23.02.2015 Views

Machine Learning - DISCo

Machine Learning - DISCo

Machine Learning - DISCo

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

It is easiest to introduce the resolution rule in propositional form, though it is<br />

readily extended to first-order representations. Let L be an arbitrary propositional<br />

literal, and let P and R be arbitrary propositional clauses. The resolution rule is<br />

P V L<br />

-L v R<br />

P V R<br />

which should be read as follows: Given the two clauses above the line, conclude<br />

the clause below the line. Intuitively, the resolution rule is quite sensible. Given<br />

the two assertions P v L and -L v R, it is obvious that either L or -L must be<br />

false. Therefore, either P or R must be true. Thus, the conclusion P v R of the<br />

resolution rule is intuitively satisfying.<br />

The general form of the propositional resolution operator is described in<br />

Table 10.5. Given two clauses C1 and C2, the resolution operator first identifies<br />

a literal L that occurs as a positive literal in one of these two clauses and as<br />

a negative literal in the other. It then draws the conclusion given by the above<br />

formula. For example, consider the application of the resolution operator illustrated<br />

on the left side of Figure 10.2. Given clauses C1 and C2, the first step of the<br />

procedure identifies the literal L = -KnowMaterial, which is present in C1, and<br />

whose negation -(-KnowMaterial) = KnowMaterial is present in C2. Thus the<br />

conclusion is the clause formed by the union of the literals C1- (L} = Pass Exam<br />

and C2 - (-L} = -Study. As another example, the result of applying the resolution<br />

rule to the clauses C1 = A v B v C v -D and C2 = -B v E v F is the clause<br />

AvCV-DvEvF.<br />

It is easy to invert the resolution operator to form an inverse entailment<br />

operator O(C, C1) that performs inductive inference. In general, the inverse entailment<br />

operator must derive one of the initial clauses, C2, given the resolvent C<br />

and the other initial clause C1. Consider an example in which we are given the<br />

resolvent C = A v B and the initial clause C1 = B v D. How can we derive a<br />

clause C2 such that C1 A C2 F C? First, note that by the definition of the resolution<br />

operator, any literal that occurs in C but not in C1 must have been present in C2.<br />

In our example, this indicates that C2 must contain the literal A. Second, the literal<br />

1. Given initial clauses C1 and C2, find a literal L from clause C1 such that -L occurs in clause C2.<br />

2. Form the resolvent C by including all literals from C1 and C2, except for L and -L. More<br />

precisely, the set of literals occurring in the conclusion C is<br />

where u denotes set union, and "-" denotes set difference.<br />

TABLE 10.5<br />

Resolution operator (propositional form). Given clauses C1 and C2, the resolution operator constructs<br />

a clause C such that C1 A C2 k C.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!