Intelligent Systems

Intelligent Systems Propositional Logic – Lecture 2 Prof. Dieter Fensel (& James Scicluna)

Agenda • Motivation • Propositional Logic • Inference by Enumeration • Validity and Satisfiability • Logical Equivalence • Proof Methods – Resolution • Soundness and Completeness • Horn Clauses • Forward Chaining • Backward Chaining • Efficient Propositional Inference – The DPLL Algorithm • Limitations of Propositional Logic • Summary

Motivation – Why Logics? • Origins of Logic • Ancient Greece • Philosophers argued about themes, but who was right in the end? • When Socrates claimed that one statement followed from another, was it actually true? • Logics today • Verification (e.g. Circuits) • Proofs • Robotics • ...others • Logics are formal languages for representing information such that conclusionscan be drawn. • Syntax defines the sentences in the language. • Semantics define the "meaning" of sentences, i.e., defines the truth of each sentence w.r.t. each possible world (model)

Motivation – Basic Ideas • Entailment means that one thing follows from another KB╞α • Knowledge base KB entails sentence α if and only if α is true in all worlds where KB is true • E.g., the KB containing “Innsbruck is in Tirol” and “Tirol is in Austria” entails “Innsbruck is in Austria” • E.g., the KB containing x+y = 4 entails 4 = x+y • Entailment is a relationship between sentences (i.e., syntax) that is based on semantics

Motivation – Models/Interpretations • Possible worlds – environments in which the agent might or might not be in • Models – mathematical abstractions, formally structured worlds with respect to which truth of sentences can be evaluated • m is a model of a sentence α if α is true in m • M(α)is the set of all models of α • KB ╞ α iffM(KB)  M(α) • Knowledge base KB entails sentence α if and only if all models of the KB are models of α

Motivation – Inference • KB ├iα : sentence α can be derived from KB by procedure i, i proves α • Soundness: i is sound if whenever KB ├iα, it is also true that KB╞ α • Completeness: i is complete if whenever KB╞ α, it is also true that KB ├iα • A sound and complete procedure answers any question whose answer follows from what is known by the KB correctly.

Solution – Propositional Logic Syntax • Propositional logic is the simplest logic and illustrates basic ideas of logic • The proposition symbols P1, P2,…are sentences (formulae) • If S is a sentence, S is a sentence (negation) • If S1 and S2 are sentences, S1 S2 is a sentence (conjunction) • If S1 and S2 are sentences, S1 S2 is a sentence (disjunction) • If S1 and S2 are sentences, S1 S2 is a sentence (implication) • If S1 and S2 are sentences, S1 S2 is a sentence (biconditional)

Solution – Propositional Logic Semantics • Each model specifies true/false values for each proposition symbol (8 possible models) • E.g. P1,2 P2,2 P3,1 false true false • Rules for evaluating truth with respect to an interpretation m: S is true iff S is false S1 S2 is true iff S1 is true and S2 is true S1 S2 is true iff S1 is true or S2 is true S1 S2 is true iff S1 is false or S2 is true S1 S2 is false iff S1 is true and S2 is false S1 S2 is true iff S1S2 is true and S2S1 is true • Simple recursive process evaluates an arbitrary sentence w.r.t. an interpretation, e.g., P1,2  (P2,2 P3,1) : true  (true  false) = true  true = true

Solution – Propositional Logic Truth Tables

Solution – Propositional Logic α1 = "[1,2] is safe"

Procedure - Inference by Enumeration • Depth-first enumeration of all models is sound and complete • PL-TRUE evaluates a sentence recursively wrt. to an interpretation • EXTEND(s,v,m) extends the partial model m by assigning value v to symbol s. • For n symbols, time complexity is O(2n), space complexity is O(n)

Definition - Validity and Satisfiability A sentence is valid if it is true in all models, e.g., True, A A, A  A, (A  (A  B))  B Validity is connected to inference via the Deduction Theorem: KB ╞ α if and only if (KB α) is valid A sentence is satisfiable if it is true in some model e.g., A B, C A sentence is unsatisfiable if it is true in no models e.g., A A Satisfibility is connected to inference via the following KB ╞ α if and only if (KBα) is unsatisfiable

Definition - Logical equivalence • Two sentences are logically equivalent iff they are true in same models: α ≡ ß iff α╞ β and β╞ α

Procedure - Proof Methods • Proof methods divide into (roughly) two kinds • Application of inference rules • Legitimate (sound) generation of new sentences from old ones • Proof = a sequence of inference rule applications • Can use inference rules as operators in a standard search algorithm • Typically requires transformation of sentences into a normal form • Model checking • Truth table enumeration (always exponential in n) • Improved backtracking, e.g., Davis-Putnam-Logemann-Loveland (DPLL) • Heuristic search in model space (sound but incomplete) e.g., min-conflicts-like hill-climbing algorithms

Procedure - Resolution Conjunctive Normal Form (CNF) conjunction of disjunctions of literals often called clauses E.g., (A B)  (B C D) • Resolution inference rule (for CNF): li… lk, m1 … mn li … li-1 li+1  … lkm1 … mj-1 mj+1... mn where li and mj are complementary literals. E.g., P1,3P2,2, P2,2 P1,3 • Resolution is sound and complete for propositional logic

Procedure - Conversion to CNF B1,1 (P1,2 P2,1) • Eliminate , replacing α  β with (α  β)  (β  α). (B1,1 (P1,2 P2,1))  ((P1,2 P2,1)  B1,1) • Eliminate , replacing α  β with α β. (B1,1 P1,2 P2,1)  ((P1,2 P2,1)  B1,1) • Move  inwards using de Morgan's rules: (B1,1  P1,2 P2,1)  ((P1,2 P2,1)  B1,1) • Apply distributivity law ( over ) and flatten: (B1,1 P1,2 P2,1)  (P1,2  B1,1)  (P2,1 B1,1)

Definition - Soundness Whenever KB ├r α, it is also true that KB╞ α li… lk, m1 … mn li … li-1 li+1  … lkm1 … mj-1 mj+1... mn li and mj are complementary literals. (li … li-1 li+1  … lk)  li mj  (m1 … mj-1 mj+1... mn) (li … li-1 li+1  … lk)  (m1 … mj-1 mj+1... mn) using the rule (a   b)  (b  a) which can be written also as (a  b)  ( a  b).

Definition - Completeness • Whenever KB╞ α, it is also true that KB ├rα • KB ╞ α if and only if (KBα) is unsatisfiable • Ground resolution theorem • If a set of clauses is unsatisfiable, then the resolution closure of those clauses contains the empty clause. • Proof by contradiction • AssumeKB ╞ α and that the closure does not contain the empty clause.

Procedure - Forward and backward chaining • Horn clauses: disjunction of literals of which at most one is positive • Important because Horn clauses can be written as an implication whose premise is a conjuction of positive literals and whose conclusion is a single positive literal • Definite clauses: exactly one positive literal • Positive literal forms the head • Negative literals form the body • Inference with Horn clauses can be done by forward or backward chaining in a time that is linear in the size of the KB.

Procedure - Forward chaining • Fire any rule whose premises are satisfied in the KB • Add its conclusion to the KB, until query is found

Procedure - Forward chaining algorithm • agenda: true symbols not yet processed • count: how many premises are unknown • for each new symbol processed, count is reduced by one for each implication in whose premise this symbol appears • Count = 0  conclusion can be added to the agenda • Forward chaining is sound and complete for Horn KB

Example - Forward Chaining

Procedure - Backward chaining • Work backwards from the query q • To prove q by BC, check if q is known already, or prove by BC all premises of some rule concluding q • Avoid loops: check if new sub-goal is already on the goal stack • Avoid repeated work: check if new sub-goal has already been proved true, or has already failed

Example – Backward Chaining

Explanation: Forward vs. Backward chaining • FC is data-driven, automatic, unconscious processing, • e.g., object recognition, routine decisions • May do lots of work that is irrelevant to the goal !!! • BC is goal-driven, appropriate for problem-solving, • e.g., Where are my keys? How do I get into a PhD program? • Complexity of BC can be much less than linear in size of KB (only relevant facts are considered)

Procedure: Efficient propositional inference • Two families of efficient algorithms for propositional inference • Complete backtracking search algorithms • DPLL algorithm (Davis, Putnam, Logemann, Loveland) • Incomplete local search algorithms • WalkSAT algorithm (Incomplete, local search algorithm)

Procedure: The DPLL algorithm Determine if an input propositional logic sentence (in CNF) is satisfiable. Improvements over truth table enumeration • Early termination • A clause is true if any literal is true (disjunction of literals) • A sentence is false if any clause is false (conjunction of clauses) • Pure symbol heuristic • Pure symbol: always appears with the same "sign" in all clauses. • e.g., In the three clauses (A B), (B C), (C  A), A and B are pure, C is impure. • If a sentence has a model then it has a model with the pure symbols assigned so as to make their literals true. • Unit clause heuristic • Unit clause: only one literal in the clause • The only literal in a unit clause must be true.

Procedure: The DPLL algorithm

KB contains “physics” sentences for every single square For every time t and every location [x,y], Rapid proliferation of clauses Limitation of propositional logic

Summary • Logical agents apply inference to a knowledge base to derive new information and make decisions • Basic concepts of logic: • syntax: formal structure of sentences • semantics: truth of sentences w.r.t. models • entailment: necessary truth of one sentence given another • inference: deriving sentences from other sentences • soundness: derivations produce only entailed sentences • completeness: derivations can produce all entailed sentences • Resolution is complete for propositional logic • Forward, backward chaining are linear-time, complete for Horn clauses • Propositional logic lacks expressive power

References • S. Russell, P. Norvig: Artificial Intelligence – A Modern Approach

Intelligent Systems