Transcript Data Modeling - Hiram College

Logic
CPSC 386 Artificial Intelligence
Ellen Walker
Hiram College
Knowledge Based Agents
• Representation
– How is the knowledge stored by the agent?
– Procedural vs. Declarative
• Reasoning
– How is the knowledge used…
• To solve a problem?
• To generate more knowledge?
• Generic Functions
– TELL (add a fact to the knowledge base)
– ASK (get next action based on info in KB)
Hunt the Wumpus
• Game is played in a MxN grid
• One player, one wumpus, one or more pits
• Goal: find gold while avoiding wumpus and
pits
• Percepts:
– Glitter (gold is in this square)
– Stench (wumpus is within 1 square N,E,S,W)
– Breeze (pit is within 1 square N, E, S, W)
Wumpus Example
[start]
stench
[Wumpus]
Glitter
[gold]
stench,
breeze
breeze
[Pit]
stench
breeze
Examples of reasoning
• If the player is in square (1, 0) and the percept is
breeze, then there must be a pit in (0,0) or a pit in
(2,0) or a pit in (1,1).
• If the player is in (0,0) [and still alive], there is not a
pit in (0,0).
• If there is no breeze in (0,0), there is no pit in (0,1)
• If there is also no breeze in (0,1) then there is no pit
in (1,1).
• Therefore, there must be a pit in (2,0)
Formalizing Reasoning
• Information is represented in sentences, which must
have correct syntax
( 1 + 2 ) * 7 = 21 vs. 2 ) + 7 = * ( 1 21
• The meaning of a sentence (semantics) defines its
truth in each model (possible world)
• One sentence entails another sentence if the second
one follows logically from the first; i.e. every model
that has the first true, also has the second true.
• Inference is the process of deriving a specific
sentence from a KB (where the sentence must be
entailed by the KB)
Desirable properties of an inference
algorithm
• Soundness
– Only sentences that are entailed by a KB will be derived by
the inference algorithm
• Completeness
– Every sentence that is entailed by a KB will be derived by
the inference algorithm (eventually)
• If these properties are true, then every sentence
derived from a true KB will be true in the world.
– All reasoning can be done in a model, not the world!
Propositional Logic
• Syntax: Sentence ->
true, false, P, Q, R …
sentence
( sentence  sentence )
( sentence  sentence )
( sentence sentence )
( sentence  sentence )
( sentence )
primitive sent.
not
and
or
implies (if)
if & only if (iff)
• Note: propositional logic can be directly implemented
in hardware using logic gates for operations
Propositional Logic Sentences
• If there is a pit at [1,1], there is a breeze at
[1,0]
P11  B10
• There is a breeze at [2,2], if and only if there
is a pit in the neighborhood
B22  ( P21  P23  P12  P32 )
• There is no breeze at [2,2]
B22
Propositional Logic Inference
• Question: Does KB entail S?
• Method 1: Truth Table Entailment
– Construct a truth table whose columns are all propositions
used in the sentences in KB.
– If S is true everywhere all sentences in KB are true, then KB
entails S (otherwise not)
• Method 2: Proof
– Proof by deduction
– Proof by contradiction
– Etc.
Truth Table Entailment
A
F
F
F
F
T
T
T
T
B
F
F
T
T
F
F
T
T
C
F
T
F
T
F
T
F
T
AB
F
F
F
F
F
F
T
T
AC
F
F
F
F
F
T
F
T
BC
F
F
F
T
F
F
F
T
A^C, C
does not
entail
BC
A,B,
Entails
AB
Truth Table Entailment is…
• Sound
– by definition, since it directly implements the
definition of entailment
• Complete
– Only when the KB (and therefore the truth table) is
finite
• Time consuming
– The truth table size is 2number of statements , and we
have to check every row!
Rules for Deductive Proofs
• Modus Ponens
– Given: S1  S2 and S1, derive S2
• And-elimination
– Given: S1  S2, derive S1
– Given: S1  S2, derive S2
• DeMorgan’s Law
– Given ( A  B) derive A  B
– Given ( A  B) derive A  B
• More on p. 210
Example Proof by Deduction
• Knowledge
S1: B22  ( P21  P23  P12  P32 )
S2: B22
rule
observation
• Inferences
S3: (B22  (P21  P23  P12  P32 ))
((P21  P23  P12  P32 )  B22) [S1,bi elim]
S4: ((P21  P23  P12  P32 )  B22)
[S3, and elim]
S5: (B22  ( P21  P23  P12  P32 )) [contrapos]
S6: (P21  P23  P12  P32 )
[S2,S6, MP]
S7: P21  P23  P12  P32
[S6, DeMorg]
Evaluation of Deductive Inference
(using p. 110 rules)
• Sound
– Yes, because the inference rules themselves are sound.
(This can be proven using a truth table argument).
• Complete
– If we allow all possible inference rules, we’re searching in an
infinite space, hence not complete
– If we limit inference rules, we run the risk of leaving out the
necessary one…
• Monotonic
– If we have a proof, adding information to the DB will not
invalidate the proof
Resolution
• Resolution allows a complete inference
mechanism (search-based) using only one
rule of inference
• Resolution rule:
– Given: P1  P2  P3 … Pn, and P1  Q1 … Qm
– Conclude: P2  P3 … Pn  Q1 … Qm
Complementary literals P1 and P1 “cancel out”
• Why it works:
– Consider 2 cases: P1 is true, and P1 is false
Resolution in Wumpus World
• There is a pit at 2,1 or 2,3 or 1,2 or 3,2
– P21  P23  P12  P32
• There is no pit at 2,1
– P21
• Therefore (by resolution) the pit must be at
2,3 or 1,2 or 3,2
– P23  P12  P32
Proof using Resolution
• To prove a fact P, repeatedly apply resolution until
either:
– No new clauses can be added, (KB does not entail P)
– The empty clause is derived (KB does entail P)
• This is proof by contradiction: if we prove that KB 
P derives a contradiction (empty clause) and we
know KB is true, then P must be false, so P must be
true!
• To apply resolution mechanically, facts need to be in
Conjunctive Normal Form (CNF)
• To carry out the proof, need a search mechanism that
will enumerate all possible resolutions.
Conjunctive Normal Form for
B22  ( P21  P23  P12  P32 )
1. Eliminate  , replacing with two implications
(B22  ( P21  P23  P12  P32 ))  ((P21  P23  P12  P32 )  B22)
2. Replace implication (A  B) by A  B
(B22  ( P21  P23  P12  P32 ))  ((P21  P23  P12  P32 ) 
B22)
3. Move  “inwards” (unnecessary parens removed)
(B22  P21  P23  P12  P32 )  ( (P21  P23  P12  P32 )
 B22)
4. Distributive Law
(B22  P21  P23  P12  P32 )  (P21  B22)  (P23  B22) 
(P12  B22)  (P32  B22)
(Final result has 5 clauses)
Resolution Example
• Given B22 and P21 and P23 and P32 ,
prove P12
• (B22  P21  P23  P12  P32 ) ; P12
• (B22  P21  P23  P32 ) ; P21
• (B22  P23  P32 ) ; P23
• (B22  P32 ) ; P32
• (B22) ; B22
• [empty clause]
Mechanical Approach to Resolution
• Use DFS in “clause space”
– Initial state is Not (whatever is being proven)
– Goal state is Empty clause
– Next state generator finds all clauses in the KB
[that include negations of one or more
propositions in the current state]; next states are
resolutions of those clauses with current state
• Like all DFS, this is worst-case exponential
(search all possible clauses)
Evaluation of Resolution
• Resolution is sound
– Becase the resolution rule is true in all cases
• Resolution is complete
– Provided a complete search method is used to find the proof,
if a proof can be found it will
– Note: you must know what you’re trying to prove in order to
prove it!
• Resolution is exponential
– The number of clauses that we must search grows
exponentially…
– If it didn’t we could use resolution to solve 3SAT in
polynomial time!
Horn Clauses
• A Horn Clause is a CNF clause with exactly one
positive literal
–
–
–
–
–
The positive literal is called the head
The negative literals are called the body
Prolog: head:- body1, body2, body3 …
English: “To prove the head, prove body1, …”
Implication: If (body1, body2 …) then head
• Horn Clauses form the basis of forward and
backward chaining
• The Prolog language is based on Horn Clauses
• Deciding entailment with Horn Clauses is linear in the
size of the knowledge base
Reasoning with Horn Clauses
• Forward Chaining
– For each new piece of data, generate all new
facts, until the desired fact is generated
– Data-directed reasoning
• Backward Chaining
– To prove the goal, find a clause that contains the
goal as its head, and prove the body recursively
– (Backtrack when you chose the wrong clause)
– Goal-directed reasoning
First Order Logic (ch. 8)
• Has a greater expressive power than
propositional logic
– We no longer need a separate rule for each
square to say which other squares are breezy/pits
• Allows for facts, objects, and relations
– In programming terms, allows classes, functions
and variables