Authorised by Academic Registrar, April 1996
Objectives At the completion of the subject, students should be able to produce simple computational models of cognitive processes; implement the above models in appropriate languages such as Prolog and LISP; use formal logic augmented with structured data objects to represent real-world knowledge; and understand the fundamentals of artificial neural nets simulations.
Synopsis This subject addresses the fundamentals of symbolic and numeric computational models of cognitive processes, and the heuristic programming techniques used to solve NP-complex problems. The syllabus covers the following topics. Computers and cognitive processes, language and programming, perception and neural nets. State-space methods to solve NP problems. Search as a starting point for heuristic programming. Constraint satisfaction, dependency-directed backtracking, means-ends analysis, depth-, breadth-, best- first- search; mini-max, alpha-beta, A* search algorithms; neural nets and optimisation. Introduction to formal logic, logical inference and LUR-resolution, procedural and declarative knowledge representation, default reasoning, semantic nets and frames. Expert systems, forward and backward chaining, management of uncertain knowledge. Machine learning, learning from examples, learning and production systems, discovery of concepts. Applications: machine vision, robotics and STRIPS-like planning, natural language processing and conceptual dependency graphs.
Assessment Examination: 60% + Assignment and practical work: 40%