Department of Computer Science Logic in Computer Science Group Assignment Section C Group 8 Name: Id: 1. Aman Hassen CIR//07 2. Jiregna Admasu CIR/272/07 3. Meron Fikadu CIR/329/07 4. Meron Rufael CIR/554/07 5. Terefe Birhanu CIR/453/07 6. Getabefikadu Temesgen CIR/212/07 Probabilistic logic Logic traditionally concerns matters immutable, necessary and certain, while probability concerns the uncertain, the random, and the capricious. If the premises of a valid argument are all certain, then so is the conclusion. Suppose, on the other hand, that the premises are not all certain, but probable to various degrees; can we then put bounds on the probability of the conclusion? Or suppose that we want the probability of the conclusion of a given valid argument to be above a particular threshold; how probable, then, must the premises be? Cont.. The aim of a probabilistic logic aka probability logic/probabilistic reasoning is to combine the capacity of probability theory to handle uncertainty with the capacity of deductive logic to exploit structure of formal argument. Probabilistic logics attempt to find a natural extension of traditional logic truth tables: the results they define are derived through probabilistic expressions instead. Cont.. A difficulty with probabilistic logics is that they tend to multiply the computational complexities of their probabilistic and logical components. In a semantical generalization of ordinary first-order logic the truth values of sentences can range between 0 and 1. The truth value of a sentence in probabilistic logic is taken to be the probability of that sentence in ordinary first-order logic.
We make precise the notion of the probability of a sentence
through a possible worlds analysis. 1. Probabilities on Possible Worlds A sentence S can be either true or false if we were concerned about just the one sentence S, we could imagine two sets of possible worlds one, say Wl, containing worlds in which S was true and one, say W2, containing worlds in which S was false. The actual world, the world we are actually in, must be in one of these two sets, but we might now know which one. We can model our uncertainty about the actual world by imagining that it is in Wl, with probability Pl, and is in W2 with some probability P2 = 1 - Pl In this sense we can say that the probability of S (being true) is p1. Cont.. Consider the sentence The sky is blue". Clearly we have (by common experience) that The sky is necessarily blue" is false (sunsets?) and The sky is possibly blue" is true. So, there are situations or possible worlds in which The sky is blue" is true, and some in which it is false. If we have more sentences, we have more sets of possible worlds. Sentences may be true in some worlds and false in others in different combinations. Cont.. Each set contains worlds with a unique and consistent set of truth values for the sentences. If we have L sentences, we might have as many as 2 L sets of possible worlds. Typically though, we will have fewer than this maximum number because some combinations of true and false values for our L sentences will be logically inconsistent. We cannot, for example, imagine a world in which S1 is false, S2 is true and S1 ^ S2 is true. That is, some sets of the 2 L worlds might contain only impossible worlds. As an example, consider the sentences P, Q, P Q The consistent sets of truth values for these two sentences are given by the columns in the following table: Cont.. P Q PQ F F T F T T T F F T T T
In this case, there are four sets of possible worlds each
one corresponding to one of these four sets of truth values. 2. Probabilities on Proofs To dene probabilities on proofs, there are at least two different options. They correspond to the approaches taken in stochastic logic programs and PRISMs. . Consider therefore the stochastic context free grammar, where we have omitted many of the terminal symbols for brevity. This context-free grammar is stochastic in that every grammar rule has a probability associated to it FIRST ORDER PROBABILISTIC LOGICS
By now, everything is in place to introduce the key
representational frameworks that combine probabilistic reasoning with logical or relational representations. In the spirit of the above presentation, we can distinguish them according to whether they dene probabilities on interpretations or on proofs. Probabilistic Logical Models The rst class of representations extends Bayesian networks with abilities to dene probabilities on rst order logical or relational interpretations. Probabilistic logic programs Probabilistic logic programs are logic programs in which some of the facts are annotated with probabilities.