Computer Science 4500 -- Artificial Intelligence

Test 2 -- Review

1. Expert systems

history; limited domain; knowledge engineering; development;
shell structure: knowledge base (rules+data), inference engine,
user interface, explanation facility, knowledge update
Knowledge Engineering, process, scoping domain, selecting expert
tools: levels

2. Knowledge representation

Knowledge representation principles: representational
adequacy, acquisitional efficiency, inferential adequacy,
inferential efficiency
Predicate logic notation, predicates, slots, relationships
OAV notation: objects, attributes, values
Conceptual graphs & semantic nets: nodes (objects/actions with
property lists), arcs/connections (relationships between nodes)
Conceptual dependency: primitives, pictures, actions, aiders,
actions/relationships/transitions (arrows), tenses
Frames: slot-filler type, classes of objects & instantiations,
frame system, procedural attachment
Scripts: tracks, props, roles, entry & exit conditions, scenes/events
Non-monotonic logic issues
Reasoning under uncertainty: Bayesian logic, Dempster-Shafer theory,
Stanford certainty factors
Relative measures and fuzzy logic/sets

3. Natural language processing

Natural vs. formal language, elements (phonemes, syllables, words)
Spoken language: speech recognition (continuous vs. isolated)
Written language: character-recognition, linguistic levels
(syntactic, semantic, pragmatic), formal grammar components & CFG,
meta-language, parsing, ambiguity, resolving ambiguity
Processing vs."understanding", TNs, RTNs, ATNs;
implementing RTNs and ATNs with recursive functions;
Stochastic (Markov model) and clustering approaches to language;
language application systems: article summaries, database front end