstrated in Eq. (8.2), proves the theorem.
F D.G/ D S Xn
(10.6)
nD1
where refers to the tensor product operation defined in de Lathauwer et al. (2000),
and:
.n/
.n/
.NC1/
147
all-orthogonality: any pair of the subtensors of S are orthogonal, i.e. for all
possible values of n; and subject to :
< Sin D ; Sin D >D 0
(10.7)
ordering: All of the subtensors of S along any given dimension n are ordered
according to their Frobenius norm, i.e. 8n D 1::N C 1:
jjSin D1 jj jjSin D2 jj : : : jjSin DIn jj 0
(10.8)
The values I1 ; : : : INC1 are the ranks of the system along each of the input
dimensions. Tensor S is referred to as the core tensor, and matrices Xi are referred
to as weighting matrices. If the values in just the pk th row of any Xk in Eq. (10.6) are
modified, then only those output values of function F D.G/ will be changed which
belong to the pk th perceptual gradation along the kth dimension of hyper-rectangular
grid G. This can be easily seen if we express a single element of F D.G/ as follows:
NC1
(10.9)
where the pk th row of matrix Xk is denoted by xk;pk (tensor multiplication with this
term can be written separately due to the commutativity of tensor multiplication).
It is obvious that if any point on hyper-rectangular grid G is chosen in which the
value of the kth dimension is not the pk th discretization point, then the output value
of the function will remain unchanged. For this reason, the manipulation of vector
xk;pk is referred to as the local tuning of the CogInfoCom channel along the kth input
dimension. The values in vector xk;pk in turn are referred to as tuning weights.
148
space does not necessarily have much to do with perceptual relationships between
the CogInfoCom messages that are generated, a mere rank reduction of the tuning
model will not necessarily result in perceptually justified tuning controls.
Thus, it is important to relax the tradeoff between interpretability and complexity.
While the user can only concentrate on the perceptual effects of changes made to
tuning weights if the number of weights is small, it is equally important that the user
be able to explore the parameter space of RH to the fullest extent possible. To this
end, a novel approach is proposed that combines rank reduction and the adaptive
modification of the basis vectors in the rank-reduced space so that the full-rank
space can be explored in a restricted, but systematic way. Such an approach can be
achieved through the following steps:
1. Assuming that the user would like to tune the system in the kth dimension, the
first step is to reduce its rank from Ik to 1. This can be achieved using HigherOrder Orthogonal Iteration (HOOI)a method which is proven to yield optimal
rank-reduction (Ishteva et al. 2008)to obtain:
argmin .F D.G/ FO D.G/ /
NC1
O n gnD1
SO;fX
(10.10)
where
NC1
On
FO D.G/ D SO X
nD1
(10.11)
such that IOk D 1; IOn D In ; 8n k, and IOk is the rank along the kth dimension of
the new system.
2. Having obtained this approximation, the goal is to increase the rank of the
vector space which can be controlled by the single tuning parameter remaining
O n; n D
in the kth dimension. In order to achieve this, weighting matrices X
O
1::.N C 1/, as well as the core tensor, S are expanded such that the obtained
Q n and core tensor SQ can be used to reconstruct the tensor
weighting matrices X
representation of the original parameter-generating function:
NC1
Qn
F D.G/ D SQ X
(10.12)
nD1
149
nD1
nD1
QC
SQ D F D.G/ X
n
nD1
(10.13)
Q NC1 jwpk 2 Rg
HL D fSQ xQ n;pn k wpk NC1 X
nD1
nk
(10.14)
nD1
nk
N
(10.15)
150
Q NC1 /C
FpD.G/
Dwpk ..BQ xQ n;pn / NC1 X
1 ;:::;pN
nD1
nk
(10.16)
Q NC1 / k rpk
..BQ c xQ n;pn / NC1 X
nD1
nk
(10.17)
where
1
0
N
(10.18)
.NC1/
151
Fig. 10.2 Proposed augmentation method used to compensate for reduced interpretability after the
tuning model is rank-reduced, assuming that k D 1 and N D 2. The light (yellow) shades contain
fix values, and the dark (mauve and green) shades contain variables which can be chosen so as to
compensate for the rank-reduction as well as to alter the slope of the hyperline that is traversed
within RH during tuning
152
and due to the fact that this weighting matrix is augmented by an identity matrix,
only a single and unique subtensor of K will affect the slope of the hyperline.
The key idea behind the implementation is that while the user changes the
weights in the first column of the weight matrix which belongs to dimension k (in
Fig. 10.2, this is denoted by XO1 ), the values of K are periodically and cyclically
modified between two extremes.
It should be noted that while the generic tuning model described in Sect. 10.3
guarantees that the effects of changing a tuning weight will be local, this property
of locality does not transfer to SDM. The reason for this is that as subtensor K
of the core tensor is modified, all gradations of the dimension which is tuned will
be affected. Nevertheless, in practical cases the effects of tuning operations can be
made to be local using a simple implementation workaround: if the user finishes
manipulating a given gradation, and wishes to save the changes made, the local
parameters can be updated in the original tensor representation of F D.G/ , and the
first two steps of SDM (i.e., rank reduction and re-augmentation) can be performed
once again. These operations are not cost-inhibitive, and in any case only need to
be effected if the user wishes to save a set of manipulations with respect to a given
gradation.
The velocity parameter of SDM is set transparently to the user, as it influences the sensitivity
of the distance parameter, but does not influence the direction of parameter discovery.
153
Fig. 10.3 The GUI used for tuning an audio-based CogInfoCom channel. Controls on the righthand side can be used to directly control generation parameters. Controls on the left-hand side
of the top window represent parameters of the generic tuning model, while the single control
on the left-hand side of the bottom window represents the distance tuning parameter of SDM
(cf. Fig. 10.1). As the parameter of SDM is moved by the user, corresponding changes in the
original generation parameters are updated automatically
was created using 24 parameters, hence, the output dimension of the parametergenerating function contained 24 dimensions. The key point was that it is difficult
for any user to understand both the individual effects of 24 tuning parameters, as
well as the effects that they may have on each other. To further complicate the task,
the goal of the tests was to create a CogInfoCom channel that consists of 20 different
CogInfoCom messages.
The test was performed by ten subjects. Each subject was given as much time
as he/she needed to use SDM in order to find 20 different CogInfoCom messages
that could be ordered along any arbitrary perceptual quality, chosen by the user
according to his or her own liking. Following the tuning process, test subjects were
allowed a few minutes to practice, after which they were presented with 30 test
questions to measure recall precision.
154
Weights
0.8
0.6
0.4
0.2
rm
2
ha
rm
ha
ur
rm
ha
ho
ns
gd
pc
de
tc
h
pi
ur
gd
ur
3
gd
ho
pc
ho
h1
h3
pc
pc
pc
ns
de
ns
in
ou
de
fa
d
fa
d
vo
fadin
1
0.5
0
fadout
1
0.5
0
pch1
1
0.5
0
pch3
1
0.5
0
gdur1
1
0.5
0
gdur3
1
0.5
0
dens
1
0.5
0
gdur
1
0.5
0
harm1
1
0.5
0
harm2
1
0.5
0
harm3
Fig. 10.4 As users alter the value of the distance parameter in SDM, all of the generation
parameters change together in accordance with the tuning model. The figure shows various
configurations in the generation vector space which were discovered by a user of SDM. The
horizontal axis refers to different dimensions in the generation vector, and the vertical axis refers
to the normalized value of the corresponding dimension. Line segments with the same color refer
to parameter configurations which were active at the same time
1
0.5
0
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
20
40
60
80
100
time
120
140
160
180
200
Fig. 10.5 As users alter the value of the distance parameter in SDM, all of the generation
parameters change together in accordance with the tuning model. The figure shows various
configurations in the generation vector space which were discovered by a user of SDM, in terms
of the values of a selection of individual dimensions as they change through time
155
14
12
10
8
6
4
2
0
1
Arbitrary dimension
Fig. 10.6 The boxplot shows error sizescounted in number of gradationsalong a single
arbitrary dimension, aggregated over ten test subjects. The red, horizontal line marks the median
error, the edges of the blue box mark the 25th and 75th percentile of error, the whisker marks errors
that lie outside of these percentiles, while the red crosses mark outliers
Figure 10.6 shows results from the tests. The figure shows that 75 % of all test
questions resulted in an error of at most two gradations. These can be regarded as
good results given that there were 20 gradations in a high-dimensional parameter
space. When it was suggested to test subjects that they could also tune the
CogInfoCom channels using the 24 original parameters (which were displayed
simultaneously with the SDM parameters), they agreed that the task would be so
difficult that it would not be even worth trying. Comparing this with the test results,
it can be seen that SDM is a viable and potentially useful approach to user-based
tunability (Fig. 10.7).
156
Response gradation
13
4.3%
3
1.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
76.5%
0.0% 23.5%
0
0.0%
8
2.7%
3
1.0%
2
0.7%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
57.1%
0.0% 42.9%
0
0.0%
4
1.3%
10
3.3%
9
3.0%
4
1.3%
2
0.7%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
33.3%
0.0% 66.7%
0
0.0%
1
0.3%
0
0.0%
6
2.0%
2
0.7%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
54.5%
0.0% 45.5%
0
0.0%
0
0.0%
1
0.3%
1
0.3%
7
2.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
77.8%
0.0% 22.2%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
3
1.0%
8
2.7%
2
0.7%
1
0.3%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
2
0.7%
0
0.0%
1
0.3%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
40.0%
0.0% 60.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
7
2.3%
1
0.3%
0
0.0%
1
0.3%
1
0.3%
0
0.0%
0
0.0%
1
0.3%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
53.8%
0.0% 46.2%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
8
2.7%
1
0.3%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
72.7%
0.0% 27.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
7
2.3%
2
0.7%
0
0.0%
3
1.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
50.0%
0.0% 50.0%
10
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
2
0.7%
7
2.3%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
63.6%
0.0% 36.4%
11
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
4
1.3%
3
1.0%
1
0.3%
3
1.0%
2
0.7%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
28.6%
0.0% 71.4%
12
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
1
0.3%
0
0.0%
2
0.7%
0
0.0%
3
1.0%
3
1.0%
2
0.7%
0
0.0%
0
0.0%
1
0.3%
2
0.7%
0
0.0%
0
20.0%
0.0% 80.0%
13
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
1
0.3%
7
2.3%
6
2.0%
1
0.3%
0
0.0%
3
1.0%
1
0.3%
0
0.0%
0
35.0%
0.0% 65.0%
14
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
2
0.7%
4
1.3%
6
2.0%
3
1.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
37.5%
0.0% 62.5%
15
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
1
0.3%
2
0.7%
1
0.3%
9
3.0%
2
0.7%
1
0.3%
1
0.3%
0
0.0%
0
47.4%
0.0% 52.6%
16
0
0.0%
2
0.7%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
2
0.7%
0
0.0%
0
0.0%
0
0.0%
4
1.3%
0
0.0%
0
0.0%
2
0.7%
0
36.4%
0.0% 63.6%
17
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
4
1.3%
10
3.3%
2
0.7%
2
0.7%
3
43.5%
1.0% 56.5%
18
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
3
1.0%
5
1.7%
1
0.3%
4
33.3%
1.3% 66.7%
19
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
1
0.3%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
5
1.7%
5
1.7%
1
38.5%
0.3% 61.5%
20
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
0
0.0%
1
0.3%
1
0.3%
2
50.0%
0.7% 50.0%
100% 44.4% 62.5% 31.6% 38.9% 72.7% 46.7% 66.7% 58.3% 46.7% 50.0% 16.7% 36.8% 30.0% 50.0% 36.4% 52.6% 29.4% 45.5% 20.0% 45.3%
0.0% 55.6% 37.5% 68.4% 61.1% 27.3% 53.3% 33.3% 41.7% 53.3% 50.0% 83.3% 63.2% 70.0% 50.0% 63.6% 47.4% 70.6% 54.5% 80.0% 54.7%
9 10 11 12 13 14 15 16 17 18 19 20
Target gradation
Fig. 10.7 Confusion matrices for the validation of SDM with arbitrary dimensions (10 test
subjects, 30 test questions)
Chapter 11
Earlier chapters have described an approach to defining icons and messages for the
design of structured multi-sensory signals capable of providing information on highlevel concepts. Various mapping techniques between generating parameters and
meaning have been presented, and the tunability of the resulting channels through
complexity-preserving, but still interpretable operations has also been discussed.
This chapter focuses on dynamic aspects of communicationi.e., aspects which
pertain to the development of communication through time. As this subject area has
more relevance to the long-term co-existence of humans and ICT than to episodic
interactions, its details are relatively less well understood. For this reason, the
chapter cannot aim to provide definitive answers; instead, its goal is to provide
an introduction to the topic by at least highlighting some of the more important
challenges and by presenting initial ideas that have been proposed for further
elaboration.
The chapter is structured as follows. Section 11.1 provides a brief introduction to
the dynamic aspects of communication considered in the chapter, and highlights two
key questions that motivate this research. This is followed by a discussion on recent
ideas obtained from low-level (i.e. non-human) communication (in Sect. 11.2)
and high-level human-human communication (in Sect. 11.3). A general outline on
how these concepts could be applied towards the design of adaptive CogInfoCom
channels is provided in Sect. 11.4.
157
158
Emotions and passions tend to be more problematic than helpful in human-human interaction.
There is no need to contaminate purely logical computers with emotional reactiveness (Picard
2003a).
159
clear from the Gartner hype cycles of the past years, such as those shown in
Fig. 1.3), such aspects of communication are almost entirely missing from the
considerations applied in mainstream engineering practices.
Based on these considerations, it is becoming increasingly important for the
engineering sciences to take the spontaneity, variability and progression of natural
communication into account when designing systems for human-ICT co-evolution.
While many authors have in the past advocated the incorporation of individual
aspects of natural communication into engineering design, these three factors have
been rarely addressed together. In our view, answering the following two questions
will be key to making further progress:
1. How does communication emerge in the first place: what phases does a series of
interactions have to go through for it to be considered as communication?
2. Once communication is established, how is it maintained, and how does it
evolve in order to remain relevant, even as the cognitive beings involved
form increasingly detailed representations of each others habits, attitudes and
motives?
Although a definitive answer to these questions currently seems elusive even in
the biological and social sciences, we believe that being aware of their existence
is already half the solution. A greater awareness of how potential answers to these
questions could contribute to engineering design may eventually lead to important
results. In the remainder of this chapter, some ideas from low-level biological
communication and high-level human communication to support future research
are highlighted.
160
the concept of CogInfoCom channels in this interpretation still retains its semantically oriented, message-like nature, it is also seen as a form of communication that
evolves from repetitive variations in lower-level signals (in much the same way as
signals are in turn built on the foundation of cues). Before providing a tentative
definition of the concepts of cues and signals, and describing their relationship
with icons and messages, a brief summary of considerations supporting such a
hierarchical approach is provided as follows:
It may be useful to model the biological transition between cues and signals in
ICT systems, as this would allow for a transitioning from interaction modes with
no implicit communication (i.e. characterized only by goal-oriented exchanges)
to interaction modes with meaningful implicit communication (characterized by
contextually supported implicit interpretations besides goal-oriented exchanges).
CogInfoCom channels can be explicitly designed and taught to users (or programmed into artificial systems), but it is unlikely that they can emerge spontaneously and still be effective. If an alternative to explicit design and training
is considered, it is likely that such channels will be formed through extended
periods of prior interaction that leads to both cognitive entities picking up on
the kinds of subtleties that can form the basis of differentiation among alternative
interpretations.
Cues, signals and channels are different and complement each other well in
the sense that their perception and interpretation is founded on different levels
of goal-directedness and also requires different levels of attention. Cues are
unique in that they emerge primarily through goalless behaviors not targeted
at communication; signals are unique in that they emerge primarily through
goal-directed behaviors without explicit shared meaning; and finally, channels
are unique in that they are produced in goal-oriented ways based on previously
established conventions of shared meaning.
161
SIGNALS
CUES
Fig. 11.1 Structural and interactive aspects of communication. Cues and icons have direct
meaning, whereas messages have abstract meaning. In some cases, cues are also icons at the same
time, while in others, cuesunlike iconsdo not give rise to immediate and unique associations
162
163
Fig. 11.2 The emergence of messages and channels from cues in CogInfoCom. Communicational
exchanges leading to specific interpretations are represented by arrows, while increasing resolutions in the dashes represent increasing resolutions in interpretation
would require the designer to also implement some way for the system to recognize
what external events (e.g., human behaviors) are useful to it and which of its
cues were successful in eliciting those events. Further, it would require mapping
those cues to appropriate modalities so as to support the emergence of semantic
associations in the most effective way possible. Clearly, just the definitions of these
stages of communication will not help engineers in making good design choices.2
Instead, finding a set of characteristics which transcend the notions of cues, signals
and channels, but which can nevertheless be brought into connection with these
categories may lead to a viable solution. More specifically, if there exist a set of
communicational features that can be associated with cues, signals and channels,
but are also distinct in each of these phases, then they can be taken as a basis for
CogInfoCom design.
In this section, three such communicational features are outlined in terms of
volition, directness and temporal reference; their possible relevance to cues and
signals is outlined in Sect. 11.4. Much inspiration for these features was obtained
from Nass and Yen (2010), in which strong arguments are presented for users
aptitude towards implicitly developing mental and emotional relationships with
machines that are analogous to relationships in human-human communication
(Csapo 2014; Csapo and Baranyi 2013). An important abstraction underlying these
discussions is the notion of triggers; although the term has been used in more than
For example, an application that requires constant feedback from users as to whether or not an
interaction was useful to them would quickly grow tedious; and finding appropriate modalities
through design would lead back to the original problem of designing CogInofCom channels.
164
11.3.1 Volition
The dimension of volition concerns the level of consciousnessor, in terms more
suited to artificial cognition: the level of explicitnessrequired for the activation
of a trigger. Thus, it can be made clear that not all forms of communication are
founded on an explicit desire, or functional requirement to communicate. Based on
this notion, voluntary and involuntary triggers are defined as follows:
Definition 11.4. A trigger is voluntary if explicit/conscious actions are required
for its activation.
Definition 11.5. A trigger is involuntary if implicit/unconscious actions are sufficient for its activation.
Responses to voluntary and involuntary triggers can be characterized as episodic,
periodic or aperiodic as described below. Further, depending on whether a trigger
is voluntary or involuntary, the temporal properties of associated response(s) can
be further qualified based on a unique set of criteria. Thus, the following response
types are defined:
Definition 11.6. The response given to a trigger is episodic if it consists of a single
response provided at a unique point in time, and aperiodic/periodic if it consists
of multiple communicational exchanges initiated by the receiving endpoints of the
trigger at varying (in the case of aperiodic responses) or regular (in the case of
periodic responses) periods of time.
165
11.3.2 Directness
Another important aspect of trigger-based communication is its level of directness.
Earlier in Sect. 7.1, directness was specified as a factor that refers to whether
the cognitive entity knows that it is communicating, and whether the act of
communication requires active involvement in terms of the dedication of cognitive
resources. Untangling from this notion those aspects that are common with volition,
an important feature which emerges is the element of unexpectedness and/or
discomfort associated with a trigger. If the trigger comes as a surprise to (or is
unexpected by) the receiving endpoints; or if the responses of those endpoints are
surprising (or unexpected) to the triggering endpoints, it can be said that the source
of information based on which the endpoints are communicating is unbalanced.
Thus, directness and indirectness can be defined as follows:
Definition 11.9. Knowledge with respect to a given information is shared
between a set of cognitive entities if all entities know with respect to each of the
other entities whether or not they are in possession of the information.
Definition 11.10. A trigger is indirect if knowledge of its occurrence is shared
between the triggering and receiving endpoints, and direct otherwise. Similarly, a
response is indirect if knowledge of its content is shared between the triggering and
receiving endpoints, and direct otherwise.
166
11.3.3 Discussion
The rationale behind these definitions can be supported by the following examples:
In human-human communication, if a stranger asks someone an unexpected
favor, the communicative exchange that ensues is caused by a voluntary and
direct trigger: the favor is asked for purposefully, and the knowledge of its
occurrence is not shared by the interlocutors. In fact, the request may come as
quite a surprise to the receiver of the exchange, prompting a direct need to address
it as well as a direct response (acceptance or rejection) that is perhaps relatively
less tactful than usual.
Conversely, if a favor is asked in friendly settings, the request is perceived as
less direct. If the receiver of the request is accustomed to receiving this request
from the same person, the trigger is indirect. These two examples highlight the
fact that the transition from direct to indirect communication is not necessarily
binary: various levels of indirectness are conceivable based on the degree to
which relevant knowledge is shared.
In human-device communication, the pressing of a button generally corresponds
to a voluntary and indirect trigger: the button is pressed purposefully, and the
user also knows that the device is capable of responding, while the button on the
device was explicitly designed to trigger the requested operations. If the same
button is pressed accidentally, the trigger is involuntary; and if the ability of the
device to respond comes as a surprise to the user, the response will be direct.
If the device performs the operation even before the user presses the button, the
response is proactiveand reactive otherwise.
If a given context (such as the users location, surrounding noise levels and other
sensor measurements) causes an ICT system to remind the user of a past event,
the trigger of the communication is involuntary and the associated response is
post hoc. If, on the contrary, the involuntary trigger causes the system to alert the
users attention to a potential future event, the response is pre hoc. Finally, if the
response refers to the current moment or otherwise has no temporal implication,
then the response is ad hoc.
The kinds of trigger and response types which have been described in this section
are summarized in Fig. 11.3.
167
Fig. 11.3 The figure provides a color coded reference on trigger and response types. Mutually
exclusive conceptssuch as voluntary and involuntary triggers, or episodic, aperiodic and periodic
responses are shown in the same color. Concepts related to level of directness are both red and blue
because both triggers and responses can be direct or indirect
its reference to past, presence and future events, and how much of the underlying
context is shared by the communicational endpoints involved in it. Two key points
are suggested by these observations:
In a way that is analogous to the emergence of biological communication, it may
be useful for long-term, entangled interactions if the patterns of communication
among cognitive entities could evolve through time;
In order for users to perceive communication as natural, it should apply different
trigger and response types at different times, as dictated by the context.
It is important to realize that there are also strong parallels between these two
points: interactions in terms of cues, messages and channels are characterized by
different kinds of triggers. For example, a link between biological concepts of
communication and volition/directness of associated triggers is shown in Fig. 11.4.
Although the figure shows tentative conclusions that should be the subject of further
investigations, it nevertheless gives a strong indicationwhich is also intuitively
clearthat:
Cueswhich are a natural by-product of autonomous behaviorscannot be
voluntarily triggered, and cannot be characterized by a priori knowledge (cues
are not communicational elements in the formal sense: therefore, all knowledge
on their occurrencewhich is in fact no knowledgeis trivially shared);
Messageswhich are emitted because they are perceived as useful based on past
interactionsare generally involuntary from the perspective of the entity whose
behavior elicits them, and at the same time rest on the shared knowledge of their
effects (otherwise there would be no purpose for their existence);
168
Fig. 11.4 The emergence of messages and channels from cues in CogInfoCom in the special case
of human-ICT communication. The figure suggests that the ritualization and differentiation process
progresses as the user and system gradually learn to understand and make increasing use of each
others communicative capabilities (although the figure depicts the users perspective, the ability
of the system to learn is also implicitly understood). In this process, what begins as a kind of trialand-error set of exchanges with increasing directness evolves into a process where increasingly
explicit control is achieved through increasingly indirect interactions
Channels are used voluntarily (i.e. for evolved functional purposes) and are also
generally indirect (given that their evolution requires shared knowledge).
Besides highlighting the above points, Fig. 11.4 also suggests a way to solve
the problem of how to achieve cue ritualization and signal differentiation in
CogInfoCom settings without requiring constant, tedious feedback from users.
Specifically, it is suggested that triggers asking for feedback should initially be
indirect, gaining directness only once the functional need for interaction becomes
apparent; whereas later the same triggers should be increasingly less direct, as signal
differentiation is expected to occur autonomously through practice. Directness, in
this case, is seen as a strongly multi-modal concept that can manifest itself in a
large variety of phenomena including word usage in spoken commands as well as
meta-communicational cues and signals.
In any case, the natural progression described here seems to be attainable if both
the humans and ICT devices involved in communication are capable of improving
the effectiveness of their communicational behavior through different forms of
learning. When considering the ways in which a channel might learn representations, it becomes clear that several forms of learning are possible depending on who
169
(or what) is the teacher, and who (or what) adapts during the process. Thus, it is
possible to distinguish among the following forms of learning in the communication
between two entities:
regressive learning: one of the two entities provides learning samples, and one
of the two entities adapts based on those samples
progressive learning: both entities provide learning samples, and both entities
adapt
Regressive learning encapsulates the experience of updating ones representations of an environment based on a set of samples (i.e., experiences) that
are either entirely self-induced, or are passively experienced under comparable
circumstances. This terminology is in keeping with the term regression as it is used
in machine learning, i.e., as a kind of categorization that occurs over a continuum
of potential categories. Progressive learningas opposed to regressive learningis
characterized by bilateral interactions in which both parties participate actively in
the generation of samples during the learning process. In the context of CogInfoCom
channels, this can be beneficial in the sense that the users decisions and actions are
supported by a broader set of feedback behaviors.
The development of suitable algorithms for the progressive evolution of CogInfoCom channels is an important challenge for the future. Based on the discussions
in this chapter, it is suggested that this will require representing and monitoring
the evolution of communicational exchanges, and also the systematic actuation
of communicational behaviors in a way that is suitable to the frameworks of cue
ritualization/signal differentiation and CogInfoCom trigger/response types.
Part IV
Chapter 12
This chapter collects the challenges of the todays Internet, and outlines the
visionary capabilities of the next generation Internet, including cognitive capabilities, virtual collaborations, 3D Internet, Internet of Things and smart applications.
Objects, devices provided by sensors and connected to the Internet become smart,
cognitive things, which can be virtually represented and collaborate with other
virtual or real cognitive things and human beings in a composed virtual 3D space.
The challenge is that billions of cognitive things are to be interconnected (Fig. 12.1),
the concept of Internet of Things was deployed (Smith 2012; Vermesan and Friess
2013). The next generation Internet, called Future Internet (FI) is intensively
investigated (Tselentis et al. 2010; Domingue et al. 2011; Alvarez et al. 2012;
Galis and Gavras 2013); the deployment of the future networked knowledge society
is prospectively based on FI technologies and applications, including advanced
community applications. This chapter also defines the research goals and sorts the
relevant research themes into clusters, pointing out the position of CogInfoCom,
in general the handling of cognitive entities, i.e. cognitive things and beings
(Sallai 2013c).
173
174
Limitations of todays Internet are (Tselentis et al. 2010; Domingue et al. 2011;
Alvarez et al. 2012; Galis and Gavras 2013; Nguyen et al. 2011; Zahariadis et al.
2011; Papadimitriou et al. 2012; Ziegler et al. 2013):
the limited identification capacity, the lack of IPv4 address space;
the essentially private wireline network concept, the lack of an inherent mobile
centric network architecture and a scalable efficient network and mobility
management;
the best effort solutions, the lack of guaranteed and differentiable quality of
services and security;
the growth rate in energy consumption and carbon dioxide emissions are not
sustainable;
the application development is inefficient and inflexible for handling large
diversity of use cases and requirements, etc.
Growth of societal role of the Internet gives rise demands as (Domingue et al.
2011; Alvarez et al. 2012; Galis and Gavras 2013; European Commission 2013b;
Karnouskos and Skarmeta 2013; WEF 2007; CogInfoCom 2012):
anywhere, anytime access (always on);
interconnection of devices, objects, sensors (networked 20100 billion things);
175
176
177
178
The FI-based applications are in the centre of the vision for a smarter world
involving applications as smart cities, smart homes and offices, smart health,
education and entertainment, smart energy, smart transportation and logistics, smart
manufacturing and agriculture, smart governance, etc. The FI applications are
software solutions for a specific task, hence they are expected to be seamlessly
adjusting to the users needs and context, while in parallel hiding the complexity of
the FI infrastructure and interactions with other applications (Alvarez et al. 2012).
In the frame of the FI PPP (public-private partnership) programme FI-ware cloudbased generic application enabler platform has been created, which is completed by
specific modules for use cases (e.g. FI-Content for social connected TV, smart city
services and augmented reality based games, FI-Space for business collaboration
networks in agri-food, transport and logistics, FI-Star for healthcare, FITMAN for
manufacturing industries, FINESCE for smart energy) (ECFI 2014).
179
A list of relevant Future Internet features is (Domingue et al. 2011; Alvarez et al.
2012; Galis and Gavras 2013; Matsubara et al. 2013; Papadimitriou et al. 2012;
Wainwright and Papanikolaou 2012):
1.
2.
3.
4.
Recent capabilities are in good coincidence with the visionary ones, what shows
the reality of the vision and the awareness of the research governance. The Future
Internet application research is focusing on smart cities (Alvarez et al. 2012), green
energy (Nguyen et al. 2011) and advanced humanoid applications (CogInfoCom
2012, 2013, 2014) especially on smart health and ambient assisted living (Minutolo
et al. 2012), safe vehicular communications (Papadimitratos et al. 2009) and
driver assistance systems (Betaille and Toledo-Moreo 2010), as well as interactive
cognitive mobile applications (Csapo and Baranyi 2010). CogInfoCom research
works and some applications, as crowdsourcing and crowdsensing community
applications, augmented reality, 3D virtual collaboration and massive open online
courses (MOOCs) are closely related to Future Internet function 6 and support the
holistic concept of Internet of all cognitive entities.
180
Studying the research themes on the Future Internet in the literature, in particularly on FIAs (Tselentis et al. 2010; Domingue et al. 2011; Alvarez et al. 2012;
Galis and Gavras 2013; Wainwright and Papanikolaou 2012), Working Programme
of Horizon2020 (European Commission 2013b) as well as the Hungarian Future
Internet research activity (Sallai 2013b; CogInfoCom 2013; Bakonyi and Sallai
2014a,b), the relevant research topics have been identified and clusterized (Sallai
2013a, 2014). The main research clusters of Future Internet research activity
has been arranged in a layered model from basic research on Internet Science
through Internet Engineering up to Internet Practice. Figure 12.3 shows the nine
clusters and their main research goals and spheres. The Internet Science, the basic
research is Cluster 1, the Internet Engineering applied research is divided into five
clusters (Cluster 2 : : : 6), the Internet Practice, the applications and experiments
are comprised in three clusters (Cluster 7, 8 and 9). Figure 12.3 also indicates the
possible relations of the clusters to other research goals and spheres.
In the following, the research clusters and their relevant topics, the elements
of clusters are presented. Each cluster has 57 elements, and 54 relevant topics
are defined in total. Relevant research topics on cognitive sciences, cognitive
infocommunications or related applications are indicated by asterisks.
181
182
183
184
185
issues. The cluster elements are defined as (Domingue et al. 2011; Alvarez et al.
2012; Galis and Gavras 2013; ITU 2011a,b, 2012a,b; Sales et al. 2012; Matsubara
et al. 2013; Liu 2013; Vermesan and Friess 2013):
A.
B.
C.
D.
Chapter 13
In the previous chapter, the concept of Future Internet was formalized through the
perspective of the next 510 years. This formalization included a vision of a future
network that aims at service and resource awareness. In this chapter, a longer term
and broader vision is outlined, which covers not only technological aspects, but
also aspects relevant to the impact of Future Internet on the physical and societal
environment. Through the convergence of TIM technologies with the cognitive
sciences and a novel concept of Global Networked Society, it is argued, customers
and their devices together will be increasingly viewed as cognitive entities capable
of bringing about unprecedented functional interactions.
The chapter first describes the ways in which the original Digital Ecosystem
conceptformulated in 2007is being incrementally reshaped through emergent
technological-social-cognitive phenomena. This is followed by a brief overview of
the newly emerging Cognitive Internet Ecosystem. It is important to clarify that
the goal of this chapter is not to provide a detailed account of definite facts, but
rather to describe the authors views on a possible long-term evolution of networked
CogInfoCom systems. Due to the rapid development of technologies and synergies,
the discussion here can by no means serve as a definite characterization of the future,
and should therefore be approached with critical caution.
187
188
(WEF 2007). At the same event, possible scenarios were formulated for the
evolution of DES into the year 2015 based on answers to the following key
questions:
Will social and economic value creation be industry controlled and led, or organic
and community-led?
Will the digital business environment evolve toward a more open or closed
system?
Guided by these questions, the following three scenarios emerged for the Digital
Ecosystem:
Safe Havens scenario (industry led value creation and closed business environment) describes a digital world in which online security issues create a need
for virtual safe havens. The industry would respond through vertical integration
by creating secure walled environments that provide all digital services, but
nevertheless lend a sense of constriction to user interaction.
Middle Kingdoms scenario (industry led value creation and open business environment) describes a digital world dominated by intermediaries that effectively
connect users to like-minded individuals and to the highly specialized suppliers
that can best meet their needs. At the center of the space between consumers and
suppliers would lie the kingdoms where the power would be concentrated.
Youniverse scenario (community-led value creation and open business environment) describes a digital world characterized by the rise of new organizational
structures, the change of economic value creation and traditional business
thinking, and by highly personalized digital experiences. To survive uncertainty
and changes, companies would have to find ways to capitalize on this distributed
innovation landscape.
During the past few years many aspects of the DES have materialized. Thus,
the current time period is often referred to as the Digital Age, in which open
access to information and knowledge is fundamental. Communication platforms
are widely seen as resources which much remain open for all in order to ensure
equitable access to the Internet. Through this vision, a reassessment of the current
regulatory framework has also begun. Digital users are starting to take control of
when, where and how they can consume digital content; hence, they are no longer
merely consumers, but also participate in the DES in a variety of other ways, e.g.
by contributing to online communities and taking part in the creation/distribution
of digital content and services. Communities are also being created around the
development of new infrastructure, as when members of a community agree to share
their wireless internet access (Yoonaidharma et al. 2014).
At the same time, the current manifestation of digital convergence rests primarily
on Internet technology; the Internet is considered as a global platform for accelerating the flow of information that is pushing many earlier forms of information
and knowledge sharing into obsolescence. Most traditional communications media,
including telephone, music, film, and television are being reshaped or redefined
so as to suit the Internet, giving birth to new services such as Voice over Internet
189
Protocol and Internet Protocol TeleVision. The Internet has become a global system,
an international network of networks that consists of millions of public and private
packet switched networks, linked by a wide range of networking technologies.
Therefore, the Digital Ecosystem concept is increasingly referred to as an Internet
Ecosystem (IES). Accordingly, the term Digital Age is being replaced by Internet
Age, so as to adequately reflect changes in the essence of the technology used in
the infrastructure that enables digital interactions; as well as to express the global
character and the principle of collectiveness realized in todays Digital Ecosystem.
Although both the Digital Ecosystem and the Internet Ecosystem are founded on
the infocommunication or ICT infrastructure (see also Chap. 3), a new wave of both
opportunities and threats has emerged, which is described by Gartner Inc. (2014):
As a consequence, rapid technological breakthroughs have been achieved, leading to radical changes in lifestyle and business landscape (the latter also through an
underestimation of the digital environment). Companies hoping to achieve success
in the DES/IES are required to find digital synergies, and more generally to adopt
appropriate digital business strategies based on a suitable analysis of the digital
environment, leading to the identification of digital opportunities and threats as
well as a better understanding of digital organizational capabilities, including digital
strengths and weaknesses.
At the same time, the advanced digital TIM technology has penetrated into
customer devices, and has integrated customers through a large variety of interaction modes, leading to an extension of relevant content space with sensory and
cognitive content. Applications can increasingly make use of and modify the context
which surrounds customerse.g. via crowdsourcing and crowdsensing community
applications. Inasmuch as this is the case, the DES/IES is progressively widening
so as to comprise cognitive infocommunication networks, the Internet of Things,
Big Data tools and solutions, 3D Internet and community applications, as well
as all of the devices, tools and opportunities that are in general provided by
the Future Internet. Expressed more primarily, the Digital/Internet Ecosystem is
radically being enlarged and changed in character through the sudden involvement
of billions of smart devices. The term Internet of Things Ecosystem is used to
express this important motivating feature (IDC 2014). Further, the holistic term
Internet of Everything Ecosystem becomes viable when the Future Internet vision
is interpreted for bringing together people and things (Kurian and Anand 2014).
190
191
Although it is mentioned in that paper any single one of these (and other)
criteria can be (and are) used to delineate sensory modalities, many of them can be
equally applied to the kinds of informational sensory modalities which are appearing
today. Through the analysis of real-time feeds of data and information, it will
someday be possible to obtain a virtual sensation of the general psychological
atmosphere in a city, the level of collective awareness on a specific issue,
or the overall conditions of health in a region. As capabilities for obtaining
such information are increasingly used and increasingly grow to become a part of
the standardized (canonical) capabilities of the CIES, they will increasingly be
viewed as (augmented virtual) sensory modalities that can be concentrated into a
single concept. Based on the services of such artificial modalities, new augmented
capabilities for the manipulation of previously inaccessible information types will
also become possible through integrated/augmented virtual realities. Such are the
perspectives which await members of generation CE.
References
193
194
References
Bakonyi P, Sallai G (2014a) Future Internet National Research Program - JINKA 2.4 (in
Hungarian), Budapest, p 60
Bakonyi P, Sallai G (eds) (2014b) Future Internet research in Hungary - FIRCC Report 2014,
Debrecen, p 128. ISBN: 978-963-473-718-6
Balata J, Franc J, Mikovec Z, Slavik P (2014) Collaborative navigation of visually impaired.
J Multimodal User Interfaces 8:175185
Baranyi P (2004) TP model transformation as a way to LMI based controller design. IEEE Trans
Ind Electron 51(2):387400
Baranyi P (2014) The generalized TP model transformation for TS fuzzy model manipulation and
generalized stability verification. IEEE Trans Fuzzy Syst 22(4):934948. ISSN: 1063-6706
Baranyi P, Csapo A (2010) Cognitive infocommunications: CogInfoCom. In: 2010 11th international symposium on computational intelligence and informatics (CINTI), Budapest,
pp 141146
Baranyi P, Csapo A (2012) Definition and synergies of cognitive infocommunications. Acta
Polytech Hung 9:6783
Baranyi P, Gilanyi A (2013) Mathability: emulating and enhancing human mathematical
capabilities. In: 2013 IEEE 4th international conference on cognitive infocommunications
(CogInfoCom). IEEE, pp 555558
Baranyi P, Yam Y, Varlaki P (2006) Tensor product model transformation in polytopic model-based
control. CRC/Taylor & Francis, Boca Raton/London
Baranyi P, Galambos P, Csapo A, Varlaki P (2012) Stabilization and synchronization of dynamicons through CogInfoCom channels. In: 2012 IEEE 3rd international conference on cognitive
infocommunications (CogInfoCom), pp 3336
Baranyi P, Csapo A, Varlaki P (2014) An overview of research trends in CogInfoCom. In: 2014
18th international conference on intelligent engineering systems (INES). IEEE, pp 181186
Barrass S (1998) Auditory information design. Ph.D. thesis, Australian National University
Benford S, Bowers J, Fahlen LE, Greenhalgh C, Snowdon D (1997) Embodiments, avatars, clones
and agents for multi-user, multi-sensory virtual worlds. Multimedia Systems 5(2):93104
Benus S (2013) Cognitive aspects of communicating information with conversational fillers
in slovak. In: 4th IEEE international conference on cognitive infocommunications. IEEE,
pp 271276
Benus S, Levitan R, Hirschberg J (2012) Entrainment in spontaneous speech: the case of
filled pauses in supreme court hearings. In: 3rd IEEE international conference on cognitive
infocommunications, pp 793797
Benus S, Gravano A, Levitan R, Levitan SI, Willson L, Hirschberg J (2014a)
Entrainment, dominance and alliance in supreme court hearings. Knowl-Based Syst.
doi:10.1016/j.knosys.2014.05.020
Benus S, Levitan R, Hirschberg J, Gravano A, Darjaa S (2014b) Entrainment in slovak collaborative dialogues. In: 5th IEEE international conference on cognitive infocommunications, Vietri
sul Mare, pp 309313
Berthelon F, Sander P (2013a) Emotion ontology for context awareness. In: 4th IEEE international
conference on cognitive infocommunications, Budapest, pp 5964
Berthelon F, Sander P (2013b) Regression algorithm for emotion detection. In: 4th IEEE
international conference on cognitive infocommunications, Budapest, pp 9196
Berthoz A, Pavard B, Young LR (1975) Perception of linear horizontal self-motion induced by
peripheral vision (linearvection) basic characteristics and visual-vestibular interactions. Exp
Brain Res 23(5):471489
Betaille D, Toledo-Moreo R (2010) Creating enhanced maps for lane-level vehicle navigation.
IEEE Trans Intell Transp Syst 11(4):786798
Bhatt GD (2001) Knowledge management in organizations: examining the interaction between
technologies, techniques, and people. J Knowl Manag 5(1):6875
Biocca F, Kim J, Choi Y (2001) Visual touch in virtual environments: an exploratory study of
presence, multimodal interfaces and cross-modal sensory illusions. Presence Teleoperators
Virtual Environ 10(3):247265
References
195
Biocca F, Inoue Y, Polinsky H, Lee A, Tang A (2002) Visual cues and virtual touch: role of visual
stimuli and intersensory integration in cross-modal haptic illusions and the sense of presence.
In: Gouveia F (ed) Proceedings of presence, Porto
Biro J, Heszberger Z, Martinecz M, Kis SM (2010) Novel equivalent capacity approximation
through asymptotic loss analysis. Comput Commun 33(1):S152S156
Blake R (2000) Computer mediated communication: a window on l2 Spanish interlanguage. Lang
Learn Technol 4(1):120136
Blattner M, Sumikawa D, Greenberg R (1989) Earcons and icons: their structure and common
design principles. Hum Comput Interact 4(1):1144
Blattner MM, Papp A, Glinert E (1994) Sonic enhancement of two-dimensional graphics displays.
In: Kramer G (ed) Auditory display: sonification, audification and auditory interfaces. Addison
Wesley, Reading, pp 447470
Blum JR, Eichhorn A, Smith S, Sterle-Contala M, Cooperstock JR (2014) Real-time emergency
response: improved management of real-time information during crisis situations. J Multimodal
User Interfaces 8:161173
Boiko B (2005) Content management bible. Wiley, Indianapolis
Bonin F, Campbell N, Vogel C (2014a) Time for laughter. Knowl-Based Syst 71:1524
Bonin F, Vogel C, Campbell K (2014b) Social sequence analysis: temporal sequences in interactional conversations. In: 5th IEEE international conference on cognitive infocommunications,
Vietri sul Mare, pp 403406
Borus G, Gilanyi A (2013) Solving systems of linear functional equations with computer. In: 4th
IEEE international conference on cognitive infocommunications (CogInfoCom), pp 559562
Brewster S, Brown L (2004) Tactons: structured tactile messages for non-visual information
display. In: Proceedings of the 5th conference on Australasian user interface (AUIC04), vol
28. Dunedin, pp 1523
Brunger-Koch M, Briest S, Vollrath M (2006) Virtual driving with different motion characteristics:
braking manoeuvre analysis and validation. In: Proceedings of the driving simulation conference, pp 6978
Campbell N (2012) Social aspects and speechability in CogInfoCom systems. In: 3rd IEEE
international conference on cognitive infocommunications, Kosice
Carazo P, Font E (2010) Putting information back into biological communication. J Evol Biol
23(4):661669
Card SK, Moran TP, Newell A (1983) The psychology of human-computer interaction. Lawrence
Erlbaum Associates, Hillsdale, NJ, USA
Cassell J, Bickmore T, Billinghurst M, Campbell L, Chang K, Vilhjalmsson H, Yan H (1999)
Embodiment in conversational interfaces: REA. In: Proceedings of the SIGCHI conference on
human factors in computing systems, pp 520527
Castro J (1995) Fuzzy logic controllers are universal approximators. IEEE Trans Syst Man Cybern
25:629635
Castrucci M, Priscoli FD, Pietrabissa A, Suraci V (2011) A cognitive future Internet architecture.
Springer, Berlin
Cerf VG (2009) The day the Internet age began. Nature 461(7268):12021203
Chaffee SH, Metzger MJ (2001) The end of mass communication? Mass Commun Soc 4(4):
365379
Chen M, Gonzalez S, Vasilakos A, Cao H, Leung V (2011) Body area networks: a survey. Mobile
Netw Appl 16(2):171193
Cheng DS, Salamin H, Salvagnini P, Cristani M, Vinciarelli A, Murino V (2014) Predicting
online lecture ratings based on gesturing and vocal behavior. J Multimodal User Interfaces
8(2):151160
Chernoff H (1973) The use of faces to represent points in k-dimensional space graphically. J Am
Stat Assoc 68(342):361368
ChuDuc H, NguyenPhan K, NguyenViet D (2013) A review of heart rate variability and its
applications. APCBEE Procedia 7:8085
CloudNet 2014 (2014) Third IEEE international conference on cloud networking
196
References
References
197
Domingue J et al (eds) (2011) The Future Internet - achievements and technological promises. In:
Future Internet assembly, Budapest. Springer, Heidelberg
Dropuljic B, Popovic S, Petrinovic D, Cosic K (2013) Estimation of emotional states enhanced by
a priori knowledge. In: 4th IEEE international conference on cognitive infocommunications,
Budapest, pp 481486
Drouin M, Kaiser DH, Miller DA (2012) Phantom vibrations among undergraduates: prevalence
and associated psychological characteristics. Comput Hum Behav 28(4):14901496
Duval T, Nguyen TTH, Fleury C, Chauffaut A, Dumont G, Gouranton V (2014) Improving awareness for 3D virtual collaboration by embedding the features of users physical environments and
by augmenting interaction tools with cognitive feedback tools. J Multimodal User Interfaces
8(2):187197
ECFI (2014) First European conference on future Internet, Brussels, 2-4 April 2014. Future Internet
PPP. www.ecfi.eu/brussels2014
Eichhorn W (1978) Functional equations in economics. Addison-Wesley, Reading
EITO (2014) European information technology observatory. ICT Market Report, Yearbooks from
2001 to 2014. ISSN: 097-4862
Ekman P (2003) Darwin, deception, and facial expression. Ann N Y Acad Sci 1000(1):205221
Ellis S (1991) Nature and origins of virtual environments: a bibliographical essay. Comput Syst
Eng 2(4):321347
Endsley M, Garland D (2000) Situation awareness: analysis and measurement. Routledge,
New York
Enriquez M, MacLean K (2003) The hapticon editor: a tool in support of haptic communication
research. In: Proceedings of the 11th symposium on haptic interfaces for virtual environment
and teleoperator systems (HAPTICS03). IEEE Computer Society, Los Angeles, pp 356362
Enriquez M, Maclean K, Chita C (2006) Haptic phonemes: basic building blocks of haptic
communication. In: Proceedings of the 8th international conference on multimodal interfaces
(ICMI 2006). ACM Press, Banff, pp 302309
Ericson A, Johansson C, Nergard H (2014) Manufacturing knowledge: going from production of
things to designing value in use. Intell Decis Technol 9(1):7989
ETSI (2011) About ETSI, 2011. European Telecommunications Standards Institute, SophiaAntipolis. http://www.etsi.org/website/aboutetsi/aboutetsi.aspx
European Commission (1997) Green paper on the convergence of the telecommunications, media
and information technology sectors, and implications for regulation. Towards an Information
Society Approach, COM, 3 December 1997
European Commission (2010) Future media networks - research challenges 2010. Future media
networks cluster of networked media systems FP7 projects. doi:10.2759/37178
European Commission (2013a) HORIZON 2020 - The Framework Programme for Research
and Innovation. Work Programme (2014-2020) 5.i. Leadership in enabling and industrial
technologies: Information and communication technologies. Annex 6 to Decision, December
2013, p 107
European Commission (2013b) International conference on Internet science. The FP7 European
Network of Excellence in Internet Science, Brussels, 9-11 April 2013. http://internet-science.
eu
Farago T, Konok V, Gacsi M (2011) Humans are able to recognize emotions by behavioural
expressions of an abstract artificial agent. In: 2nd international conference on cognitive
infocommunications
Fekete K, Csorba K (2014) Power extraction of human behavior in mobile environment. In: 5th
IEEE international conference on cognitive infocommunications, Vietri sul Mare, pp 481486
Fenn J, Raskino M (2008) Mastering the hype cycle: how to choose the right innovation at the right
time. Harvard Business Press Review, Boston
Fischer A, Beck MT, De Meer H (2013) An approach to energy-efficient virtual network
embeddings. In: 2013 IFIP/IEEE international symposium on integrated network management
(IM 2013), pp 11421147
198
References
Flanagan JR, Vetter P, Johansson RS, Wolpert DM (2003) Prediction precedes control in motor
learning. Curr Biol 13(2):146150
Flanagan JR, Bowman MC, Johansson RS (2006) Control strategies in object manipulation tasks.
Curr Opin Neurobiol 16(6):650659
Foldesi P, Botzheim J (2012) Computational method for corrective mechanism of cognitive
decision-making biases. In: 2012 IEEE 3rd international conference on cognitive infocommunications (CogInfoCom), pp 211215
Fortuna C, Mohorcic M (2009) Trends in the development of communication networks: cognitive
networks. Comput Netw 53(9):13541376
Fransman M (2002) Mapping the evolving telecoms industry: the uses and shortcomings of the
layer model. Telecommun Policy 26(9):473483
Freeman C (1994) The diffusion of information and communication technology in the world
economy in the 1990s. In: The management of information and communication technologies:
emerging patterns of control. Aslib, London, pp 841
Fuchs S, Hale K, Axellson P (2007) Augmented cognition can increase human performance in
the control room. In: 2007 IEEE 8th human factors and power plants and HPRCT 13th annual
meeting, Monterey, pp 128132
FutureRFID (2014) First international conference on future RFID technologies, Eger, 6-7 November 2014. www.futurerfid.ektf.hu
Gal Z, Almasi B, Daboczi T, Vida R, Oniga S, Baran S, Farkas I (2014) Internet of things:
application areas and research results of the FIRST project. Infocommun J 6(3):3744
Galambos P (2012) Vibrotactile feedback for haptics and telemanipulation: survey, concept and
experiment. Acta Polytech Hung 9(1):4165
Galambos P, Baranyi P (2011a) Vibrotactile force feedback for telemanipulation: concept and
applications. In: 2011 2nd international conference on cognitive infocommunications (CogInfoCom). IEEE, Budapest, pp 16
Galambos P, Baranyi P (2011b) VirCA as virtual intelligent space for RT-middleware. In: 2011
IEEE/ASME international conference on advanced intelligent mechatronics (AIM), Budapest,
pp 140145. doi:10.1109/AIM.2011.6027133
Galambos P, Weidig C, Baranyi P, Aurich JC, Hammann B, Kreylos O (2012a) VirCA NET:
a case study for collaboration in shared virtual space. In: 3rd IEEE international conference
on cognitive infocommunications. IEEE, Kosice, pp 273277
Galambos P, Weidig C, Zentay P, Csapo d, Baranyi P, Aurich JC, Hammann B, Kreylos O
(2012b) VirCA NET: a collaborative use case scenario on factory layout planning. In: 3rd IEEE
international conference on cognitive infocommunications. IEEE, Kosice, pp 467468
Galambos P, Csapo d, Zentay P, Fulop IM, Haidegger T, Baranyi P, Rudas IJ (2015) Design,
programming and orchestration of heterogeneous manufacturing systems through VR-powered
remote collaboration. Robot Comput Integr Manuf 33:6877
Galanis D, Karabetsos S, Koutsombogera M, Papageorgiou H, Esposito A, Riviello MT (2013)
Classification of emotional speech units in call centre interactions. In: 4th IEEE international
conference on cognitive infocommunications, pp 403406
Galis A, Gavras A (eds) (2013) The future Internet validated results and new horizons. In: Future
Internet assembly, Dublin. Springer, Heidelberg
Galis A, Denazis S et al. (eds) (2004) Programmable networks for IP service deployment. Artech
House Books, Norwood
Ganter B, Wille R (1999) Formal concept analysis. Springer, Berlin
Garriott R (1985) Ultima IV: quest of the avatar. Origin Systems US Copyright Office PA317504
Gartner Inc. (2014) Lead and Transform Your enterprise into the Digital Age, presented by L.
Weldon, Budapest
Garzo A, Benczur AA, Sidlo CI, Tahara D, Wyatt EF (2013) Real-time streaming mobility
analytics. In: 2013 IEEE international conference on big data, pp 697702
Gaver W (1986) Auditory icons: using sound in computer interfaces. Hum Comput Interact
Elsevier Science, Amsterdam 2(2):167177
References
199
Gaver W (1988) Everyday listening and auditory icons. Ph.D. thesis, University of California, San
Diego
Gaver W (1989) The SonicFinder: an interface that uses auditory icons. Hum Comput Interact
4(1):6794
Gaver W (1997) Auditory interfaces. In: Helander HG, Landauer TK, Prabhu P (eds) Handbook of
human-computer interaction. Elsevier, Amsterdam
Gazebo (2012) GAZEBO robot simulation environment. http://www.gazebosim.org/
Gilnyi A (1998) Solving linear functional equations with computer. Math Pannonica 9(1):5770
Giles J (2009) What is ICT? Michalsons. http://www.michalsons.co.za/what-is-ict/2525
Gilovich T, Griffin D, Kahneman D (2002) Heuristics and biases: the psychology of intuitive
judgement. Cambridge University Press, Cambridge
Granstrom B, House D, Swerts M (2002) Multimodal feedback cues in human-machine interactions. In: International conference on speech prosody 2002
Gray J (2000) The Hilbert problems 19902000. Newsletter 36:1013
Greitzer FL, Griffith D (2006) A human-information interaction perspective on augmented
cognition. In: Abstract submitted to augmented cognition international, CA[PNNL-SA-49657],
San Francisco
Griffith D, Greitzer FL (2007) Neo-symbiosis: the next stage in the evolution of human information
interaction. Int J Cogn Inform Nat Intell 1(1):3952
Gripenberg P (2011) The walking video interview (WVI) as a potential technique to tap into the
everyday experiences of ICTs. In: Information systems research seminar in Scandinavia (IRIS),
Turku, pp 139155
Gurkok H, Nijholt A (2012) Brain-computer interfaces for multimodal interaction: a survey and
principles. Int J Hum Comput Interact 28(5):292307
Hale KS, Fuchs S, Berka C (2008) Driving EEG cognitive assessment using eye fixations. In: 2nd
international conference on applied human factors and ergonomics, Las Vegas
Halpern JY (1990) An analysis of first-order logics of probability. Artif Intell 46(3):311350
Hangya V, Farkas R (2013) Target-oriented opinion mining from tweets. In: 4th IEEE international
conference on cognitive infocommunications, pp 251254
Hanson MA, Powell HC Jr, Barth AT, Ringgenberg K, Calhoun BH, Aylor JH, Lach J (2009) Body
area sensor networks: challenges and opportunities. Computer 42(1):5865
Hzy A (2004) Solving linear two variable functional equations with computer. Aequationes Math
67(12):4762. doi:10.1007/s00010-003-2703-9
Hearst MA (1997) Dissonance on audio interfaces. IEEE Expert 12(5):1016
Hecht D, Reiner M (2009) Sensory dominance in combinations of audio, visual and haptic stimuli.
Exp Brain Res 193:307314
Heder M (2014) Emergence and tacit knowledge in machines. Ph.D. thesis. http://www.omikk.
bme.hu/collections/phd/Gazdasag_es_Tarsadalomtudomanyi_Kar/2014/Heder_Mihaly/tezis_
eng.pdf
Hemingway CJ, Gough TG (1998) A socio-cognitive theory of information systems. Research
Report Series, University of Leeds, School of Computer Studies (25):140
Henten A, Samarajiva R, Melody W (2003) Designing next generation telecom regulation: ICT
convergence or multi-sector utility? Info 5(1):2633
Hercegfi K (2011a) Heart rate variability monitoring during human-computer interaction. Acta
Polytech Hung 8(5):205224
Hercegfi K (2011b) Improved temporal resolution heart rate variability monitoring-pilot results
of non-laboratory experiments targeting future assessment of human-computer interaction.
Int J Occup Saf Ergon 17(2):105117
Hermann T (2002) Sonification for exploratory data analysis. Ph.D. thesis, University of Bielefeld
Hermann T (2008) Taxonomy and definitions for sonification and auditory display. In: 14th
international conference on auditory display, pp 18
Hermann T, Ritter H (1999) Listen to your data: model-based sonification for data analysis. In:
Lasker GE (ed) Advances in intelligent computing and multimedia systems. The International
Institute for Advanced Studies in System Research and Cybernetics, Baden-Baden, pp 189194
200
References
Hermann T, Hunt A, Neuhoff J (2011) The sonification handbook. Logos Verlag, Berlin
Hewes D (1995) The cognitive bases of interpersonal communication. Routledge, New York
Hirata-Kohno N, Petho A (2013) On a key exchange protocol based on diophantine equations.
Infocommun J 5(3):1721
Hochberg L, Donoghue J (2006) Sensors for brain-computer interfaces. IEEE Eng Med Biol Mag
25(5):3238
Hodder I (2012) Entangled: An Archaeology of the Relationships between Humans and Things.
Blackwell Publishing Wiley-Blackwell, Chichester, UK. ISBN: 978-0470-67211-2
Hodder I (2014) The entanglements of humans and things: a long-term view. New Lit Hist
45(1):1936
Hokanson B, Hooper S (2000) Computers as cognitive media: examining the potential of
computers in education. Comput Hum Behav 16(5):537552
Hollan J, Hutchins E, Kirsh D (2000) Distributed cognition: toward a new foundation for humancomputer interaction research. ACM Trans Comput Hum Interact 7(2):174196
Hunt A, Hermann T, Pauletto S (2004) Interacting with sonification systems: closing the loop.
In: Banissi E, Brner K (eds) IV04: proceedings of the 8th international conference on
information visualisation (IV04). IEEE Computer Society, London, pp 879884
Hurley S, Noe A (2003) Neural plasticity and consciousness. Biol Philos 18:131168
IDC 2014 (2014) International data corporation: European Internet of things ecosystem and trends.
http://www.idc.com/getdoc.jsp?containerId=IDC_P29863
IDATE (2009) DigiWorld Yearbook 2009, Montpellier. ISBN:978-2-84822-143-4
Ishteva M, Lathauwer L, Absil P, van Huffel S (2008) Dimensionality reduction for higher-order
tensors: algorithms and applications. Int J Pure Appl Math 42(3):337343
ITU (1999) Convergence and regulation. Volume of trends in telecommunication reform. International Telecommunication Union, Geneva
ITU (2009) Measuring the information society: the ICT development index. International Telecommunication Union, Geneva
ITU (2011a) ITU-T recommendation Y.3001: future network vision objectives and design goals.
International Telecommunication Union, Geneva
ITU (2011b) ITU-T recommendation Y.3011: framework of network virtualization for future
networks. International Telecommunication Union, Geneva
ITU (2012a) ITU-T recommendation Y.3021: framework of energy saving for future networks.
International Telecommunication Union, Geneva
ITU (2012b) ITU-T recommendation Y.3031: identification framework in future networks. International Telecommunication Union, Geneva
Jamsa J, Sukuvaara T, Luimula M (2014) Vehicle in a cognitive network. Intell Decis Technol
9(1):1727
Jokinen K (2008) User interaction in mobile navigation applications. In: Meng L, Zipf A, Winter S
(eds) Map-based mobile services. Lecture Notes in geoinformation and cartography. Springer,
Berlin/Heidelberg, pp 168197
Joosten S (1994) Trigger modelling for workflow analysis. In: Proceedings of CON94: workflow
management, challenges, paradigms and products, pp 236247
Kahneman D (2011) Thinking, fast and slow. Farrar, Straus and Giroux, New York, NY, USA
Kalmar A, Vida R, Maliosz M (2013) Context-aware addressing in the Internet of things
using bloom filters. In: 4th IEEE international conference on cognitive infocommunications,
pp 487492
Kapahnke P, Liedtke P, Nesbigall S, Warwas S, Klusch M (2010) An open platform for semanticbased 3D simulations in the 3D Internet. Lect Notes Comput Sci 6497:161176
Karnouskos S, Skarmeta A (2013) The future Internet of things. Introduction to chapter on Internet
of things. In: FIA 2013. Springer, Heidelberg, pp xxvxxvii
Kaye J (2004) Making scents: aromatic output for HCI. Interactions 11:4861.
doi:http://doi.acm.org/10.1145/962342.964333
References
201
Keszei B, Dull A, Logo E, Hamornik BP, Koles M, Tovolgyi S, Hercegfi K (2014) Visual attention
and spatial behavior in VR environment: an environmental psychology approach. In: 2014 5th
IEEE conference on cognitive infocommunications (CogInfoCom). IEEE, pp 247250
Kifor T, Gottdank T, Hajnal A, Baranyi P, Korondi B, Korondi P (2011) Smartphone emotions
based on human-dog interaction. In: 2011 2nd international conference on cognitive infocommunications (CogInfoCom), pp 16
Kimura M, Uwano H, Ohira M, Matsumoto KI (2009) Toward constructing an electroencephalogram measurement method for usability evaluation. In: Jacko JA (ed) Human-computer
interaction. New trends. Springer, Berlin, pp 95104
Kiss G, Vicsi K (2014) Physiological and cognitive status monitoring on the base of acousticphonetic speech parameters. In: Besacier L, Dediu A-H, Martin-Vide C (eds) Statistical
language and speech processing. Springer, Berlin, pp 120131
Kiss G, Sztaho D, Vicsi K, Golemis A (2014) Connection between body condition and speech
parameters - especially in the case of hypoxia. In: 5th IEEE international conference on
cognitive infocommunications, Vietri sul Mare, pp 333336
Kitagawa M, Dokko D, Okamura A, Yuh D (2005) Effect of sensory substitution on suturemanipulation forces for robotic surgical systems. J Thorac Cardiovasc Surg 129(1):151158
Koles M, Hercegfi K, Logo E, Tovolgyi S, Szabo B, Hamornik BP, Komlodi A, Baranyi
PZ, Galambos P, Persa G (2014) Collaboration experience in immersive VR environment
in the frame of the NeuroCogSpace project. In: 2014 5th IEEE conference on cognitive
infocommunications (CogInfoCom). IEEE, pp 373376
Kolmogorov A (1957) On the representation of continuous functions of many variables by
superposition of continuous functions of one variable and addition. Dokl Akad Nauk USSR
114:953956
Koltay K, Borbely M, Dani E, Balazs L, Csernoch M, Bujdoso G (2013) Users retrieval skills in
library environment: ways to measure and improve them. In: 4th IEEE international conference
on cognitive infocommunications, Budapest, pp 319323
Kondor D, Csabai I, Dobos L, Szule J, Barankai N, Hanyecz T, Sebok T, Kallus Z, Vattay G (2013)
Using robust PCA to estimate regional characteristics of language use from geo-tagged twitter
messages. In: 4th IEEE international conference on cognitive infocommunications, pp 393398
Kovcs S, Vincze D, Gcsi M, Miklosi A, Korondi P (2009) Interpolation based fuzzy automaton
for human-robot interaction. In: Preprints of the 9th international symposium on robot control
(SYROCO09). The International Federation of Automatic Control (IFAC), Nagaragawa
Convention Center, Gifu, pp 451456
Kovesdan G, Asztalos M, Lengyel L (2014) Fast android application development with component
modeling. In: 5th IEEE international conference on cognitive infocommunications, Vietri sul
Mare, pp 515520
Kozma R (1991) Learning with media. Rev Educ Res 61(2):179212
Krafft J (2003) Vertical structure of the industry and competition: an analysis of the evolution of
the info-communications industry. Telecommun Policy 27(8):625649
Krafft J (2010) Profiting in the info-coms industry in the age of broadband: lessons and new
considerations. Technol Forecast Soc Chang 77(2):265278
Kurian G, Anand M (2014) GENKEY - 2400 - The Internet of everything ecosystem - bringing IT
and OT together with the Internet of things, San Francisco. https://www.ciscolive.com/online/
connect/sessionDetail.ww?SESSION_ID=78188
Lakatos G, Miklosi A (2012) How can the ethological study of dog-human companionship inform
social robotics? In: Birke L, Hockenhull J (eds) Crossing boundaries: investigating humananimal relationships. Brill, Leiden, pp 187208
de Lathauwer L, Moor BD, Vandewalle J (2000) A multi linear singular value decomposition.
SIAM J Matrix Anal Appl 21(4):12531278
Latre B, Braem B, Moerman I, Blondia C, Demeester P (2011) A survey on wireless body area
networks. Wirel Netw 17(1):118
Lederman S (2004) Haptic identification of common objects: effects of constraining the manual
exploration process. Percept Psychophys 66(4):618628
202
References
Lee EA, Seshia SA (2011) Introduction to embedded systems - a cyber-physical systems approach.
LuLu, Berkeley. ISBN:978-1312-42740-2. http://LeeSeshia.org
Lee H, Lee J, Seo S (2009) Brain response to good and bad design. In: Jacko JA (ed) Humancomputer interaction. New trends. Springer, Berlin, pp 111120
Lemmens P, Bussemakers M, De Haan A (2001) Effects of auditory icons and earcons on visual
categorization: the bigger picture. In: Proceedings of the international conference on auditory
display (ICAD01), Helsinki, pp 117125
Leung WK, Yuen KW, Wong KH, Meng H (2013) Development of text-to-audiovisual speech
synthesis to support interactive language learning on a mobile device. In: 4th IEEE international
conference on cognitive infocommunications, pp 583588
Lewis TG (2009) Network science: theory and applications. Wiley, Hoboken, NJ, USA. ISBN:
978-0470331880
Li Z, Giannini F, Falcidieno B, Pernot JP, Veron P (2013) Towards a new approach for the
description of shapes from multimodal data. In: 4th IEEE international conference on cognitive
infocommunications, Budapest, pp 109114
Liao Sh (2003) Knowledge management technologies and applications-literature review from 1995
to 2002. Expert Syst Appl 25(2):155164
Liu Yl (2013) Convergence in the digital age. Telecommun Policy 37(8):611614.
doi:10.1016/j.telpol.2013.04.012
Logo E, Hamornik BP, Koles M, Hercegfi K, Tovolgyi S, Komlodi A (2014) Usability related
human errors in a collaborative immersive VR environment. In: 2014 5th IEEE conference
oncognitive infocommunications (CogInfoCom). IEEE, pp 243246
Lorentz G (1966) Approximation of functions. Holt, Reinhard and Winston, New York
Maclean K, Enriquez M (2003) Perceptual design of haptic icons. In: Proceedings of eurohaptics
2003, pp 351363
Madhyastha T, Reed A (1994) A framework for sonification design. In: Kramer G (ed) Auditory
display. Addison-Wesley, Reading
Marentakis G, Brewster SA (2005) A comparison of feedback cues for enhancing pointing
efficiency in interaction with spatial audio displays. In: Proceedings of the 7th international
conference on human computer interaction with mobile devices & services, pp 5562
Massimino M (1992) Sensory substitution for force feedback in space teleoperation. Ph.D. thesis,
MIT, Department of Mechanical Engineering
Matsubara D, Egawa T, Nishinaga N, Shin MK, Kafle VP, Galis A (2013) Open the way to future
networks a viewpoint framework from ITU-T. Springer, Berlin
Matuszka T, Vinceller Z, Laki S (2013) On a keyword-lifecycle model for real-time event detection
in social network data. In: 4th IEEE international conference on cognitive infocommunications,
pp 453458
McGee M (2002) Investigating a multimodal solution for improving force feedback generated
textures. Ph.D. thesis, University of Glasgow
Melody WH (1997) Telecom reform: principles, policies and regulatory practices. Den Private
Ingenirfond, Technical University of Denmark
Mihalydeak T, Csajbok ZE (2013) Membranes with boundaries. In: Membrane computing.
Springer, Berlin, pp 277294
Mikhaylov K, Pitkaaho T, Tervounen J, Niemela M (2013) Wireless sensor glove interface
and its application in digital holography. In: 4th IEEE international conference on cognitive
infocommunications, Budapest, pp 325329
Miklosi A, Soproni K (2006) A comparative analysis of animals understanding of the human
pointing gesture. Anim Cogn 9(2):8193
Miklosi A, Kubinyi E, Topal J, Gacsi M, Viranyi Z, Csanyi V (2003) A simple reason for a big
difference: wolves do not look back at humans, but dogs do. Curr Biol 13(9):763766
Minutolo A, Esposito M, De Pietro G (2012) Development and customization of individualized
mobile healthcare applications. In: 3rd IEEE international conference on cognitive infocommunications, Kosice, pp 321326
References
203
Mitola J, Maguire G (1999) Cogitive radio: making software radios more personal. IEEE Pers
Commun 6(4):1318
Mori M (1970) Bukimi no tani (Uncanny valley effect). Energy 7(4):3335
Mulas F, Pilloni P, Manca M, Boratto L, Carta S (2013) Linking human-computer interaction with
the social web: a web application to improve motivation in the exercising activity of users.
In: 4th IEEE international conference on cognitive infocommunications, pp 351356
Mustonen M (2008) A review-based conceptual analysis of auditory signs and their design.
In: Proceeding of ICAD
Nannicelli T, Taberham P (eds) (2014) Cognitive media theory, 1st edn. Routledge, New
York/London
Nass CI, Yen C (2010) The man who lied to his laptop: what machines teach us about human
relationships. Current Trade Penguin Group, New York, NY, USA
Navarretta C (2014a) Feedback facial expressions and emotions. J Multimodal User Interfaces
8(2):135141
Navarretta C (2014b) Predicting emotions in facial expressions from the annotations in naturally
occurring first encounters. Knowl-Based Syst 71:3440
Navarretta C (2014c) Speech, emotions and facial expressions in dyadic spontaneous conversations. Intell Decis Technol 8(4):255263
Newman M, Barabasi AL, Watts DJ (2006) The structure and dynamics of networks. Princeton
University Press, Princeton
Nguyen KK, Cheriet M, Lemay M, Arnaud BS, Reijs V, Mackarel A, Minoves P, Pastrama A,
Van Heddeghem W (2011) Renewable energy provisioning for ICT services in a future Internet.
Springer, Berlin
Niitsuma M, Hashimoto H (2009) Observation of human activities based on spatial memory in
intelligent space. J Rob Mechatronics 21(4):515523
Niitsuma M, Hashimoto H, Hashimoto H (2007) Spatial memory as an aid system for human
activity in intelligent space. IEEE Trans Ind Electron 54(2):11221131
Nijholt A, Tan D (2008) Brain-computer interfacing for intelligent systems. IEEE Intell Syst
23(3):7279
Ning H, Wang Z (2011) Future Internet of things architecture: like mankind neural system or social
organization framework? IEEE Commun Lett 15(4):461463
Nishinaga N (2010) NICT new-generation network vision and five network targets. IEICE Trans
Commun 93(3):446449
Norman D (1991) Cognitive artifacts. Cambridge University Press, New York, pp 1738. http://
www.portal.acm.org/citation.cfm?id=120352.120354
OECD (1992) Telecommunications and broadcasting: convergence or collision? Organization for
Economic Cooperation and Development, Paris
OECD (2011) Guide to measuring the information society. Organization for Economic Cooperation and Development, Paris
Origlia A, Galata V, Cutugno F (2014) Introducing context in syllable-based emotion tracking.
In: 5th IEEE international conference on cognitive infocommunications, Vietri sul Mare,
pp 337342
Ousterhout T, Dyrholm M (2013) Cortically coupled computer vision with emotiv headset using
distractor variables. In: 4th IEEE international conference on cognitive infocommunications,
Budapest, pp 245249
Pallot M, Trousse B, Senach B (2012) A tentative design of a future Internet networking domain
landscape. In: The future Internet. Springer, Berlin, pp 237249
Palovics R, Daroczy B, Benczur AA (2013) Temporal prediction of retweet count. In: 4th IEEE
international conference on cognitive infocommunications, pp 267270
Pang ASK (2013) The distraction addiction: getting the information you need and the communication you want, without enraging your family, annoying your colleagues, and destroying your
soul. Hachette, New York
204
References
Papadimitratos P, La Fortelle A, Evenssen K, Brignolo R, Cosenza S (2009) Vehicular communication systems: enabling technologies, applications, and future outlook on intelligent
transportation. IEEE Commun Mag 47(11):8495
Papadimitriou D, Zahariadis T, Martinez-Julia P, Papafili I, Morreale V, Torelli F, Sales B,
Demeester P (2012) Design principles for the future Internet architecture. In: lvarez F et
al (eds) The future Internet. Springer, Berlin, pp 5567
Pavani F, Spence C, Driver J (2000) Visual capture of touch: out-of-the-body experiences with
rubber gloves. Psychol Sci 11(5):353359
Pentland A (2007) Social signal processing. IEEE Signal Process Mag 24(4):108
Pentland A (2008) Honest signals: how they shape our world. MIT Press, London
Perera C, Zaslavsky A, Christen P, Georgakopoulos D (2014) Context aware computing for the
Internet of things: a survey. IEEE Commun Surv Tutorials 16(1):414454
Persa G, Csapo A, Baranyi P (2012) CogInfoCom systems from an interaction perspective a pilot
application for EtoCom. J Adv Comput Intell Intell Inform 16(2):297304
Persa G, Torok A, Galambos P, Sulykos I, Kecskes-Kovacs K, Czigler I, Honbolygo F, Baranyi
P, Csepe V (2014) Experimental framework for spatial cognition research in immersive virtual
space. In: 2014 5th IEEE conference on cognitive infocommunications (CogInfoCom). IEEE,
pp 587593
Petcu D, Galis A (2013) Introduction to chapter on computing and networking clouds. In: The
future Internet: cloud computing, networking and mobility. Springer, Heidelberg, pp xiiixv
Picard RW (1995) Affective computing. The MIT Press, Cambridge
Picard R (1997) Affective computing. The MIT Press, Cambridge
Picard RW (2003a) Affective computing: challenges. Int J Hum Comput Stud 59(1):5564
Picard RW (2003b) What does it mean for a computer to have emotions. In: Trappl R, Petta P,
Payr S (eds) Emotions in humans and artifacts. MIT Press, Cambridge, pp 213235
Pieska S, Kaarela J, Luimula M (2014) Enhancing innovation capability with cognitive infocommunications. Intell Decis Technol 9(1):6778
Pinto M, Cavallo V, Ohimann T, Espie S, Roge J (2004) The perception of longitudinal
accelerations: what factors influence braking manoeuvres in driving simulators? In: Confrence
simulation de conduite, pp 139151
Pirhonen A (2006) Non-speech sounds as elements of a use scenario: a semiotic perspective.
In: Proceedings of the 12th international conference on auditory display (ICAD2006, London
Platzer E, Petrovic O (2011) An experimental deprivation study of mobile phones, Internet and
TV. Comput Technol Appl 2(8):600606
Pontonnier C, Dumont G, Samani A, Madeleine P, Badawi M (2014) Designing and evaluating
a workstation in real and virtual environment: toward virtual reality based ergonomic design
sessions. J Multimodal User Interfaces 8(2):199208
Preece J, Rogers Y, Sharp H, Benyon D, Holland S, Carey T (1994) Human-computer interaction.
Addison-Wesley Longman Ltd, Harlow, UK
Preissl B, Muller J (2006) Governance of communication networks: connecting societies and
markets with IT. Physica, Heidelberg
Prekopcsak Z, Halacsy P, Gaspar-Papanek C (2008) Design and development of an everyday hand
gesture interface. In: Proceedings of the 10th international conference on human computer
interaction with mobile devices and services, pp 479480
Prekopcsak Z, Makrai G, Henk T, Gaspar-Papanek C (2011) Radoop: analyzing big data with
rapidminer and hadoop. In: Proceedings of the 2nd RapidMiner community meeting and
conference (RCOMM 2011), pp 865874
Prinz J (2006) Putting the brakes on enactive perception. Psyche 12:119
Puszta A, Szule J, Laki S (2013) Near real-time thematic clustering of web documents and other
Internet contents. In: 4th IEEE international conference on cognitive infocommunications,
Budapest, pp 307312
Raptis S (2013) Exploring latent structure in expressive speech. In: 4th IEEE international
conference on cognitive infocommunications. IEEE, pp 741746
References
205
Rasmussen J (1982) Human errors: a taxonomy for describing human malfunction in industrial
installations. J Occup Accid 4(2):311333
Reason J (1990) Human error. Cambridge University Press, Cambridge
Recker M, Ram A, Shikano T, Li G, Stasko J (1995) Cognitive media types for multimedia
information access. J Educ Multimedia Hypermedia 4(23):183210
Riecke BE, Schulte-Pelkum J, Caniard F, Bulthoff HH (2005) Towards lean and elegant selfmotion simulation in virtual reality. In: Proceedings of virtual reality, 2005 (VR 2005). IEEE,
pp 131138
Riva G, Davide F (2001) Communications through virtual technologies. Identity, community and
technology in the communication age. IOS Press, Amsterdam, pp 124154
Rochlis J (2002) Human factors and telerobotics: tools and approaches for designing remote
robotics workstation displays. Ph.D. thesis, Massachusetts Institute of Technology
Romportl J, Zackova E, Kelemen J (eds) (2015) Beyond artificial intelligence the disappearing
human-machine divide. Topics in intelligent engineering and informatics, vol 9. Springer
International Publishing, Switzerland
Roschelle J (1996) Designing for cognitive communication: epistemic fidelity or mediating
collaborative inquiry? In: Computers, communication and mental models. Taylor & Francis,
London, UK pp 1527
RTC (2008) Robot technology component specification. http://www.omg.org/spec/RTC/
Ruxton GD, Schaefer HM (2011) Resolving current disagreements and ambiguities in the
terminology of animal communication. J Evol Biol 24(12):25742585
Ryan M et al. (eds) (2003) The EU regulatory framework for electronic communications and
related EU legislation. Handbook. Arnold and Porter, London
Saito T (1994) An evolving scenario of communication network towards b-ISDN. In: Iversen
VB (ed) Integrated broadband communication networks and services, vol 18. North Holland,
Amsterdam
Sales B, Darmois E, Papadimitriou D, Bourse D (2012) A systematic approach for closing the
research to standardization gap. In: lvarez F et al (eds) The future Internet. Springer, Berlin,
pp 1829
Sallai G (2007) Converging information, communication and media technologies. In: Banse G
(ed) Assessing societal implications of converging technological development. Edition Sigma,
Berlin, pp 2543
Sallai G (2012a) The cradle of cognitive infocommunications. Acta Polytech Hung 9(1):171181
Sallai G (2012b) Defining infocommunications and related terms. Acta Polytech Hung 9(6):515
Sallai G (2013a) Chapters of Future Internet Research. In: 2013 IEEE 4th international conference
on cognitive infocommunications (CogInfoCom). IEEE, New York, pp 161166
Sallai G (2013b) The FIRST project and the Future Internet National Research Programme. In:
Future Internet PPP workshop on building an eco-system for delivering innovative future
Internet services and applications, IEEE ICC 2013. IEEE, Budapest
Sallai G (2013c) From telecommunications to cognitive infocommunications and Internet of things
- phases of digital convergence. In: 2013 IEEE 17th international conference on intelligent
engineering systems (INES), pp 1317
Sallai G (2014) Future Internet visions and research clusters. Acta Polytech Hung 11(7):524
Sayrafian-Pour K, Yang WB, Hagedorn J, Terrill J, Yazdandoost KY, Hamaguchi K (2010) Channel
models for medical implant communication. Int J Wireless Inf Networks 17(34):105112
Sayrafian-Pour K, Hagedorn J, Barbi M, Terrill J, Alasti M (2013) A simulation platform to study
inter-BAN interference. In: 4th IEEE international conference on cognitive infocommunications, pp 345350
Scaletti C (1994) Sound synthesis algorithms for auditory data representations. In: Kramer G (ed)
Auditory display. Addison-Wesley, Reading
Schmorrow D, Stanney KM, Wilson G, Young P (2006) Augmented cognition in human-system
interaction. In: Handbook of human factors and ergonomics, 3rd edn. Wiley, New York,
pp 13641383
206
References
Scott-Phillips TC, Blythe RA, Gardner A, West SA (2012) How do communication systems
emerge? Proc R Soc B Biol Sci 279(1735):19431949
Sharmin S (2003) Evaluating non-visual feedback cues for touch input device. In: Evreinov G (ed)
New interaction techniques. University of Tampere, Tampere. pp 39. ISBN:951-44-5188-0
Sharples M, Jeffery N, Du Boulay JBH, Teather D, Teather B, Du Boulay GH (2002) Sociocognitive engineering: a methodology for the design of human-centred technology. Eur J Oper
Res 136(2):310323
Sheridan T (1992) Musings on telepresence and virtual presence. Presence Teleoperators Virtual
Environ 1(1):120126
Sheridan T (1994) Human factors considerations for remote manipulation, In: Proceedings of the
advanced guidance and control aspects in robotics, Lisbon, pp 124 ISBN: 92-835-0751-7
Sheth AP (2009) Citizen sensing, social signals, and enriching human experience. IEEE Internet
Comput 13(4):87
Shneiderman B (1998) Designing the user interface: strategies for effective human-computer
interaction, 3rd edn. Addison-Wesley, Reading
Skinner A, Long L, Vice J, Blitch J, Fidopiastis CM, Berka C (2013) Augmented interaction:
applying the principles of augmented cognition to human-technology and human-human
interactions. In: Foundations of augmented cognition. Springer, Berlin, pp 764773
Smith D (1975) Pygmalion: a computer program to model and stimulate creative thought. Ph.D.
thesis, Stanford University, Department of Computer Science
Smith IG (2012) The Internet of things 2012: new horizons. CASAGRAS2, Halifax
Smith C, Kisiel K, Morrison J (2009) Working through synthetic worlds. Ashgate, London
Solvang B, Sziebig G (2012) On industrial robots and cognitive info-communication. In: 3rd IEEE
international conference on cognitive infocommunications, Kosice, pp 459464
Spitters S, Sanders M, op den Akker R, Bruijnes M (2013) The recognition of acted interpersonal
stance in police interrogations. In: 4th IEEE international conference on cognitive infocommunications, Budapest, pp 6570
Sprecher D (1965) On the structure of continuous functions of several variables. Trans Am Math
Soc 115:340355
St John M, Kobus DA, Morrison JG, Schmorrow D (2004) Overview of the DARPA augmented
cognition technical integration experiment. Int J Hum Comput Interact 17(2):131149
Staal MA, Bolton AE, Yaroush RA, Bourne LE Jr (2008) Cognitive performance and resilience to
stress. In: Lukey B, Tepe V (eds) Biobehavioral resilience to stress. Francis & Taylor, London,
pp 259299
Stanney KM, Schmorrow DD, Johnston M, Fuchs S, Jones D, Hale KS, Ahmad A, Young P (2009)
Augmented cognition: an overview. Rev Hum Factors Ergon 5(1):195224
Stein B, Wallace M, Meredith A (1995) Neural mechanisms mediating attention and orientation to
multisensory cues. In: Gazzaniga M (ed) The cognitive neurosciences. MIT Press, Cambridge,
pp 683702
Stone L (2011) Just breathe: building the case for email apnea (the huffington post). http://www.
huffingtonpost.com/linda-stone/just-breathe-building-the_b_85651.html
Stratogiannis DG, Tsiropoulos GI, Vinel A, Koucheravy Y, Tsiropoulou EE (2013) Special issue
on mobile computing and networking technologies. Telecommun Syst 52(4):17142145
Streitz N, Nixon P (2005) The disappearing computer. Commun ACM 48(3):3235
Sun R, Merrill E, Peterson T (2001) From implicit skills to explicit knowledge: a bottom-up model
of skill learning. Cognit Sci 25(2):203244
Szabo C, Roka A, Farago T, Gacsi M, Miklosi A, Korondi P (2012) Building a human-dog
interaction inspired emotional engine model. In: IECON 2012 38th annual conference on IEEE
industrial electronics society, pp 55165522
Szabo R, Farkas K, Ispany M, Benczur AA, Batfai N, Jeszenszky P, Laki S, Vagner A, Kollar L,
Sidlo C (2013) Framework for smart city applications based on participatory sensing. In: 4th
IEEE international conference on cognitive infocommunications, pp 295300
References
207
Szegletes L, Koles M, Forstner B (2014) The design of a biofeedback framework for dynamic
difficulty adjustment in games. In: 5th IEEE international conference on cognitive infocommunications, Vietri sul Mare, pp 295299
Szucs G (2013) Decision trees and random forest for privacy-preserving data mining. In: Tarnay K,
Imre S, Xu L (eds) Research and development in e-business through service-oriented solutions.
IGI Global, Hershey, pp 7190
Stevenson D (ed) (1997) Information and communications technology in UK schools, an Independent Inquiry. The Independent ICT in Schools Commission, London
Takagi H (2001) Interactive evolutionary computation: fusion of the capabilities of EC optimization and human evaluation. Proc IEEE 89(9):12751296
Tan D, Nijholt A (2010) Brain-computer interfaces: applying our minds to human-computer
interaction. Springer, Berlin
Tanaka H, Sakti S, Neubig G, Toda T, Campbell N, Nakamura S (2012) Non-verbal cognitive skills
and autistic conditions: an analysis and training tool. In: 3rd IEEE international conference on
cognitive infocommunications, Kosice, pp 4146
Tapolcai J, Gulyas A, Heszbergery Z, Biro J, Babarczi P, Trossen D (2012) Stateless multi-stage
dissemination of information: source routing revisited. In: 2012 IEEE Global communications
conference (GLOBECOM), pp 27972802
Tapolcai J, Heszberger Z, Retvari G, Biro J (2013) Reduced information scenario for shared
segment protection. In: 4th IEEE international conference on cognitive infocommunications,
pp 589594
Taylor TL (2002) Living digitally: embodiment in virtual worlds. In: The social life of avatars.
Springer, London, pp 4062
Telpolicy (1994) Special issue on competition and convergence. Telecommun Policy 18(8)
Thomas RW, Friend DH, Dasilva LA, Mackenzie AB (2006) Cognitive networks: adaptation and
learning to achieve end-to-end performance objectives. IEEE Commun Mag 44(12):5157
Thomessen T, Kosicki T (2011) Cognitive audio-visual infocommunication applied in remote
support for industrial robot systems. In: 2011 2nd international conference on cognitive
infocommunications (CogInfoCom), pp 15
Thomessen T, Niitsuma M (2013) Cognitive human-machine interface with multi-modal manmachine communication. In: 4th IEEE international conference on cognitive infocommunications, Budapest, pp 873876
Todorov E (2004) Optimality principles in sensorimotor control. Nat Neurosci 7(9):907915
Topal J, Miklosi A, Csanyi V, Doka A (1998) Attachment behavior in dogs (canis familiaris): a new
application of Ainsworths (1969) strange situation test. J Comp Psychol 112(3):219
Torok M, Kosa Z (2015) Association in knowledge management technologies. In: 2015 IEEE
13th international symposium on applied machine intelligence and informatics (SAMI). IEEE,
pp 301306
Torok M, Toth MJ, Szollosi A (2013) Foundations and perspectives of mathability in relation to the
CogInfoCom domain. In: 4th IEEE international conference on cognitive infocommunications,
Budapest, pp 869872
Torok A, Sulykos I, Kecskes-Kovacs K, Persa G, Galambos P, Kobor A, Czigler I, Csepe V,
Baranyi P, Honbolygo F (2014) Comparison between wireless and wired EEG recordings in a
virtual reality lab: case report. In: 2014 5th IEEE conference on cognitive infocommunications
(CogInfoCom). IEEE, pp 599603
Toth B, Nagy P, Nemeth G (2012) New features in the voxaid communication aid for speech
impaired people. Springer, Berlin
Trinh TA, Gyarmati L, Sallai G (2011) Understanding the impact of loyal user behaviour on
Internet access pricing: a game-theoretic framework. Telecommun Syst 48(12):4361
Trung DN, Jung JJ, Kiss A (2013) Towards modeling fuzzy propagation for sentiment analysis in
online social networks: a case study on TweetScope. In: 4th IEEE international conference on
cognitive infocommunications, pp 331338
Tselentis G et al (eds) (2010) Towards the future Internet: emerging trends from European research.
In: Future Internet assembly, Valencia. IOS Press, Amsterdam
208
References
References
209
Wang Y, Kinsner W (2006) Recent advances in cognitive informatics. IEEE Trans Syst Man
Cybern 36(2):121123
Watabe T, Niitsuma M (2013) Mental map generation assistance tool using relative pitch difference
and angular information for visually impaired people. In: 4th IEEE international conference on
cognitive infocommunications, pp 255260
WEF (2007) World Economic Forum: digital ecosystem - convergence between IT, Telecoms,
Media and Entertainment: Scenarios to 2015. World Scenario Series
Weikum G, Ntarmos N, Spaniol M, Triantafillou P, Benczr A et al (2011) Longitudinal analytics
on web archive data: its about time! 5th Biennial Conference on Innovative Data Systems
Research (CIDR 2011), 912 January 2011, pp 199202. Asilomar
Welch R, Warren D (1986) Intersensory interactions. In: Boff L, Thomas J (eds) Handbook of
perception and human performance, vol 1. Wiley, New York, pp 2536
Wilson L (2003) Interactivity or interpassivity: a question of agency in digital display. In:
Proceedings of the 5th international digital arts and culture conference, Melbourne, pp 14
Wolpert DM, Ghahramani Z (2000) Computational principles of movement neuroscience. Nat
Neurosci 3:12121217
Wolpert DM, Kawato M (1998) Multiple paired forward and inverse models for motor control.
Neural Netw 11(7):13171329
WUN CogCom (2008) Worldwide universities network cognitive communications consortium established. http://www.wun-cogcom.org/
Yang GZ (2014) Body sensor networks, 2nd edn. Springer, Berlin
Yang WB, Sayrafian-Pour K (2012) Interference mitigation using adaptive schemes in body area
networks. Int J Wireless Inf Netw 19(3):193200
Yao Y (2009) Interpreting concept learning in cognitive informatics and granular computing. IEEE
Trans Syst Man Cybern 39(4):855866
Yoonaidharma S, Kilenthong W, Bunaramrueang P, Kesavayuth D (2014) Special issue on moving
forward with future technologies: opening a platform for all. Telecommun Policy 38(89):
659850
Yuce MR (2010) Implementation of wireless body area networks for healthcare systems. Sensors
Actuators A Phys 162(1):116129. doi:10.1016/j.sna.2010.06.004
Zahariadis T, Daras P, Bouwen J, Niebert N, Griffin D, lvarez F et al (2010) Towards a contentcentric Internet. In: Tselentis G et al (eds) Towards the future Internet - emerging trends from
European Research. IOS Press, Amsterdam, pp 227236
Zahariadis T, Papadimitriou D, Tschofenig H, Haller S, Daras P, Stamoulis GD, Hauswirth M
(2011) Towards a future Internet architecture. Springer, Berlin
Zander TO, Kothe C (2011) Towards passive brain-computer interfaces: applying brain-computer
interface technology to human-machine systems in general. J Neural Eng 8(2):025005
Zeller D, Olsson M, Blume O, Fehske A, Ferling D, Tomaselli W, Godor I (2013) Sustainable
wireless broadband access to the future Internet-the earth project. Springer, Berlin
Zhang D, Guo B, Yu Z (2011) The emergence of social and community intelligence. Computer
44(7):2128
Ziegler S, Crettaz C, Ladid L, Krco S, Pokric B, Skarmeta AF, Jara A, Kastner W, Jung M (2013)
IoT6 moving to an IPv6-based future IoT. In: FIA. Springer, Berlin
Zimmerman T (1996) Personal area networks: near-field intrabody communication. IBM Syst
J 35(3):609617
Zimmerman TG (1999) Wireless networked digital devices: a new paradigm for computing and
communication. IBM Syst J 38(4):566574
Glossary
This glossary summarizes those terms that were either newly introduced, or
considered from a novel perspective in the book. The numbers listed next to each of
the items refer to the sections which may be consulted for further details (in some
cases, more than one section is indicated if the same term is discussed in different
places from different perspectives).
3D Internet 12.4
CogInfoCom channels 5.1, 8.1.2
Cognitive capabilities 2.3.2
Cognitive entity 2.3.2
Cognitive Internet Ecosystem 13.2
Cognitive networks 3.2.3, 4.6
Conceptual mapping (direct/analogy-based) 9.2
Cues 11.2
Data and content technologies 12.4
Digital convergence 3.1, 3.3
Digital Ecosystem 3.1, 3.3, 13.1
Electronic communications 3.2.2
Ethologically inspired CogInfoCom (EtoCom) 6.4
Future Internet 12
FI PPP (Future Internet Public-Private Partnership) 12.2
Generation CE 2.3.2
211
212
Glossary
Index
A
Abuczki, A., 65
Affective computing, 37, 4041
Ainsworths strange situation test
Fuzzy automaton and Fuzzy rule
interpolation, 8991
human-dog interactions, 91
test procedure, 8889
Analogy-based mapping
CogInfoCom channels, 138
corroborative stimulation, 137
parameter mapping, 136
scenario-based orchestration, 138
structural mapping, 136137
Arnold, V., 97
Auditory icons, 108, 112
Augmented cognition (AugCog), 37, 4142
Augmented reality, 179
Auvray, M., 50
B
Bach-y-Rita, P., 50, 51
Bajk, S., 98
Baranyi, P., 13, 69, 110
Barrass, S., 124
Benus, S., 60, 67
Berthelon, F., 69
Bjorn Solvang, 91
Blake, R., 164
Blattner, M., 115
Blum, J.R., 70
Body area networks (BANs), 37, 43
Body icons, 114, 116
Bonin, F., 64
Borus, G., 98
C
Campbell, N., 13, 60
Cerf, V.G., 173
Cheng, D.S., 67
Chernoff, H., 116
CogInfoCom channels
bounded-input POMF, 125
channel differentiation, 58
context-based approaches and applications
in multimodal feedback, 69
in verbal signals, 6768
in written text, 68
DBIPOMFs, 125
definition, 121122
discretized bounded-input POMF, 125
generation vector, 124
generative model, 126127
high-level human communication
CogInfoCom trigger, 164
definitions, 166
directness, 165
mental and emotional relationships, 163
trigger and response types, 166, 167
volition, 164165
human-ICT communication, 158
icon-based approaches and applications
(see Icons)
interaction and communications, 57
low-level biological communication
213
214
cues, 159162, 167168
messages, 162, 163, 167168
ritualization, 159
signals, 159, 161, 162, 167168
mapping techniques
analogy-based mapping, 136138
automated reasoning systems,
139
corroborative stimulation, 138
direct mapping, 135136
high-level direct mapping
techniques, 138
modality, 122123
natural communication, 158159
nomenclature and notations, 123124
ORA model (see Object attribute relation
(ORA) model)
parameter-generating function, 124
pattern-based approaches and applications,
6970
perceptual concepts, 120121
perceptual gradation vector, 124
progressive learning, 169
regressive learning, 169
ritualization, 58
socio-cognitive ICT, 6061
speechability, 5960
spiral discovery method, 58
structural and semantic elements, 58
tunability
generic tuning model, 146147
HOSVD-based canonical form,
145146
POTPFs, 145
SDM (see Spiral discovery method
(SDM) method)
tensor product functions, 144145
Cognitive being, 20
Cognitive capabilities, 173
Cognitive capability, 106
Cognitive entity, 20
Cognitive infocommunications
(CogInfoCom), 4, 14
channels (see CogInfoCom channels)
cognitive capability, 4
cues, 159162, 167168
definition, 4
defnition, 14
design issues relevant, 92
engines, 8182
functionality, 4
generic perspective, 4
humanoid applications, 179
Index
icons
auditory icons, 112
body icons, 114
dynamic icons, 112113
emotional icons, 113114
haptic icons, 112
kinesthetic icons, 114
network packet icons, 114
olfactory icons, 112
visual icons, 112
implicit and explicit assumptions, 1
cognitive capability, 1820
emerging and entanglement,
1618
industrial robotics and production
management, 9293
messages, 110, 162, 163,
167168
qualitative and quantitative analysis,
4
scientific priming effects (see Scientific
priming effects)
signals, 159, 161, 162, 167168
streams, 110111
synergic contribution, 40
tanglefaces, 8182
transition, 2122
trigger, 164
VirCA, industrial capabilities, 9395
Cognitive informatics (CI), 44
Cognitive Internet Ecosystem (CIES)
DES (see Digital Ecosystem (DES))
IES, (see Internet Ecosystem (IES))
Internet of Things, 189
long-term vision, 190191
3D Internet and community
applications, 189
Cognitive networks, 30, 190
definition, 45
Future Internet
cognitive content space, 45
Internet of Things (IoT), 4647
3D Internet (3DI), 46, 47
Cognitive thing, 20
Compact form of higher-order singular value
decomposition (CHOSVD), 146
Corroborative stimulation, 137
Crowdsourcing and crowdsensing community
applications, 179
Csapo, A., 13, 69, 110
Csorba, K., 69
Cyber devices (CDs), 74
Czirbusz, S., 98
Index
D
Dahlbom, B., 16
DBIPOMFs. See Discretized bounded-input
partially ordered multivariate
functions (DBIPOMFs)
Deference thesis, 50
Digital Age, 188189
Digital convergence, 1
Internet, 25
Digital convergence prism, 34
Digital ecosystem (DES), 35
Digital Age, 188189
middle kingdoms scenario, 188
safe havens scenario, 188
youniverse scenario,
188
Direct mapping, 135136
Discretized bounded-input partially ordered
multivariate functions (DBIPOMFs),
126
Dominance thesis, 50
Dropuljic, B., 68
Dynamic icons, 112113, 115116
E
Earcons, 115
e-content, 29
Eichhorn, W., 99
Electronic communications, 2728
Emergent cognitive capabilities,
190
Emoticons, 116
Emotional icons, 113114
Enriquez, M., 109, 115
Ericson, A., 93
Esposito, A., 60
Ethologically informed CogInfocom
(EtoCom)
Ainsworths strange situation test
Fuzzy automaton and Fuzzy rule
interpolation, 8991
human-dog interactions, 91
test procedure, 8889
CogInfoCom tanglefaces and engines,
8182
EtoMotor, 8385
EtoPhone, 8587
human-ICT co-evolution, 80
EtoMotor, 8385
EtoPhone, 8587
Eyecons, 115
215
F
Fekete, K., 69
FI public-private partnership (PPP), 178
Fixed and mobile telephony convergence
(FMC), 27
Foldesi, P., 42
Future Internet (FI), 37
cognitive capabilities, 173
cognitive content space, 45
features, 179
functions, 178179
Internet of Things (IoT), 4647, 173, 175
limitations, 174175
networking, evolution of, 173174
recent capabilities, 179
research goals and spheres, 179180
community applications, 184
cyber-physical systems and
applications, 184
data and content technologies, 182
experimentation, standardization,
regulation, 184185
Internet of Things (IoT), 183
Internet Science, 181
modeling, analysis and design, 181
network architectures, 181182
3D internet and cognitive
infocommunications, 183
smart applications, 173
technological opportunities, 175
3D Internet (3DI), 46, 47, 173
Vinton G. Cerf and Robert E. Kahn (fathers of
the Internet), 173
virtual collaborations, 173
visions
data and content awareness, 177
FI-based applications, 178
intelligent, innovative and secure society,
177
ITU-T, 175176
NICT, 175176
NWGN, 175176
scalable, service aware network, 176177
sustainability, environmental awareness,
177
virtual, resource aware network, 177
G
Galanis, D., 67
Ganter, B., 130
Garriott, R., 52
216
Index
H
Haptic icons, 108, 112, 115
Hashimoto, H., 13
Hzy, A., 98
Hemingway, C.J., 61
Hercegfi, K., 78
Hermann, T., 110
Higher-order singular value decomposition
(HOSVD), 142143
High-level direct mapping, 135136
Hodder, I., 17
Human-computer interactions (HCI), 4748
intersensory integration, 49
sensory dominance, 49
sensory modalities, 49
Human-device communication, 166
Human-human communication, 166
J
Jokinen, K., 13
Joosten, S., 164
I
Icons
auditory icons, 108
body icons, 114
dynamic icons, 112113
emotional icons, 113114
haptic icons, 108, 112
kinesthetic icons, 114
in multimodal feedback, 6566
network packet icons, 114
olfactory icons, 112
in physiological signals, 63
tactile icon, 108
in verbal signals, 6465
visual icons, 112
in written text, 65
Infocommunications (Infocom), 33
Information and communications technology
(ICT), 1
Information Society Technologies (IST), 33
Inter-cognitive communication, 1415
International Telecommunication Union (ITU),
28, 34
L
Lathauwer, L., 146, 149, 150
Lee, J.-H., 13
Linear functional equations, 9899
Li, Z., 92
Logo, E., 79
Low-level direct mapping, 135
K
Kaczmarek, K., 50
Kahn, R.E., 173
Kato, T., 13
Kifor, T., 85
Kinesthetic icons, 114, 116
Kiss, G., 64
Knowledge-based errors, 79
Kolmogorov, A., 97
Komlodi, A., 78
Kondor, D., 68
Kotoku, T., 13
Kovacs, S., 82, 89
M
MacLean, K., 109, 115
Maguire, G., 6
Magyar, G., 13
Marentakis, G., 162
Massive open online courses (MOOCs), 179
Mathability
grading system, 97
granular and analytical representations,
9798
granular formulations, 96
ICT devices, 96
Linear functional equations, 9899
mathematical capabilities, 95
Index
Mathability (cont.)
proper solution, 9596
sets of functional equations, 102
Theorems, 99102
McGee, M., 107
Meng, H., 13
Miklosi, A., 80, 82
Mitola, J., 6
Myin, E., 50
N
Nass, C.I., 163, 164
National Institute of Information and
Communications Technology
(NICT), 175
Nmeth, G., 14
Network packet icons, 114, 116
New-generation networks (NWGN), 175
Next generation Internet. See Future
Internet (FI)
Niitsuma, M., 14, 47
Nixon, P., 48
Norman, D., 135
O
Object attribute relation (ORA) model
definitions, 130131
interactive concepts, 134
message generated perceptual concepts
fully specified, 132133
interactive concepts, 134
perceptual concepts, 132
perceptual evaluation function, 133134
Olfactory icons, 112
ORA model. See Object attribute relation
(ORA) model
Origlia, A., 67
P
Ples, Z., 98
Plovics, R., 68
Pang, A.S.K., 16, 17
Parameter mapping, 136137
Partially ordered multivariate (POMFs)
functions, 125
Partially ordered multivariate tensor product
(TP) functions (POTPFs), 145
Perceptual evaluation function, 133134
Persa, G., 85
Personal area networks (PANs), 43
Picard, R., 40
217
Pieska, S., 93
POMFs functions. See Partially ordered
multivariate (POMFs) functions
Progressive learning, 169
Puszta, A., 68
R
Raptis, S., 64
Rasmussen/Reason framework errors, 79
Regressive learning, 169
Representation-bridging communication, 15
Representation-sharing communication, 15
Ritter, H., 110
Romportl, J., 16
RT-Middleware (RTM), 7475
Rule-based errors, 7980
S
Sallai, G., 7, 14
Scenario-based mapping, 137138
Schmorrow, D., 41
Scientific priming effects
cognitive communication, 6
cognitive informatics, 6
cognitive media, 5
convergence process, 7
infocommunication, 4, 78
media informatics and media
communications, 5
Sensor-bridging communication, 15
Sensor-sharing communication, 15
Sensory dominance, 5051
Sensory modality, 106
Sensory signals structure
auditory icons and earcons, 109
CogInfoCom icons
auditory icons, 112
body icons, 114
dynamic icons, 112113
emotional icons, 113114
haptic icons, 112
kinesthetic icons, 114
network packet icons, 114
olfactory icons, 112
visual icons, 112
CogInfoCom messages, 110, 114116
CogInfoCom streams, 110111
cognitive capability, 106
context of communication
directness, 106
mode, 105
218
temporal reference, 106
type, 106
voluntariness, 106
modality, 108
olfactory icons and smicons, 109
sensory modality, 106
sonification, 110
Sharmin, S., 162
Sharples, M., 61
Shneiderman, B., 112
Skill-based errors, 79
Smart applications, 173
Smicons, 115
Smith, D.C., 112
Social signal processing (SPS), 37, 5152
Socio-cognitive ICT, 37
Solvang, W.D., 91
Speechability, 37
Spiral discovery method (SDM) method,
58, 144
audio-based CogInfoCom channels,
152153
CogInfoCom application scenarios, 142
cognitive artifact, 142
confusion matrices, 155156
error sizes, 155
generate CogInfoCom messages, 141
generation parameter configurations, 152,
154
generic model, 152153
HOOI, 143
HOSVD, 142143
implementation, 151152
interpretability and complexity, 147148
lemma and theorem, 149150
minimal-rank orthogonal system, 147
original parameter-generating function, 148
yield optimal rank-reduction, 148
Spitters, S., 70
Streitz, N., 48
Structural mapping technique, 136137
Szabo, C., 84
Szegletes, L., 63
T
Tactile icon, 108
Technological priming effects
discrete gestures, 10
elementary capabilities, 9
enlightenment, 8
inflated expectations, 8
plateau of productivity, 9
Index
R&D initiatives, 10
speech utterances/facial information, 10
technological hype cycle, 8
technology trigger, 8
trough of disillusionment, 8
Telecommunications
cognitive infocommunications, 3132
communication sectors, 2627
content space, 3132
electronic communications, 2728
infocommunications, 2830
ISP, 29
Three dimensional (3D) Internet, 31, 173
3D virtual collaboration, 179
Topl, J., 86, 91
U
Uckelmann, D., 46
V
Vinciarelli, A., 51, 69
Vincze, D., 89
Virtual and augmented avatars, 37
definition, 52
episodic interactions, 53
identity, 53
metacommunicational fidelity, 53
objects and information, 53
presence, 53
qualifications, 52
Virtual and augmented reality, 37
communication capabilities, of human, 54
encumberment, 54
sensory capabilities, 55
situation awareness, 54
Virtual Collaboration Arena (VirCA) platform
CDs, 74
CogInfoCommediated neuroscientific
analysis, 77
experiment descriptor, 78
industrial capabilities, 9395
MTA SZTAKI, 74
multi-user scenarios, 7576
neuroscience inspired tools, 77
new synergies, development of, 7677
Rasmussen/Reason framework errors,
79
RTM, 7475
structure, 75
wired and wireless EEG configurations,
78
Index
219
W
Wang, Y., 130, 131
Wersenyi, G., 110
Y
Yen, C., 163, 164