Download presentation
Presentation is loading. Please wait.
1
1 Lecture 34 Introduction to Knowledge Representation & Expert Systems Overview Lecture Objective Introduction to Knowledge Representation Knowledge Representation Languages Introduction to Expert Systems Preview: Brief Introduction the other AI Areas
2
2 Lecture 34 Lecture Objective Present an overview of Knowledge Representation and what is entails Outline the different knowledge representation languages Motivate, discuss, exemplify pointing out the limitations of expert systems
3
3 Lecture 34 Introduction to Knowledge Representation AI is based on the computer’s ability to learn and to improve performance, based on past errors. This exists to some extent in all areas of AI. Two key components are needed to make machine intelligent, 1. Knowledge Representation: How to store knowledge and relationships... knowledge base - set of facts and rules Inference: How to make new facts from already available knowledge. inference engine - computer that accesses, selects and interprets rules to make new facts E.g., Fact: Ali is shorter than Ahmad Rule: If X is shorter than Y, then Y is taller than X
4
4 Lecture 34 Knowledge Representation To represent knowledge we need a representation language. Each representation language has syntactic and semantic conventions that makes it possible to describe things. The syntax of a representation specifies the symbols that maybe used and the way this symbols may be arranged or used. The semantics of a representation specifies how meaning is embodied in the symbols. Examples of Knowledge Representation Languages Prepositional Logic Predicate Logic IF THEN Rules (Production Rules) Semantic Nets
5
5 Lecture 34 Knowledge Representation using Logic Predicate Logic Syntax Constant symbols ( Khalid, Ali, car, 3,..) Function symbols (mapping to a constant) Father(khalid)=Ali Predicate symbols (mapping to truth values) greater(5,3) Variable symbols. E.g., x, y Connectives. not (~), and (^), or (v), implies (=>), if and only if ( ) Quantifiers: Universal ( A) and Existential (E)
6
6 Lecture 34 Knowledge Representation using Logic (cont’d) Not all cars are BMWs. ~(forall x)[Car(x) ==> BMW(x)] (exists x)[Car(x) ^ ~BMW(x)] Some numbers are not real. ~(forall x)[Number(x) ==> Real(x)] (exists x)[Number(x) ^ ~Real(x)] Every number is either negative or has a square root. (forall x)[Number(x) ==> (negative(x) v has-sqrt(x))] ~(exists x)[Number(x) ^ ~negative(x) ^ ~has-sqrt(x)]
7
7 Lecture 34 Knowledge Representation using Logic (cont’d) Example Statements Khalid is a Student Khalid is Muslim Every Muslim Prays Every Student has ID Any person that doesn't pray is not a Muslim Representation using Predicate Logic Student(Khalid) Muslim(Khalid) (for all x) [muslim(x) -> prays(x)] (for all x) [~pray(x)->~Muslim(x)]
8
8 Lecture 34 Knowledge Representation using Symantec Networks Semantic networks (Entities and Relationships) Entities are the basic objects of the system and relationships indicates how these objects are related. Is-a relationship means inheritance and can help us to infer the properties of an entity from another entity. This inference is done through searching.
9
9 Lecture 34 Expert System (ES) An ES is computer application that performs a task that would otherwise be performed by a human expert. Knowledge from human experts in a specific field is encoded to be accessible. It is essentially a specialized DBMS. It consists of a knowledge base of facts and inference engine, combining rules of fact and rules with which to discover new facts. Expert System examples… 1. Medical diagnosis (MYCIN) 2. Geological data interpretation 3. Tax preparation 4. Configuring/troubleshooting PCs A typical ES architecture consists of: knowledge base module working memory module (for the current data) inference engine forward chaining (inductive, data driven) backward chaining (deductive, goal driven) user interface (possibly a NLI, menu, windows, etc) explanation module
10
10 Lecture 34 Expert System Design To design an expert system, one needs a knowledge engineer, an individual who studies how human experts make decisions and translates the rules into terms that a computer can understand.
11
11 Lecture 34 Expert Systems: Strengths & Limitations Problems... Not all cases are the same; exceptions Hard to explain “how” an expert works Have to quantify qualitative data Knowledge could be difficult to extract and represent Regardless of these problems, Expert Systems have the following advantages: Can produce results much faster than a human expert. Error rate in a successful ES is low and can be lower than in the case of a human expert. ES can make consistent recommendations. An ES can operate in an environment that is hazardous to human. ES can capture the scarce expertise of a uniquely qualified human expert.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.