Vous êtes sur la page 1sur 4

Voice Recognition and Voice Recognition History

Research Paper

Voice Recognition is a technology that is in our every-day lives. Whenever we call to
make a payment or pay a bill, often times the company already has a voice recognition program
set up, eliminating the need to talk to a real-life person. Much similar technology is used in many
smart phones today.
Generally speaking, speech recognition programs fall into two separate categories: Small-
vocabulary with many users, and large-vocabulary with limited or often times single users.
Small-vocabulary-many user systems are ideal for automated telephone answering, a
system many large companies use. This type of system has a limited small number of
predetermined commands and inputs, such as basic menu options or numbers. However the
benefit to having a small vocabulary is that these systems are often able to pick up on different
variations in accent, and speech patterns.
Large vocabulary-limited user systems are usually in the business place or at home for
personal usage. They can identify more words with a high success rate, but can only to a few
users. The more users that use this system, the less success rate. Many smart phones use this type
of system.
The famous Siri is an intelligent personal assistant, and knowledge navigator which
works on many of apples famous iPhones would be classified as a large vocabulary with limited
users system. You simply can ask a question, make a recommendation, or tell her to perform an
action, and she will be able to recognize, and interpret what was said.
Siri, used by apples iPhones uses a natural language user interface. Naural language
interfaces operate using three different steps: The first step is to analyze. Siri analyzes the input
that is spoken, using a robust linguistic understanding library to understand and then derive the
meaning. The second step Siri uses is the use of reason. In this step, the application determines
the most apporiate way to react, taking in count different factors such as time, location, and
previous dialogues and data looked up by the user. The final step is to react. This is where Siri
will give output information, by either opening a webpage, asking for more information, or
giving a response.
Voice recognition technology started way back in the 1950s. The first system was Bell
Laboratories Audrey. This system could only understand digits. Ten years later in 1962, IBM
developed Shoebox which could understand 16 English words.
In the 1970s however is when speech recognition really took off. The U.S. Departement
of Defense

However with this type of technology arises many problems. The first being different
enunciations of words. The famous quote of You say po-tay-to, I say po-tah-to, you say
tomayto, I say tomahto comes into play. People often times annunciate words differently. This
could than trick said application, resulting in something completely different than what was
said.
Another problem with this type of technology is the many homonyms in the English
language. There are so many words that sound alike, but have different meanings. Some
examples could be accept and except, acts and ax, to, two, and too, with what seems like never
ending list.

Vous aimerez peut-être aussi