STatistical AI Reading Group previous readings: September-December, 2003

---

September 10 (Stairmaster: Mike):

The Information Bottleneck Method, (1999), Naftali Tishby, Fernando C. Pereira, William Bialek Proc. of the 37-th Annual Allerton Conference on Communication, Control and Computing Citeseer .

September 18 (Stairmaster: Natalia & Luke):

First-order probabilistic inference. David Poole. International Joint Conference in Artificial Intelligence 2003 (IJCAI-03) html

September 25 (Stairmaster: Natalia):

Logical Filtering. E. Amir and S. Russell. International Joint Conference in Artificial Intelligence 2003 (IJCAI-03) Postscript

October 2 (Stairmaster: Hanna):

Generalizing Plans to New Environments in Relational MDPs. Carlos Guestrin, Daphne Koller, Chris Gearhart and Neal Kanodia. International Joint Conference in Artificial Intelligence 2003 (IJCAI-03). Postscript .

October 9 (Stairmaster: Leon):  Overiew of RL and active vision.  Suggested papers: 

 "A Reinforcement Learning Model of Selective Visual Attention", Silviu Minut and Sridhar Mahadevan,  Proceedings of the 5-th International Conference on Autonomous Agents, pp 457-464, Montreal, Canada, 2001 http://www.cs.umass.edu/~mahadeva/papers/aa2001.ps.gz

"Learning Visual Routines with Reinforcement Learning" Andrew McCallum, ftp://ftp.cs.rochester.edu/pub/papers/robotics/96.mccallum-aaai-e.ps.gz
 
"Learning to generate artificial fovea trajectories for target detection" Schmidhuber, Juergen and Huber, R., International Journal of Neural Systems, 2(1 & 2):135-141, 1991,  ftp://ftp.idsia.ch/pub/juergen/attention.ps.gz

"Residual Q-Learning Applied to Visual Attention" Bandera, C. ., Vico, F. J., Bravo, J. M., Harmon, M. E., & Baird, L. C. Proceedings of the Thirteenth International Conference on Machine Learning, Bari, Italy, 3-6 July (1996) http://www.leemon.com/papers/icml96/ICML96.pdf

 "Active object recognition by view integration and reinforcement learning", Lucas Paletta, Axel Pinz,  Int.J. Robotics and Autonomous Systems, 31(1-2):71-86, 2000. http://www.emt.tu-graz.ac.at/~pinz/onlinepapers/RAS00.pdf

October 16 (Stairmaster: Georgios): 

Factored Planning. E. Amir and B. Engelhard, International Joint Conference in Artificial Intelligence 2003 (IJCAI-03) Postscript

October 30 (Stairmaster: Yu-Han): 

Distributed Planning in Hierarchical Factored MDPs, Carlos Guestrin and Geoffrey Gordon, UAI-2002. Postscript .

November 6 (Stairmaster: Kurt): 

"APRICODD: Approximate policy construction using decision diagrams", Robert St-Aubin , Jesse Hoey and Craig Boutilier, Advances in Neural Information Processing 13 ( NIPS 2000). ps.gz.

"Optimal and Approximate Stochastic Planning using Decision Diagrams (2000) (Make Corrections) Jesse Hoey, Robert St-Aubin, Alan Hu, Craig Boutilier, tech report. ps.gz

"SPUDD: Stochastic Planning using Decision Diagrams", Jesse Hoey , Robert St. Aubin, Alan Hu and Craig Boutilier, Proceedings of  UAI 99. ps.gz.

The focus of this talk is on the approximate dynamic programming algorithm known as approximate linear programming. I  will present background on the algorithm and then discuss my work,  including performance and approximation error bounds, an efficient  constraint sampling scheme for dealing with the large number of constraints  the algorithm typically involves, and application to queueing networks.

November 20 (Stairmaster: Georgios): 

"A Nonlinear Predictive State Representation", Matthew R. Rudary and Satinder Singh, NIPS-03. PDF

December 4  (Stairmaster: Bruno): 

Eyal Even-Dar, Shie Mannor, Yishay Mansour: Action Elimination and Stopping Conditions for Reinforcement Learning. ICML 2003: 162-169. PDF