ITCS 3153 Artificial Intelligence
Lecture 13 First-Order Logic
Chapter 8
First-order logic
We saw how propositional logic can create intelligent behavior But propositional logic is a poor representation for complex environments
Why?
First-order logic is a more expressive and powerful representation
What do we like about propositional logic?
It is:
Declarative
Relationships between variables are described
A method for propagating relationships Expressive Can represent partial information using disjunction Compositional
If A means foo and B means bar, A ^ B means foo and bar
What dont we like about propositional logic?
Lacks expressive power to describe the environment concisely
Separate rules for every square/square relationship in Wumpus world
Natural Language
English appears to be expressive
Squares adjacent to pits are breezy
But natural language is a medium of communication, not a knowledge representation
Much of the information and logic conveyed by language is dependent on context
Information exchange is not well defined Not compositional (combining sentences may mean something different) It is ambiguous
But we borrow representational ideas from natural language
Natural language syntax
Nouns and noun phrases refer to objects
People, houses, cars
Verbs and verb phrases refer to relationships btw objects Red, round, nearby, eaten
Some relationships are clearly defined functions where there is only one output for a given input
Best friend, first thing, plus
We build first order logic around objects and relations
Ontology
a specification of a conceptualization A description of the objects and relationships that can exist
Propositional logic had only true/false relationships
First-order logic has many more relationships The ontological commitment of languages is different How much can you infer from what you know? Temporal logic defines additional ontological commitments because of timing constraints
Higher-order logic
First-order logic is first because you relate objects (the first-order entities that actually exist in the world)
There are 10 chickens [Link]=10 There are 10 ducks [Link]=10
You cannot build relationships between relations or functions
There are as many chickens as ducks [Link] = [Link] the number of objects belonging to a group must be a property of the group, and not the objects themselves Cannot represent Leibnizs law: If x and y share all properties, x is y
Another characterization of a logic
Epistemological commitments
The possible states of knowledge permitted with respect to each fact In first-order logic, each sentence is a statement that is True, false, or unknown
Formal structure of first-order logic
Models of first-order logic contain:
A set of objects (its domain) Alice, Alices left arm, Bob, Bobs hat Relationships between objects Represented as tuples Sibling (Alice, Bob), Sibling (Bob, Alice) On head (Bob, hat) Person (Bob), Person (Alice) Some relationships are functions if a given object is related to exactly one object in a certain way Alice -> Alices left arm
First-order logic syntax
Constant Symbols
A, B, Bob, Alice, Hat
Predicate Symbols
Is, onHead, hasColor, person
Function Symbols
Mother, leftLeg
Each predicate and function symbol has an arity
A constant the fixes the number of arguments
First-order logic syntax
Names of things are abitrary
Knowledge base adds meaning
Number of possible domain elements is unbounded
Number of models is unbounded Checking enumeration by entailment is impossible
Syntax
Term
A logical expression that refers to an object
Constants
We could assign names to all objects, like providing a name for every shoe in your closet Function symbols Used in place of a constant symbol OnLeftFoot(John))
Atomic Sentences
Formed by a predicate symbol followed by parenthesized list of terms
Sibling (Alice, Bob) Married (Father(Alice), Mother(Bob))
An atomic sentence is true in a given model, under a given interpretation, if the relation referred to by the predicate symbol holds among the objects referred to by the arguments
Complex sentences
We can use logical connectives
~Sibling(LeftLeg(Alice), Bob)
Sibling(Alice, Bob) ^ Sibling (Bob, Alice)
Quantifiers
A way to express properties of entire collections of objects
Universal quantification (forall, )
The power of first-order logic Forallx King(x) => Person(x)
x : King ( x) Person( x)
x is a variable
Universal Quantification
Forall x, P
P is true for every object x
Forall x, King(x) => Person(x)
Richard the Lionheart King John Richards left leg Johns left leg
The crown
Universal Quantification
Note that all of these are true
Implication is true if premise is false Using AND instead of implication is overly strong
By asserting a universally quantified sentence, you assert a whole list of individual implications
Existential Quantification
There exists,
There exists an x such that Crown(x) ^ OnHead(x, John)
It is true for at least one object
AND, ^, is the appropriate connective
Existential Quantification
What if we used implication as the connective?
Implication is true if both premise and conclusion are true or if premise is false
Richard the Lionheart is not a crown, first assertion is true, and existential is satisfied
Nested Quantifiers
Building more complex sentences
Everybody loves somebody
There is someone who is loved by everyone
Use unique variable names and parentheses when appropriate
Combining
Everyone who dislikes parsnips == there does not exist someone who likes parsnips
Everyone likes ice cream == there is no one who does not like ice cream
Combining
De Morgans rules apply
Equality
Two terms refer to the same object
Father (John) = Henry Richard has at least two brothers Notice this sentence is not the same
An Example
A Tell/Ask interface for a first-order knowledge base
Sentences are added with Tell
These are called assertions
Tell (KB, King(John)) Tell (KB, Forall x: King(x) => Person(x))
Queries are made with Ask
Ask (KB, King(John)) Ask (KB, Person(John))
An Example
Quantified queries Ask (KB, exist x: Person(x))
KB should return a list of variable/term pairs that satisfy the query
The Wumpus World
More precise axioms than with propositional logic
Percept has five values
Time is important
A typical sentence Percept ([Stench, Breeze, Glitter, None, None], 5)
The actions are terms
Turn(right), Turn(left), Forward, Shoot, Grab, Release Computing best action with a query
Exist a: BestAction(a, 5)
The Wumpus World
After executing query, KB responds with variable/term list: {a/Grab}
Then tell the KB the action taken
Raw percept data is easily encoded Reflexes are easily encoded
Wumpus World
Defining the environment with[x,y] reference instead of alternative atomic name
Adjacency between two squares
Location of Wumpus is constant: Home (Wumpus) Location of agent changes: At (Agent, [ ], t)
Diagnostic Rules
Rules leading from observed effects to hidden causes
Breezy implies pits
Not breezy implies no pits Combining
Causal Rules
Some hidden property causes percepts to be generated
A pit causes adjacent squares to be breezy
If all squares adjacent to a square a pitless, it will not be breezy
Causal Rules
The causal rules formulate a model
Knowledge of how the environment operates
Model can be very useful and important and replace straightforward diagnostic approaches
Conclusion
If the axioms correctly and completely describe the way the world works and the way percepts are produced,
then any complete logical inference procedure will infer the strongest possible description of the world state given the available percepts The agent designer can focus on getting the knowledge right without worrying about the processes of deduction
