Higher Computing: Artificial Intelligence

The candidate must demonstrate knowledge and understanding, practical skills and problem solving based on the following content statements:

 Identification of applications of NLP (including automatic The development of artificial intelligence translation, speech driven software, NL search engines, NL  Definitions of human intelligence and artificial intelligence database interfaces)  Descriptions of aspects of intelligence (including language, learning, cognitive ability, problem solving skills, memory, Smart/embedded technology: creativity)  Description of examples of the use of intelligent software to  Explanation of the difficulties of determining an accurate control devices (including car engine control systems; and agreed definition of intelligence. domestic appliances) Intelligent robots: Explanation of the  Explanation of the inherent flaws of the Turing test as a difference between dumb and intelligent robots method for determining the existence of artificial  Description of contemporary research and developments intelligence  Description of possible social and legal implications of the  Description of the change in emphasis from modelling the increasing use of intelligent robots human brain to producing systems exhibiting ‘intelligent  Descriptions of practical problems (including processor behaviour’ power, power supply, mobility, vision recognition,  Description of the need for knowledge representation navigation, path planning, pick and place, and strategies techniques (including semantic nets and logic programming) used to overcome these problems)  Explanation of the need for a restricted domain  Identification of languages: LISP (functional), Prolog Expert systems (declarative/logic)  Description of the components of an expert system  Description of difference between declarative and (knowledge base, inference engine, user interface with imperative languages justification/explanation, working memory)  Explanation (with examples) of:  Distinction between an expert system and an expert system o the success and failures of game playing shell programs from simple early examples to  Description of contemporary applications of expert systems contemporary complex examples exhibiting  Description of advantages of expert systems (including intelligence permanence, cost effectiveness, consistency, portability) o the successes and failures of language processing  Description of disadvantages of expert systems (including (including Eliza, SHRDLU, chatterbots and narrow domain, lack of ‘common sense’, need for expertise contemporary applications) to set up and maintain, inability to acquire new knowledge, o the scope and limitations of expert systems inflexibility)  Explanation of the effects of hardware developments  Description of moral issues (including medical implications) (including faster processors, more memory, and increasing  Description of legal issues (including responsibility when backing store capacity) on the field of AI advice is wrong)  Description of the implementation and advantages of parallel processing Search techniques  Description of the practical problems associated with AI  Comparison of depth-first and breadth-first search (order of despite advances in hardware/software visiting nodes, memory implications, advantages and disadvantages, need for backtracking), and exemplification Applications and uses of artificial intelligence on a search tree Artificial neural systems (ANS):  Description and exemplification of combinatorial explosion  Comparison of a human neuron with an artificial neuron  Description and exemplification of use of heuristics to  Description of the structure of a neural net (including reduce search time/space artificial neuron, links, weights, layers)  Comparison of a neural net with the human brain Knowledge representation  Description of ‘learning’ through iterative process as  Description of the software development process as it opposed to algorithmic programming applies to declarative language programming  Explanation that a neural net may be a software model or  Creation of a semantic net from given problem statement hard-wired  Description and exemplification of the following features in Prolog (or similar declarative language): Vision systems o multi–argument clauses  Description of the problems of interpreting 2D images of 3D o recursive and non recursive rules objects o complex queries: (multiple variable, conjunction  Description of the stages of computer vision (image of queries) acquisition, signal processing, edge detection, object o negation recognition, image understanding) o inheritance  Explanation of the concepts of goal, sub-goal, instantiation, Natural language processing (NLP): matching  Identification of the main stages of NLP (speech  Explanation of complex manual trace: multiple level recognition, natural language understanding (NLU), natural including backtracking language generation, speech synthesis)  Explanation of the importance of the order of rules  Explanation of some difficulties in NLP (including ambiguity of meaning; similar sounding words; inconsistencies in grammar of human language; changing nature of language)