FUNDAMENTALS OF

ARTIFICIAL INTELLIGENCE

CHAPTER FOUR –KNOWLEDGE AND

REASONING
2 CONTENTS

 Introduction
 Logical Agents
 Propositional Logic
 Predicate (First-Order) Logic
 Inference in First-Order Logic
 Knowledge Representation
 Knowledge-based Systems
3 INTRODUCTION

 A logical, knowledge-based agent begins with some knowledge of the world and of its own
actions.
 It uses logical reasoning to maintain a description of the world as new percepts arrive, and to
deduce a course of action that will achieve its goals.
 The knowledge-based approach is a particularly powerful way of constructing an agent
program.
 It aims to implement a view of agents in which they can be seen as knowing about their world,
and reasoning about their possible courses of action.
4 INTRODUCTION

 Knowledge-based agents are able to accept new tasks in the form of explicitly described goals;
 They can achieve competence quickly by being told or learning new knowledge about the
environment;
 And they can adapt to changes in the environment by updating the relevant knowledge.
 A knowledge-based agent needs to know many things:
 the current state of the world;
 how to infer unseen properties of the world from percepts;
 how the world evolves over time; what it wants to achieve;
 and what its own actions do in various circumstances.
5 A KNOWLEDGE-BASED AGENT

 Knowledge-based agents are agents that know about their world and reason about their courses
of action.
 A knowledge-based agent consists of two parts:
 A knowledge base and
 An inference engine

Knowledge base  domain-specific content

Inference engine  domain-independent algorithms

 The central component of a knowledge-based agent is its knowledge base, or KB.

 Informally, a knowledge base is a set of representations of facts about the world.
 Each individual representation is called a sentence.
6 A KNOWLEDGE-BASED AGENT

 Inference Engine (IE) = reasons with knowledge to produce new knowledge.

 There must be a way to add new sentences to the knowledge base, and a way to query what is
known.
 How to interact with KB?
 Tell it what it needs to know.
 Ask itself what to do - answers should follow from the KB.
 Each time the agent program is called, it does two things.
 First, it TELLS the knowledge base what it perceives.
 Second, it ASKS the knowledge base what action it should perform.
7 A KNOWLEDGE-BASED AGENT

 The sentences in the KB must therefore contain a range of information.

 The agent must be able to:
 Represent and update states of the environment,
 Represent its own actions, Incorporate new precepts into the KB,
 Update internal representations of the world,
 Deduce hidden properties of the environment and
 Deduce what the ‘right’ action is at any given time.
 In addition, the KB must contain the goal(s) of the agent.
8 A KNOWLEDGE-BASED AGENT

 Entailment means that one thing follows from another:

 One sentence follows from another.
 The symbol for entailment is ‘╞’. 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 “Leqa won” and “Gimbi won” entails “Either Leqa won or Gimbi
won”
 E.g., x+y=4 entails 4=x+y
 Entailment is a relationship between sentences (i.e., syntax) that is based on semantics.
9 A KNOWLEDGE-BASED AGENT

 We said that it is possible to understand the operation of a knowledge-based agent in terms of

what it knows.
 It is possible to construct a knowledge-based agent by TELLing it what it needs to know.
 The agent's initial program, before it starts to receive percepts, is built by adding one by one
the sentences that represent the designer's knowledge of the environment.
 Provided that the representation language makes it easy to express this knowledge in the form
of sentences, this simplifies the construction problem enormously.
 This is called the declarative approach to system building.
 Also, one can design learning mechanisms that output general knowledge about the
environment given a series of percepts.
10 REPRESENTATION, REASONING, AND LOGIC

 The object of knowledge representation is to express knowledge in computer-tractable form,

such that it can be used to help agents perform well.
 A knowledge representation language is defined by two aspects:
 The syntax:
 describes the possible configurations that can constitute sentences.
 describe syntax in terms of how sentences are represented on the printed page, but the real
representation is inside the computer:
 The semantics:
 determines the facts in the world to which the sentences refer.
11 REPRESENTATION, REASONING, AND LOGIC

 We want to generate new sentences that are necessarily true, given that the old sentences are
true.
 This relation between sentences is called entailment.
 An inference procedure that generates only entailed sentences is called sound or truth-
preserving.
 Once a sentence is given an interpretation by the semantics, the sentence says that the world is
this way and not that way.
 Note that truth depends both on the interpretation of the sentence and on the actual state of the
world.
 The terms "reasoning" and "inference" are generally used to cover any process by which
conclusions are reached.
12 REPRESENTATION, REASONING, AND LOGIC

 A logic is deduction or derivation of new true statements from assumptions.

 Syntax defines the sentences in the language.
 Define formal structure of sentences.
 Semantics define the "meaning" of sentences;
 Define truth of a sentence in a world.
 E.g., the language of arithmetic
 x+2 ≥ y is a sentence; x2+y > {} is not a sentence
 x+2 ≥ y is true if the number x+2 is not less than the number y
13 PROPOSITIONAL LOGIC

 Propositional logic- Ways of combining or altering statements or propositions to form more

complicated statements.
 Propositional logic (also known as Boolean logic):
 Consists

 symbols P1, P2

 connectives, ,,, , which can be used to combine propositions into

more complex sentences.
The syntax of propositional logic can be summarized as follows:
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)
14 PROPOSITIONAL LOGIC

 Truth tables for connectives

15 PROPOSITIONAL LOGIC

 Two sentences are logically equivalent if they are true in the same models: α ≡ ß if α╞ β and
β╞ α.
16 PROPOSITIONAL LOGIC

 A sentence is valid if it is true in all models

 e.g., True A A, A  A,
 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., e.g., A,A B,

 A sentence is unsatisfiable if it is true in no models

e.g.,AA, can never be true;

17 PROOF BY APPLYING INFERENCE RULES

 Instead of enumerating truth tables, can do proof by successively applying sound inference
rules:
 Inference is the processing of finding what sentences are entailed.
 inference: deriving sentences from other sentences.
 Modus Ponens:

 Modus Tollens: The Modus Tollens rule state that if P→ Q is true and ¬ Q is true, then ¬ P
will also true. It can be represented as:
18 PROOF BY APPLYING INFERENCE RULES

 Resolution is a sound and complete inference algorithm:

 whenever KB ⊧ α, it is also true that KB ⊢iα, Meaning "a is derived from KB by i" or “i
derives a from KB.“
 The Resolution rule state that if P ∨Q and ¬ P ∧R is true, then Q ∨R will also be true. It can be
represented as

 Resolution works best when the formula is of the special form: it is an ∧ of ∨s of (possibly
negated, ¬) variables (called literals ).

Uses Conjunctive Normal Form (CNF) conjunction of disjunctions of
literals, e.g., (A  B)  (B  C  D)
 Conjunctive normal form (CNF) is an important normal form for propositional logic. A logic
formula is in conjunctive normal form if it is a single conjunction of disjunctions of (possibly
negated) literals.
19 PROOF BY APPLYING INFERENCE RULES

 Conversion to CNF.
 B  (P  Q)
 Eliminate , replacing A  B with (A  B)  A)
 B  (P  Q) (P  Q)  B
 Eliminate , replacing A  B with  B
  (P  Q)) (P  Q)  B)
 Move inwards using de Morgan's rules and double-negation
 ( P  Q) P Q)  B)
 Apply distributive law ( over ):
 ( P  Q) P  B) Q  B) IN CNF.
 TRY This : (A∨B)  (C  D)