CSE 574 - Automated Planning Winter 2003

Welcome to the 574 Home Page.

Instructor: Dan Weld

Automated planning is one of the oldest areas in AI. We'll review all the major approaches (state-space search, partial-order planning, graphplan derivatives, SAT and CSP compilation, etc.) and discuss speedup techniques ranging from heuristic derivation to EBL and reachability analysis. Given this foundation, the course will focus on the two hottest areas of current research:

• Metric temporal and resource planning (planning graph methods, integration with linear programming solvers, etc).
• Planning under uncertainty, including Markov decision processes, POMDPs, PRM-inspired methods, and representations for correct but incomplete knowledge such as local closed world reasoning.

The course has two goals: covering the material from a number of influential and interesting papers on these topics, teaching you to synthesize material on a topic and lead a class discussion on it, and suggesting anumber of research projects suitable for quals or generals (in the past, several of these have led to AAAI, IJCAI or ICAPS papers).

Accordingly, the class will comprise two paper-related components:

1. For each class, you will need to read one or two papers, prepare paper summaries (CSE551 style) by 11am the day of class, and come to class ready for discussion. Here are your instructions for reviewing papers.
2. For one class, you will be responsible for reading a small set of papers, picking the one or two papers everybody else will read, and preparing material to lecture and guide discussion. Here are your guidelines for preparing for your lecture/discussion.

In addition, there will be a project component.

Administrivia

• Here are instructions on joining the class mailing list, as well as an archive of the class mailing list.
• The class mailing list address is cse574@cs.washington.edu.

Your final grade will be assigned based on the following:

• 25% your lecture
• 35% project
• 20% class participation
• 20% paper summaries

Paper Schedule

• Instructions for reviewing papers

• Reading list and paper summaries

Computer Science & Engineering Department
University of Washington
PO Box 352350
Seattle, WA 98195-2350 USA