CS 786L W97 Home Page

Intelligent Software Agents

CS 786L is a graduate level topics in Artificial Intelligence course. This term the course will be a seminar style course that will examine current research in Intelligent Software Agents.

Meeting Time

• Thursday 4:30--7:00pm. Room 3307DC.
• Commencing Thursday Jan 16th.

Timetable

During the first part of the course we will be reading various papers in the area. Each week will cover a collection of papers on a particular sub-topic of the area. The papers to be covered each week are listed in the weekly schedule below. (The majority of the papers will be available on-line, and will be cached locally with pointers to the original source.) Our meeting times will be occupied by discussing these papers. Prior to each meeting all persons attending will be expected to have read the papers and to be prepared for the discussion. During the last few weeks of the course the meeting times will be devoted to student presentations and discussion of those presentations.

Weekly Schedule

1. Jan 16th. What are Software Agents?
   
   • Intelligent Agents: Theory and Practice, by M. Wooldridge and N. Jennings; The Knowledge Engineering Review 10 (2) 115-152.
     This paper comes from Michael Wooldridge's home page, where a number of other papers (follow the link to his publications) are also available.
   • Software Agents: An Overview, by Hyacinth Nwana. This is an online html document, so no local postscript copy available.
     
     
2. Jan 23rd. A Range of Applications
   
   • An Open Agent Architecture, by P. R. Cohen, A. Cheyer, M. Wang, and S. C. Baeg. AAAI Spring Symposium, pp. 1--8, Mar. 1994.
     This paper comes from SRI's Open Agent Architecture page. Other publications are also available there. Also, see Phil Cohen's new home page, for other work on agents.
     To be presented by David Kidston
     
     
   • Software Agents, by M. R. Genesereth and S. P. Ketchpel.
     This paper comes from Stanford's knowledge sharing effort. (It is their overview paper on software agents.
     To be presented by Kevin Chow
     
     
   • OS Agents: Using AI Techniques in the Operating System Environment, by O. Etzioni, H. M. Levy, R. B. Segal, and C. A. Thekkath. IEEE Software, 12(4):42--51, 1995.
     This paper comes from the University of Washington's softbots project. (Specifically, from Oren Etzioni's home page. (Under the additional papers link).
     To be presented by Tali Zvi
     
     
   • Mobile Agents: Are they a good idea?, by C. G. Harrison, D. M. Chess, and A. Kershenbaum.
     This paper comes from IBM's Massively Distributed Systems home page . (Under the intelligent agents, mobile agents subsection).
     To be presented by Blair Conrad
     
     
3. Jan 30th. Planning Based Agents
   This week, and perhaps the next, we will be looking at agents designed around the ability to plan. The planning component of these agents is based on AI planning techniques. If you are not familiar with these techniques you can read An Introduction to Least Commitment Planning by D.S. Weld. (All of the papers this week come from the University of Washington's softbots project.)
   
   
   • Planning-Based Control of Software Agents, by D.S. Weld. Proceedings of AIPS-96, May 1996.
     I will present this paper.
     
     
   • Planning with Execution and Incomplete Information, by K. Golden, O. Etzioni, and D. Weld. University of Washington Technical Report TR96-01-09, February 1996
     To be presented by Ron Petrick.
     
     
   • Planning to Gather Information, by C.T. Kwok and D.S. Weld. Proceedings of AAAI-96, Portland, OR, August 1996.
     To be presented by Michael Fleming.
     
     
4. Feb 6th. Learning/Interface Agents
   This week, we will be looking at methods for learning as applied to the retrieval of information and learning about the user. (All of the papers this week come from the University of Washington's softbots project, via their people page.)
   
   
   • Category Translation: Learning to understand information on the Internet, by M. Perkowitz and O. Etzioni D.S. Weld. Proceedings of IJCAI-95, pp. 930-936, August 1995.
     To be presented by Zhe Liu.
     
     
   • A sound and fast goal recognizer, by N. Lesh and O. Etzioni. Proceedings of IJCAI-95.
     To be presented by Itai Danan.
     
     
   • Scaling up goal recognition, by N. Lesh and O. Etzioni. Proceedings of AAAI-96, Portland, Principles of Knowledge Representation and Reasoning (KR-96), 1996.
     To be presented by Charlie Xu.
     
     
   • Interactive assessment of user preference models: the automated travel assistant., by G. Linden, S. Hanks, and N. Lesh Submitted to Sixth International Conference on User Modeling.
     To be presented by Edwin Chung.
     
     
5. Feb 13th. Learning Agents and Information Retrieval
   This week, we will be looking at methods for learning as applied to the retrieval of information. (Papers 1, 2 and 4 week come from the SIMS project at ISI.)
   
   
   • Using Inductive Learning to Generate Rules for Semantic Query Optimization , by Chun-Nan Hsu and Craig A. Knoblock. Advances in Knowledge Discovery and Data Mining, AAAI Press, 1995
     To be presented by Kevin McGillivray.
     
     
   • Discovering Robust Knowledge from Dynamic Closed-World Data., by Chun-Nan Hsu and Craig A. Knoblock. Proceedings of the Thirteenth National Conference on Artificial Intelligence (AAAI-96), Portland, Oregon, 1996.
     To be presented by Lan Wang.
     
     
   • Dynamic Reference Sifting: A Case Study in the Homepage Domain, by Jonathan Shakes, Marc Langheinrich & Oren Etzioni. Submitted to WWW6. This an on line HTML document.
     To be presented by David Kidson.
     
     
   • Reconciling Distributed Information Sources, by Jose Luis Ambite and Craig A. Knoblock. Working Notes of the AAAI Spring Symposium on Information Gathering in Distributed Heterogeneous Environments, Palo Alto, CA, 1995
     To be presented by Tali Zvi.
     
     
6. Feb 20th. Information Retrieval Agents
   This week, we will be looking at methods for information retrieval. Papers 2 and 3 come from the SIMS project, and Paper 1 comes from the University of Washington
   
   
   • Multi-Service Search and Comparison Using the MetaCrawler, by Erik Selberg and Oren Etzioni. Paper comes from the University of Washington SoftBot project.
     To be presented by Kevin Chow.
     
     
   • Retrieving and Integrating Data from Multiple Information Sources, by Yigal Arens, Chin Y. Chee, Chun-Nan Hsu, and Craig A. Knoblock. International Journal of Intelligent and Cooperative Information Systems. Vol. 2, No. 2. Pp. 127-158, 1993.
     To be presented by Blair Conrad.
     
     
   • Cooperating Information Retrieval Agents, by Craig. A. Knoblock, Yigal Arens, and Chun-Nan Hsu. Proceedings of the Second International Conference on Cooperative Information Systems, Toronto, Ontario, Canada, University of Toronto Press, 1994.
     To be presented by Ron Petrick.
     
     
7. Feb 27th. Information Retrieval Agents and KQML
   This week, we will examine one more paper on the SIMS system and then two introductory papers on KQML.
   
   
   • The SIMS Manual, by Jose-Luis Ambite et at.
     Note that this is a manual for the SIMS system. You don't have to worry about the syntactic details of the systems input and output. In particular Sections 6-9 can be simply skimmed. However the first five sections presents an nice overview of the architecture of their system, and some insights into how useable the system would be.
     To be presented by Zhe Liu.
     
     
   • KQML-A Language and Protocol for Knowledge and Information Exchange, Tim Finin, Rich Fritzson, Don, Mckay, and Robin McEntire. From Tim Finin's agents page.
     To be presented by Michael Fleming.
     
     
   • Evaluation of KQML as an Agent Communication Language, by James Mayfield, Yannis Labrou, and Tim Finin. From Tim Finin's agents page.
     To be presented by Itai Danan.
     
     
8. March 6th. KQML and Knowledge Sharing Software.
   This week we will continue our examination of KQML, and look at proposals for how KQML communication can be used for coordinating software programs.
   
   
   • A semantics approach for KQML-a general purpose communication language for software agents., by Yannis Labrou and Tim Finin.
     To be presented by Charlie Xu.
     
     
   • Secret Agents-A Security Architecture for KQML, by Chelliah Thirunavukkarasu, Tim Finin, and James Mayfield.
     To be presented by Edwin Chung.
     
     
   • Coordinating Distributed Objects with Declarative Interfaces, by Narinder Singh and Mark Gisi
     To be presented by Kevin McGillivray.
     
     
   • The Information Agent: An Infrastructure Agent Supporting Collaborative Enterprise Architectures, Mihai Barbuceanu and Mark Fox.
     To be presented by Lan Wang.
     
     
9. March 13th. Agent Oriented Programming.
   This week we will look at proposals for programming using an agent oriented pradigm.
   
   
   • Agent Oriented Programming, by Y. Shoham.
     There is no on-line version of this paper. It appeared in Artificial Intelligence, Volume 60, page 51-92. Copies of this paper will be placed on reserve in the library.
     I will present this paper.
     
     
   • Agent-K: An Integration of AOP and KQML, by Winton Davies and Peter Edwards.
     To be presented by David Kidston.
     
     
   • A Survey of Concurrent MetateM-The Language and its Applications, by Michael Fisher
     To be presented by Kevin Chow.
     
     
10. March 20th. Project Presentations.
    Presenting this week will be.
    
    1. Lan Wang.
    2. Kevin McGillivray.
    3. Edwin Chung.
    4. Charlie Xu.
    
    
11. March 27th. Project Presentations.
    Presenting this week will be.
    
    1. Itai Danan.
    2. Zhe Liu.
    3. Michael Fleming.
    4. Ron Petrick.
    
    
12. April 3rd. Project Presentations.
    Presenting this week will be.
    
    1. Blair Conrad.
    2. Tali Zvi.
    3. Kevin Chow.
    4. David Kidston.
    
    

Links to other pages on agents

• Sverker Janson's page on Intelligent Software Agents.
• Ralph Becket's page on Intelligent Software Agents.
• Tim Finin's page on Intelligent Software Agents.

Course Evaluation (Changes Made on Jan 13rd)

Finally, everyone will be required to complete a project involving either a paper investigating in greater depth a subtopic they are interested, or if feasible an implementation that utilizes some agent theory.

• 10% Class participation.
• 20% Class presentations.
• 20% Presentation of Project.
• 50% Project writeup.

Information about Paper Presentations

Here is some information about what I will be looking for when you present overviews and commentaries on the papers to the class.

I would like to see a short summary/overview of the paper along with some analysis. The analysis could cover various things, and it is intended to be short so you will probably be forced to decide what things to discuss. Possible things you could mention as the analysis component:

1. What you think are the key contributions of the paper and why.
2. What you think are the key weaknesses of the paper and why.
3. What you think are errors or flaws in the paper and why.
4. What you think are the key assumptions being made in the paper, and whether or not you think these assumptions are valid (and why).
5. What you think are the key directions for future research to prove or realize the promise of the paper's suggestions (and why).
6. How you think the ideas of the paper could be applied to a different problem and what this might promise.

Class Participants

For Credit

1. David Kidston
2. Kevin Chow
3. Tali Zvi
4. Blair Conrad
5. Ron Petrick
6. Michael Fleming
7. Zhe Liu
8. Itai Danan
9. Charlie Xu
10. Edwin Chung
11. Kevin McGillivray
12. Lan Wang

Auditing/Attending Class

1. Trang Dang
2. Valdo Keselj
3. Forbes Burkowski
4. Ion Vasilian
5. Stuart Gill
6. Eric Demaine
7. Nancy Moussa
8. Eric Leung

More to come...

Instructor: Fahiem Bacchus, DC2510, x4670, fbacchus@logos.uwaterloo.ca

To comment on this web page send mail to CS786L