277-3204

Textbook (L&S, for reference purposes in the following descriptions):

Week 1: Artificial Intelligence, its roots and scope (L&S, ch1, Intro Part II)

• AI, an attempted definition

• Historical foundations

• Overview of application areas

• An introduction to representation and search

Week 2: The Predicate Calculus (L&S, ch2)

• Representation languages

• The propositional calculus and its semantics

• The predicate calculus: syntax & semantics

• Inference: soundness, completeness

• The unification algorithm

Week 3: Structures and strategies for state space search (L&S ch 3)

• Quick review of graphs

• State space search

• Data driven and goal driven search

• Breadth first, depth first, and depth first iterative deepening search

Week 4: Heuristic search (L&S, ch 4).

• Priority queues

• A*

• Iterative deepening A*

• Beam search

• Two-person games

• Mini-Max and alpha-beta

Week 5: Architectures for AI problem solving (L&S, ch5)

• Recursive specification for queues, stacks, and priority queues

• The production system

• The blackboard

• Planning

Week 6: PROLOG and LISP (L&S, chs 9 & 10)

• The PROLOG/LISP environments

• Relational specifications and rule based constraints

• Graph search with the production system

Weeks 7 & 8: The knowledge based expert systems (L&S, ch 6)

• Production system based search: data-driven, goal-driven

• Rule stacks and the "why" query, proof trees and the "how" query

• Case Based Reasoning

• Model Based Reasoning

• Comparative approaches to knowledge based problem solving

Weeks 9 & 10: Reasoning in uncertain situations (L&S, ch 7)

• Bayes Rule; Bayesian Belief Networks

• Abductive inference, Causal networks

• Stanford Certainty Factor Algebra

• Fuzzy systems

Week 11: Building a rule based expert system in PROLOG & LISP (L&S, chs 9 & 10)

• Meta-predicates in PROLOG; a meta-interpreter: PROLOG in PROLOG

• Rule-stacks, proof-trees, and certainty factor algebras

• Exshell, a back-chaining rules interpreter in PROLOG

• Lispshell, a back-chaining rules interpreter in LISP

Week 12: Introduction of structured AI representational Schemes (L&S, ch 8)

• Issues in knowledge representation

• Semantic networks

• Conceptual dependencies

• Frames, scripts, and object systems

• The hybrid design: objects with rule sets

Weeks 13 & 14: Advanced Topics in Artificial Intelligence (L&S, chs 11 - 15)

• Programs that understand Natural Language: knowledge-based and Markovian

• Automated Reasoning

• Models for machine learning: symbol based, connectionist, and genetic

Week 15: Course summary and review (L&S, ch 16)

• The possibility of a science of intelligence

• Limitations and future research

There are two examinations, a mid-term and a final, each one hour long

There will be three or four programming tasks; possible assignments include:

1. Building graph search algorithms in PROLOG

a) depth first

b) breadth first

c) best search

2. Building graph search algorithms in LISP

a) depth first

b) breadth first

c) best search

3. Using EXSHELL or Lispshell to build a rule based expert reasoning system

There will also be an 8-10 page paper, on a topic in AI of the students choice

Course credit: Mid-term and final 35% each, programming assignments 20%, paper 10%