Change History for These Pages

August 22, 2006. Added quick jump links at top of page. Updated CV and publications.

September 28, 2005. Updated CV and publications.

July 16, 2005. Updated current and former students and project descriptions and my new office number.

January 22, 2004. Updated current and former students. Updated conferences. Added link to IMLS web site.

January 4, 2004. Updated office hours, teaching, and conferences.

January 6, 2003. Updated Conference/Workshop list. Added CS534 for Winter 2003.

April 18, 2002. Updated pointers to several papers. These replace the versions listed below.

• Dietterich, T. G., Becker, S., and Ghahramani, Z. (eds.) (2002). Advances in Neural Information Processing Systems 14. MIT Press, Cambridge, MA. Online Proceedings.

• Margineantu, D., Dietterich, T. G. (2002). Improved Class Probability Estimates from Decision Tree Models. To appear in Lecture Notes in Statistics, © Springer-Verlag. Postscript Preprint. Postscript Preprint.

• Dietterich, T. G., Busquets, D., Lopez de Mantaras, R., Sierra, C. (2002). Action Refinement in Reinforcement Learning by Probability Smoothing. In Proceedings of the International Conference on Machine Learning. Postscript Preprint. PDF Preprint.

• Zubek, V. B., Dietterich, T. G. (2002). Pruning Improves Heuristic Search for Cost-Sensitive Learning. Postscript Preprint. PDF Preprint.

February 6, 2002. Added papers submitted to ICML-2002:

• Zubek, V. B., Dietterich, T. G. (submitted). Pruning Improves Heuristic Search for Cost-Sensitive Learning. Postscript Preprint.

• Dietterich, T. G., Busquets, D., Lopez de Manataras, R., Sierra, C. (submitted). Action Refinement in Reinforcement Learning by Probability Smoothing. Postscript Preprint.

November 6, 2001. Added paper submitted to AAMAS-2002 conference:

• Busquets, D., Lopez de Mantaras, R., Sierra, C., and Dietterich, T. G. (submitted). Reinforcement learning for landmark-based robot navigation. Postscript Preprint

September 23, 2001. Added links for CS361. Added updated references to NIPS papers:

• Dietterich, T. G. and Wang, X. (to appear). Batch value function approximation via support vectors. Accepted for publication in Dietterich, T. G., Becker, S., Ghahramani, Z. (Eds.) Advances in Neural Information Processing Systems 14. Cambridge, MA: MIT Press. Postscript preprint.

• Wang, X. and Dietterich, T. G. (to appear). Stabilizing value function approximation with the BFBP algorithm. Accepted for publication, Dietterich, T. G., Becker, S., Ghahramani, Z. (Eds.) Advances in Neural Information Processing Systems 14. Cambridge, MA: MIT Press. Postscript preprint.

August 19, 2001. Updated links and deleted most dead links.

December 6, 2000. Updated publications to include the following:

• Leen, T. K., Dietterich, T. G., and Tresp, V. (2001) Advances in Neural Information Processing Systems, 13, Cambridge, MA: MIT Press.

• Dietterich, T. G. (2000). Hierarchical reinforcement learning with the MAXQ value function decomposition. Journal of Artificial Intelligence Research, 13, 227-303. Compressed postscript. Also available from my FTP directory as Gzipped postscript

• Dietterich, T. G. (2000). Machine Learning. In David Hemmendinger, Anthony Ralston and Edwin Reilly (Eds.), The Encyclopedia of Computer Science, Fourth Edition, Thomson Computer Press. 1056--1059.

• Dietterich, T. G. (2000). An Overview of MAXQ Hierarchical Reinforcement Learning. In B. Y. Choueiry and T. Walsh (Eds.) Proceedings of the Symposium on Abstraction, Reformulation and Approximation SARA 2000, Lecture Notes in Artificial Intelligence (pp. 26-44), New York: Springer Verlag. Postscript preprint © Springer-Verlag.

May 15, 2000. Added papers to appear at MCS and SARA and updated the publication information for PRICAI-2000 paper with Bayer:

• Dietterich, T. G. (2000). Ensemble Methods in Machine Learning. To appear in F. Roli (Ed.) First International Workshop on Multiple Classifier Systems, Lecture Notes in Computer Science, New York: Springer Verlag. Postscript preprint © Springer-Verlag.

• Dietterich, T. G. (2000). An Overview of MAXQ Hierarchical Reinforcement Learning. To appear in B. Y. Choueiry and T. Walsh (Eds.) Proceedings of the Symposium on Abstraction, Reformulation and Approximation SARA 2000, Lecture Notes in Artificial Intelligence, New York: Springer Verlag. Postscript preprint © Springer-Verlag.

• Bayer, V. and Dietterich, T. G. (2000) A POMDP Approximation Algorithm that Anticipates the Need to Observe. To appear in Proceedings of the Pacific Rim Conference on Artificial Intelligence (PRICAI-2000); Lecture Notes in Computer Science. Springer-Verlag. Postscript Preprint.

• Dietterich, T. G. (2000). Ensemble Methods in Machine Learning. In J. Kittler and F. Roli (Ed.) First International Workshop on Multiple Classifier Systems, Lecture Notes in Computer Science (pp. 1-15). New York: Springer Verlag. Postscript preprint © Springer-Verlag.

April 27, 2000. Updated publications with final versions of papers for ICML and KDD:

• Fountain, T., Dietterich, T. G., Sudyka, B. (2000). Mining IC Test Data to Optimize VLSI Testing. To appear in Proceedings of the Sixth ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. ACM Press. PDF preprint.

• Chown, E., Dietterich, T. G. (2000). A divide-and-conquer approach to learning from prior knowledge. To appear International Conference on Machine Learning, ICML-2000. Postscript preprint.

• Margineantu, D. D., Dietterich, T. G. (In Press). Bootstrap Methods for the Cost-Sensitive Evaluation of Classifiers . To appear, International Conference on Machine Learning, ICML-2000. Postscript preprint

February 7, 2000. Posted final version of NIPS99 paper.

• Dietterich, T. G. (2000). State abstraction in MAXQ hierarchical reinforcement learning. In Advances in Neural Information Processing Systems, 12. S. A. Solla, T. K. Leen, and K.-R. Muller (eds.), 994-1000, MIT Press. Postscript Preprint.

Posted ICML Submissions.

• Margineantu, D. D., Dietterich, T. G. (Submitted). Bootstrap Methods for the Cost-Sensitive Evaluation of Classifiers . Postscript preprint
• Bayer, V., Dietterich, T. G. (Submitted). A POMDP Approximation Algorithm that Anticipates the Need to Observe. Technical Report 00-30-01, Department of Computer Science, Oregon State University Postscript Preprint.

July 26, 1999. Posted revised version of paper accepted to Machine Learning:

• Dietterich, T. G., (To appear). An experimental comparison of three methods for constructing ensembles of decision trees: Bagging, boosting, and randomization. Machine Learning Postscript preprint.

May 18, 1999. Posted submission to NIPS-99:

• Dietterich, T. G. (submitted). State abstraction in MAXQ hierarchical reinforcement learning. Submitted to NIPS-99. Postscript Preprint.

Posted revised version of my comprehensive paper on the MAXQ formalism:

• Dietterich, T. G. (submitted). Hierarchical reinforcement learning with the MAXQ value function decomposition. Submitted for journal publication. Postscript Preprint.

March 10, 1999. Posted submissions to ICML-99:

• Margineantu, D., Dietterich, T. G. (1999). Learning Decision Trees for Loss Minimization in Multi-Class Problems. Technical report. Postscript file.
• Wang, X., Dietterich, T. G. (1999). Efficient Value Function Approximation Using Regression Trees. Technical report. Postscript file.

July 1, 1998. Posted a corrected version of the MAXQ paper that will appear in the ICML-98 proceedings:

• Dietterich, T. G. (to appear). The MAXQ method for hierarchical reinforcement learning. 1998 International Conference on Machine Learning. Postscript preprint. Note: This version has some errors corrected compared to the version that appears in the proceedings. In particular, Figure 1 is fixed.

June 17, 1998. Moved CS533 pages and updated entries for the reading group and the group meeting.

February 27, 1998. Posted two new papers (submissions to ICML98):

• Dietterich, T. G. (submitted). The MAXQ method for hierarchical reinforcement learning. Postscript preprint.

• Chown, E., Dietterich, T. G. (submitted). A divide-and-conquer approach to learning from prior knowledge. Postscript preprint.

February 3, 1998. Posted a new paper:

• Zhang, W., Dietterich, T. G. (submitted). Solving Combinatorial Optimization Tasks by Reinforcement Learning: A General Methodology Applied to Resource-Constrained Scheduling. Postscript preprint.

January 29, 1998. Posted a new paper:

• Dietterich, T. G., (submitted). An experimental comparison of three methods for constructing ensembles of decision trees: Bagging, boosting, and randomization. Postscript preprint.

January 29, 1998. Posted the final version of the statistics paper that will appear in Neural Computation:

• Dietterich, T. G., (In press). Approximate Statistical Tests for Comparing Supervised Classification Learning Algorithms. Neural Computation. Postscript preprint. (Revised December 30, 1997).

December 30, 1997. Moved CS430/530 pointer to Previous Courses section. Updated conference section. Updated citations for the following two papers:

• Chown, E., Dietterich, T. G. (submitted). A comparison of neural network and process-based models for vegetation distribution under global climate change. Postscript preprint.

• Dietterich, T. G., (1997). Machine Learning Research: Four Current Directions AI Magazine. 18 (4), 97-136. Postscript preprint.

November 20, 1997. Revised the research project descriptions and related materials. I also posted the slides and text for an ONR presentation.

October 20, 1997. Merged all of the publications into a single page sorted in reverse chronological order. I hope this will make it easier for people to search to find things.

October 4, 1997: Many changes--I've been lazy about keeping up this list.

• CS430/530 page added.
• Updated list of current students.
• Added Hierarchical RL MAXQ decomposition paper.
• Added Machine Learning Research: Four Current Directions preprint of a review article to appear in AI Magazine.

June 17, 1997: End of spring quarter. Moved CS534 pages to courseware section. Added pages for ONR Contractor's Meeting. Updated conference listing. Added publication:

• Kong, E. G., and Dietterich, T. G. (1997). Probability estimation using error-correcting output coding. IASTED International Conference: Artificial Intelligence and Soft Computing, Banff, Canada. Postscript preprint.

May 12, 1997: Updated the publications to reflect papers accepted to and rejected from ICML-97. Updated mailing addresses for Machine Learning Journal.

March 23, 1997: Created pages for CS534 (Machine Learning). These will be changing constantly throughout the quarter.

March 20, 1997: Posted old 1991 tech report:

• Dietterich, T. G. (1991). Do Hidden Units Implement Error-Correcting Codes?. Technical Report. Postscript file

March 13, 1997: Posted final version of EBRL paper:

• Dietterich, T. G., Flann, N. S. (In Press). Explanation-based Learning and Reinforcement Learning: A Unified View. Machine Learning . Postscript preprint.

March 5, 1997: Fixed a small bug in the Musk paper:

• Dietterich, T. G., Lathrop, R. H., Lozano-Perez, T. (1997) Solving the multiple-instance problem with axis-parallel rectangles. Artificial Intelligencem, 89(1-2), 31-71. Postscript preprint.

January 22, 1997: Posted 3 submissions to the Machine Learning Conference:

• Chown, E., Dietterich, T. G. (submitted). Learning in the Presence of Prior Knowledge: A Case Study Using Model Calibration. Submitted to the Fourteenth International Conference on Machine Learning. Postscript preprint.
• Margineantu, D., Dietterich, T. G. (submitted). Pruning Adaptive Boosting. Submitted to the Fourteenth International Conference on Machine Learning. Postscript preprint.
• Tadepalli, P., Dietterich, T. G. (submitted). Hierarchical Explanation-Based Reinforcement Learning. Submitted to the Fourteenth International Conference on Machine Learning. Postscript preprint.

January 20, 1997: Updated the publication information for

• Dietterich, T. G., Lathrop, R. H., Lozano-Perez, T. (1997) Solving the multiple-instance problem with axis-parallel rectangles. Artificial Intelligencem, 89(1-2), 31-71. Postscript preprint.

December 16, 1996: Updated research descriptions. Moved CS519 to Previous Courses section. Added Office Hours for Winter Quarter. Revised conference section to delete very old conferences.

October 31, 1996: Posted updated version of the Explanation-Based Reinforcement Learning paper:

• Dietterich, T. G., Flann, N. S., (submitted). Explanation-based Learning and Reinforcement Learning: A Unified View. Submitted to Machine Learning . Postscript preprint.

October 16, 1996: Posted the first version of a paper on statistical testing for learning algorithms:

• Dietterich, T. G., (1996). Statistical Tests for Comparing Supervised Classification Learning Algorithms. Submitted. Postscript preprint.

October 16, 1996: Changed office hours and course information for CS519.

July 11, 1996: Updated several references to include page numbers. Adjusted the list of conferences. Updated pointers to former students. Posted the final version of the multiple-instance learning paper:

• Dietterich, T. G., Lathrop, R. H., Lozano-Perez, T. (In Press) Solving the multiple-instance problem with axis-parallel rectangles. Artificial Intelligence. Postscript preprint (The last 3 figures are not available online).

April 5, 1996: Posted final version of ML96 paper with Kearns and Mansour and moved the Chown/Dietterich paper to the technical reports page.:

• Chown, E., Dietterich, T. G., (1996) Learning in the presence of prior knowledge: A case study using model calibration. Postscript file.
• Dietterich, T. G., Kearns, M., Mansour, Y., (to appear). Applying the weak learning framework to understand and improve C4.5. Accepted to ICML-96. Postscript preprint.

March 20, 1996: Moved CS450 to the previous courses area. Updated office hours for spring quarter, 1996.

January 19, 1996: Added submissions to ML96:

• Chown, E., Dietterich, T. G., (submitted). Learning in the presence of prior knowledge: A case study using model calibration. Submitted to ICML-96. Postscript preprint.
• Dietterich, T. G., Kearns, M., Mansour, Y., (submitted). Applying the weak learning framework to understand and improve C4.5. Submitted to ICML-96. Postscript preprint.

December 11, 1995: Removed old pages for CS450/550 from Spring 1995 and replaced them with new pages for CS450/550 for Winter, 1996. Changed office hours entry.

December 11, 1995: Added conference listing for Snowbird and AAAI Spring Symposium. Changed area codes for my telephone numbers to 541.

September 19, 1995: Created pages for CS 533, Fall 1995, Applied Artificial Intelligence for Engineeers.

July 19, 1995: Added pointer to Eric Chown's home page. Moved pointer to CS450/550. Updated the entries for the papers that appeared in the Machine Learning conference proceedings.

June 7, 1995: The CS450/550 pages are completed. Note particularly that I have made available MPEG movies from some of the course projects.

May 31, 1995: Changed CV to reflect promotion to Full Professor. Added references to two new papers:

• Zhang, W., Dietterich, T. G., (1995). Value Function Approximations and Job-Shop Scheduling. Submitted to the Workshop on Value Function Approximation in Reinforcement Learning at ICML-95. Postscript preprint.
• Zhang, W., Dietterich, T. G., (1995). High-Performance Job-Shop Scheduling With A Time-Delay TD(lambda) Network. Submitted to NIPS-95. Postscript preprint.

May, 1995: I've made many changes to the CS450 web pages. Too numerous to list here. I also added a link to the home page so that from the OSU machines you can access these pages through the local file system. This is much faster than going through the WWW server.

April 28, 1995: Reorganized some of the headings on my home page. Moved the CS450 page and installed many pages of information for that class. Corrected a bug in the solutions to Homework 3.

April 21, 1995: Added a page for CS450. CS 450/550 (Introduction to Computer Graphics). Also fixed a bug in a pointer on the journal articles page.

April 20, 1995: Final camera-ready versions of our three conference papers have been posted online:

• Dietterich, T. G., Flann, N. S., (To appear). Explanation-based Learning and Reinforcement Learning: A Unified View. Accepted to ML95. Poscript preprint. Long version submitted to Machine Learning
• Kong, E. B., Dietterich, T. G., (To appear). Error-Correcting Output Coding Corrects Bias and Variance. Accepted to ML95. Poscript preprint.
• Zhang, W., Dietterich, T. G., (To appear). A Reinforcement Learning Approach to Job-shop Scheduling. Accepted to IJCAI95. Postscript preprint.

April 5, 1995: Updated the conference paper listing to show the three papers that were accepted. Moved one paper to the Tech report page.

March 10, 1995: Kong has graduated. Updated paper pointers. Fixed a bug in the pointer to the ML Journal version of the ID3 vs. Backpropagation comparison paper. Many minor changes. Added pointer to our papers on Explanation-Based Reinforcement Learning: Full version submitted to Machine Learning . and Short version submitted to Machine Learning Conference. Added pointer to Monash University Minimum Message Length group .

February 1, 1995: Posted ML95 submissions. See Conference Papers.

January 12, 1995: Posted IJCAI paper. Added pointer to Wei Zhang's page.

January 3, 1995: Posted final version of JAIR ECOC paper. Postscript file.

January 3, 1995: Updated classes.

December 30, 1994: Updated classes and ofice hours for Winter Quarter.

December 14, 1994: Revised version of ECOC paper installed.

December 12, 1994: Updated conference pointers.

December 12, 1994: Added courseware from CS519 and courseware entry to home page.

December 7, 1994: Added this change history page.

December 7, 1994: Added reference and preprint for Bakiri and Dietterich, 1993 Cairo Conference paper.