First-Order Logic PDF

First-Order Logic  Limitations of propositional logic Suppose you want to say “All humans are mortal” – In propositional logic, you would need ~6.7 billion statements Suppose you want to say “Some people can run a marathon” – You would need a disjunction of ~6.7 billion statements...

First-Order Logic  Limitations of propositional logic Suppose you want to say “All humans are mortal” – In propositional logic, you would need ~6.7 billion statements Suppose you want to say “Some people can run a marathon” – You would need a disjunction of ~6.7 billion statements  First-order logic Propositional logic assumes the world consists of atomic facts First-order logic assumes the world contains objects, relations, and functions  Syntax of FOL Constants: John, Sally, 2,... Variables: x, y, a, b,... Predicates: Person(John), Siblings(John, Sally), IsOdd(2),... Functions: MotherOf(John), Sqrt(x),... Connectives: , , , ,  Equality: = Quantifiers: ,  Term: Constant or Variable or Function(Term1,... , Termn) Atomic sentence: Predicate(Term1,... , Termn) or Term1 = Term2 Complex sentence: made from atomic sentences using connectives and quantifiers  Semantics of FOL Sentences are true with respect to a model and an interpretation Model contains objects (domain elements) and relations among them Interpretation specifies referents for constant symbols → objects predicate symbols → relations function symbols → functional relations An atomic sentence Predicate(Term1,... , Termn) is true iff the objects referred to by Term1,... , Termn are in the relation referred to by predicate  Universal quantification x P(x) Example: “Everyone at UNC is smart” x At(x,UNC)  Smart(x) Why not x At(x,UNC)  Smart(x)? Roughly speaking, equivalent to the conjunction of all possible instantiations of the variable: At(John, UNC)  Smart(John) ... At(Richard, UNC)  Smart(Richard) ... x P(x) is true in a model m iff P(x) is true with x being each possible object in the model  Existential quantification x P(x) Example: “Someone at UNC is smart” x At(x,UNC)  Smart(x) Why not x At(x,UNC)  Smart(x)? Roughly speaking, equivalent to the disjunction of all possible instantiations: [At(John,UNC)  Smart(John)]  [At(Richard,UNC)  Smart(Richard)]  … x P(x) is true in a model m iff P(x) is true with x being some possible object in the model  Properties of quantifiers x y is the same as y x x y is the same as y x x y is not the same as y x x y Loves(x,y) “There is a person who loves everyone” y x Loves(x,y) “Everyone is loved by at least one person” Quantifier duality: each quantifier can be expressed using the other with the help of negation x Likes(x,IceCream) x Likes(x,IceCream) x Likes(x,Broccoli) x Likes(x,Broccoli)  Equality Term1 = Term2 is true under a given model if and only if Term1 and Term2 refer to the same object E.g., definition of Sibling in terms of Parent: x,y Sibling(x,y)  [(x = y)  m,f  (m = f)  Parent(m,x)  Parent(f,x)  Parent(m,y)  Parent(f,y)] Using FOL: The Kinship Domain Brothers are siblings x,y Brother(x,y)  Sibling(x,y) “Sibling” is symmetric x,y Sibling(x,y)  Sibling(y,x) One's mother is one's female parent m,c (Mother(c) = m)  (Female(m)  Parent(m,c))  Why “First order”? FOL permits quantification over variables Higher order logics permit quantification over functions and predicates: P,x [P(x)  P(x)] x,y (x=y)  [P (P(x)P(y))]  Inference in FOL All rules of inference for propositional logic apply to first-order logic We just need to reduce FOL sentences to PL sentences by instantiating variables and removing quantifiers  Reduction of FOL to PL Suppose the KB contains the following: x King(x)  Greedy(x)  Evil(x) King(John) Greedy(John) Brother(Richard,John) How can we reduce this to PL? Let’s instantiate the universal sentence in all possible ways: King(John)  Greedy(John)  Evil(John) King(Richard)  Greedy(Richard)  Evil(Richard) King(John) Greedy(John) Brother(Richard,John) The KB is propositionalized – Proposition symbols are King(John), Greedy(John), Evil(John), King(Richard), etc.  Reduction of FOL to PL What about existential quantification, e.g., x Crown(x)  OnHead(x,John) ? Let’s instantiate the sentence with a new constant that doesn’t appear anywhere in the KB: Crown(C1)  OnHead(C1,John)  Substitution Substitution of variables by ground terms: SUBST({v/g},P) – Result of SUBST({x/Harry, y/Sally}, Loves(x,y)): Loves(Harry,Sally) – Result of SUBST({x/John}, King(x)  Greedy(x)  Evil(x)): King(John)  Greedy(John)  Evil(John)  Universal instantiation (UI) A universally quantified sentence entails every instantiation of it: v P(v) SUBST({v/g}, P(v)) for any variable v and ground term g E.g., x King(x)  Greedy(x)  Evil(x) yields: King(John)  Greedy(John)  Evil(John) King(Richard)  Greedy(Richard)  Evil(Richard) King(Father(John))  Greedy(Father(John))  Evil(Father(John))  Existential instantiation (EI) An existentially quantified sentence entails the instantiation of that sentence with a new constant: v P(v) SUBST({v/C}, P(v)) for any sentence P, variable v, and constant C that does not appear elsewhere in the knowledge base E.g., x Crown(x)  OnHead(x,John) yields: Crown(C1)  OnHead(C1,John) provided C1 is a new constant symbol, called a Skolem constant  Propositionalization Every FOL KB can be propositionalized so as to preserve entailment – A ground sentence is entailed by the new KB iff it is entailed by the original KB Idea: propositionalize KB and query, apply resolution, return result Problem: with function symbols, there are infinitely many ground terms – For example, Father(X) yields Father(John), Father(Father(John)), Father(Father(Father(John))), etc.  Propositionalization Theorem (Herbrand 1930): – If a sentence α is entailed by an FOL KB, it is entailed by a finite subset of the propositionalized KB Idea: For n = 0 to Infinity do – Create a propositional KB by instantiating with depth-n terms – See if α is entailed by this KB Problem: works if α is entailed, loops if α is not entailed Theorem (Turing 1936, Church 1936): – Entailment for FOL is semidecidable: algorithms exist that say yes to every entailed sentence, but no algorithm exists that also says no to every nonentailed sentence Review: FOL inference  Propositionalization Example KB: x King(x)  Greedy(x)  Evil(x) y Greedy(y) King(John) Brother(Richard,John) What will propositionalization produce? King(John)  Greedy(John)  Evil(John) King(Richard)  Greedy(Richard)  Evil(Richard) Greedy(John) Greedy(Richard) King(John) Brother(Richard,John) But what if all we want is to prove Evil(John)?  Generalized Modus Ponens (GMP) p1', p2', … , pn', (p1  p2  …  pn q) such that SUBST(θ, pi')= SUBST(θ, pi) for all i SUBST(θ,q) Used with definite clauses (exactly one positive literal) All variables assumed universally quantified Example: p1' is King(John) p1 is King(x) p2' is Greedy(y) p2 is Greedy(x) θ is {x/John,y/John} q is Evil(x) SUBST(θ,q) is Evil(John)  Unification UNIFY(α,β) = θ means that SUBST(θ, α) = SUBST(θ, β) p q θ Knows(John,x) Knows(John,Jane) {x/Jane} Knows(John,x) Knows(y,Mary) {x/Mary, y/John} Knows(John,x) Knows(y,Mother(y)) {y/John, x/Mother(John)} Knows(John,x) Knows(x,Mary) {x1/John, x2/Mary} Knows(John,x) Knows(y,z) {y/John, x/z} Standardizing apart eliminates overlap of variables Most general unifier  Inference with GMP p1', p2', … , pn', (p1  p2  …  pn q) such that SUBST(θ, pi')= SUBST(θ, pi) for all i SUBST(θ,q) Forward chaining – Like search: keep proving new things and adding them to the KB until we can prove q Backward chaining – Find p1, …, pn such that knowing them would prove q – Recursively try to prove p1, …, pn  Example knowledge base The law says that it is a crime for an American to sell weapons to hostile nations. The country Nono, an enemy of America, has some missiles, and all of its missiles were sold to it by Colonel West, who is American. Prove that Col. West is a criminal  Example knowledge base It is a crime for an American to sell weapons to hostile nations: American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Nono has some missiles x Owns(Nono,x)  Missile(x) Owns(Nono,M1)  Missile(M1) All of its missiles were sold to it by Colonel West Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missiles are weapons: Missile(x)  Weapon(x) An enemy of America counts as “hostile”: Enemy(x,America)  Hostile(x) West is American American(West) The country Nono is an enemy of America Enemy(Nono,America)  Forward chaining proof American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America)  Forward chaining proof American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America)  Forward chaining proof American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America) Backward chaining example American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America) Backward chaining example American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America) Backward chaining example American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America) Backward chaining example American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America) Backward chaining example American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America) Backward chaining example American(x)  Weapon(y)  Sells(x,y,z)  Hostile(z)  Criminal(x) Owns(Nono,M1)  Missile(M1) Missile(x)  Owns(Nono,x)  Sells(West,x,Nono) Missile(x)  Weapon(x) Enemy(x,America)  Hostile(x) American(West) Enemy(Nono,America) Backward chaining algorithm  Resolution: FOL version p1  ···  pk, q1  ···  qn such that UNIFY(pi, qj) = θ SUBST(θ, p1  ···  pi-1  pi+1  ···  pk  q1  ···  qj-1  qj+1  ···  qn) For example, Rich(x)  Unhappy(x) Rich(Ken) Unhappy(Ken) with θ = {x/Ken} Apply resolution steps to CNF(KB  α); complete for FOL Resolution proof: definite clauses