Computer Science 375:
Optimization and Uncertainty

Fall 2012

Mon/Wed, 10:00 - 11:25 pm, Searles 126
There are many views of what artificial intelligence is. In one view, artificial intelligence attempts to represent desires and goals in a way that can be "understood" by a computer and to design algorithms that help the computer avoid undesirable actions and achieve the desired goal. For many real-world tasks, this means solving optimization problems and coping with uncertainty. A numeric framework, rather than the symbolic one of traditional artificial intelligence, is useful for expressing and attacking such problems. We will explore a number of artificial intelligence topics in this numeric framework, with particular attention to planning under uncertainty.


ADMINISTRATIVE INFORMATION:
Instructor: Stephen Majercik, 222 Searles Hall, 725-3106, smajerci@bowdoin.edu
Office Hours: Mon/Tues, 2:00-3:30 pm, or by appointment 
Class E-mail: csci375@bowdoin.edu


TEXT (optional):
Stuart Russel and Peter Norvig
Artificial Intelligence: A Modern Approach, Third Edition
Prentice-Hall, 2010.


PREREQUISITES:
 Computer Science 210


REQUIREMENTS :
Problem Sets and Small Programming Projects 65-70%
Final Project30-35%
Class NotesPossible Extra Credit
Class ParticipationTiebreaker

COURSE OUTLINE:
 See reverse.











TENTATIVE COURSE OUTLINE:
Dates Topic Readings
Sep 3 Introduction NA
Sep 5 Deterministic and Stochastic Planning 10.1, 10.2
Sep 10, 12, 17, 19 Markov Decision Processes Ch 17.1-17.3, 21.1-21.3
Sep 24, 26 Neural Networks Ch 18.7
Oct 1, 3 Probability Ch 13.1-13.5
Oct 10, 15, 17 Bayes Networks Ch 14.1-14.4
Oct 22, 24 Hidden Markov Models Ch 15.1-15.3
Oct 29, 31 Machine Vision Ch 24
Nov 5 Project Possibilities NA
Nov 7, 12 Partial Order Planning Ch 10.4.4
Nov 14, 19 Planning as Graph Analysis Ch 10.3
Nov 26, 28 Planning as Satisfiability Ch 10.4.1
Dec 3, 5 Philosophical, Social, and Ethical Issues in AI Ch 26 and Handouts
Exam Period Final Project Presentations NA