Each student is responsible for his/her own work. The standard departmental rules for academic dishonesty apply to all assignments in this course. Collaboration on homeworks and programs should be limited to answering questions that can be asked and answered without using any written medium (e.g., no pencils, pens, or email). This means that no student should read any code written by another student.
wangha@cs.orst.edu
.
INTRODUCTION Sept 25 AI = the design of rational agents [1.1, 1.4] 27 Structure of intelligent agents and environments [2 (all)] Simple reflex agents, agents with memory, goal-based agents, utility-based agents. SEARCH-BASED AGENTS (environment is accessible, deterministic, static, discrete) (environment is known and modeled using arbitrary code) Sept 29 Problems and problem spaces [3.1, 3.2, 3.3] Intro to lisp (lisp warmup exercise?) Oct 2 Tree search algorithms: breadth-first, uniform-cost [3.4, 3.5] 4 Depth first, Iterative deepening, Repeated states [3.6] 6 Informed search: best-first, greedy, A* [4.1, 4.2] 9 Local search: hill-climbing, gradient ascent, simulated annealing [4.4] LOGICAL AGENTS (environment is inaccessible) (environment is known and modeled using logical inference) 11 Knowledge bases, Wumpus world [6.1, 6.2] 13 Propositional logic, Horn clauses [6.3, 6.4] 16 Inference, building an agent [6.5, 6.6] DECISION-THEORETIC AGENTS (environment is accessible but nondeterministic) (environment is known and modeled using belief networks) 18 Introduction to probability 20 MIDTERM EXAM 23 More probability, Bayes theorem [14 all] 25 Belief networks and the SPI algorithm [15.1, 15.2, handout] 27 Belief networks and the SPI algorithm (2) Oct 30 Influence diagrams and agent design [16.1, 16.2] Nov 1 Utility functions [16.3] 3 Sequential decision-making. No sensors. MDP's. [17.1, 17.2] 6 Value Iteration, Policy iteration, Rational agent design [17.3, 17.4] 7 POMDP's, Dynamic belief networks and decision networks [17.5, 17.6] LEARNING AGENTS (environment unknown, accessible) 10 Adaptive DP [20.2, 20.3, 20.4, 20.5] 13 temporal averaging, Q learning [20.6] 15 Supervised learning in deterministic environments [18.1, 18.2] 17 Learning decision trees [18.3] 20 Learning neural networks [19.1, 19.2, 19.3] 22 Backpropagation [19.4] 24 No Class (Thanksgiving Holiday) 27 Learning belief networks [20.6] OPEN PROBLEMS Nov 29 Meta-level architectures; reasoning about reasoning Dec 1 Can machines think? The future of AI. 5 2:00 FINAL EXAM
Graduate students will be expected to submit a term project. This can either be an expansion of one of the programming assignments, or a report on a set out additional readings. Graduate student project proposals (one paragraph, hardcopy or email) must be submitted at or before the midterm.