AI, Fuzzy Logic and Neural Network
 
 
Subject Code: ECP4176
Aim of Subject: To expose students to the fundamentals of AI, fuzzy logic and neural network and their engineering applications.Discussion on the introductory aspects of definition, models, process and applications of artificial intelligence, fuzzy logic and neural networks.
Learning Outcome of Subject: At the completion of the subject, students should be able to:
  • Have a clear understanding of AI definitions, its knowledge representations, use of logic and structural representations.
  • Define the expert system structure, development process, knowledge acquisition and development tools.
  • Understand the fundamentals of Fuzzy approaches; - Fuzzy Set Theory, Classical Sets, Fuzzy Sets, Types of Membership Functions, Fuzzy Sets Representation, Fuzzy Sets Terms and Notations, Power of a Fuzzy Set, Fuzzification, Level Fuzzy Sets, Properties of Fuzzy Sets and Extension Principle.
  • Define and manipulate the Fuzzy Relations, its representations, relations, properties, basic operations and compositions.
  • Understand Fuzzy Linguistic descriptions and their Analytical Forms.
  • Identify the basic structure of Fuzzy Logic Systems.
  • Understand the features of Neural Networks, types of activation functions and its classifications.
  • Understand perceptrons concepts and apply the Error back-Propagation algorithms.
  • Have an overview on several applications of AI in engineering and its design issues.
Programme Outcomes:
  • Ability to acquire and apply fundamental principles of science and engineering(50%)
  • Capability to communicate effectively(10%)
  • Acquisition of technical competence in specialised areas of engineering discipline(10%)
  • Ability to identify, formulate and model problems and find engineering solutions based on a systems approach(10%)
  • Ability to conduct research in chosen fields of engineering(5%)
  • Ability to work independently as well as with others in a team(10%)
  • Capability and enthusiasm for self-improvement through continuous professional development and life-long learning(5%)
Assessment Scheme:
  • Lab Experiment - work in group of 2,written and oral assessment at the end of lab(10%)
  • Tutorial / Assignment - group assignment,to enhance understanding of basic concepts in lecture(15%)
  • Test/Quiz - written exam(15%)
  • Final Exam - written exam(60%)
Teaching and Learning Activities : 51 hours (lectures,tutorials and laboratory experiments)
Credit Hours: 3
Pre-Requisite: None
References:
  • Patrick H. Winston, "Artificial Intelligence", Addison Wesley, 3rd Edition, 1992.
  • L. X. Wang, "Adaptive Fuzzy Systems and Control", Prentice Hall, 1994.
  • S. Y. Kung, "Digital Neural Network", Prentice Hall, 1993.
  • R. L. Harvey, "Neural Network Principles", Prentice Hall, 1994.
  • S. Hekmatpour, "LISP: A Portable Implementation", Prentice Hall, 1989.

Subject Contents

  • Artificial Intelligence

    • Definition and objectives of AI, signs of intelligence, Turing test. Brief history of AI. Major areas of AI. Application areas of AI. AI vs. conventional programming. Knowledge representation, use of logic and structural representation. Problem representation, problem solving methods. AI languages, computer architectures for AI applications, signal and image processing. Definition of expert system, structure of an expert system. Expert system development process, knowledge acquisition, development tools.
     
  • Fuzzy Logic

    • Introduction to fuzzy set theory, knowledge base problem, objective and subjective knowledge, crisp sets, fuzzy sets, linguistic variables, membership functions. Set theoretic operations, comparison between crisp sets and fuzzy sets. Law of Contradiction and Law of Excluded Middle, fuzzy intersection, union and complement, and other fuzzy operators. Fuzzy relations and compositions on the same and different product spaces. Max-Min composition, Max-Product composition, fuzzy relational matrix, sup-star composition. Hedges or modifiers of linguistic variables, fuzzy logic vs. probability. Fuzzy reasoning and implication, the fuzzy truth tables, traditional propositional logic and the rule of inference, the Modus Ponens and Modus Tollens, fuzzy modeling with causal IF-THEN statements. Fuzzy Models, fuzzy logic systems, combination of fuzzy basis functions, universal approximator, fuzzy neural network, fuzzy associate memory matrix, self-learning fuzzy systems. Fuzzy logic system applications.
     
  • Neural Network

    • Definition of artificial neural network. Similarities of neural network with human brain. Classification of ANN. Terms used in ANN: Input/output sets, weights, bias or threshold, supervised learning, network training, Convergence process, single layer vs. multilayer perceptron, Forward and Backward propagation, gradient descent rule. Back-propagation neural network, Variable term used in back propagation neural network: learning rate, momentum, hidden nodes, sigmoid activation function. Back propagation algorithm of ANN. Design of ANN model, training sets for ANN, test sets for ANN, network testing and performance. Application of ANN in engineering.

Laboratory

1. Fuzzy Logic systems for Function approximation.
2. Training of Multilayer feed forward Neural Networks.