Shailesh Kumar
Masters Alumni
Shailesh worked with NNRG on reinforcement learning algorithms for packet routing. He went on to get his PhD in the UT Electrical Engineering Department, with Joydeep Ghosh as his advisor. After graduation he went to work in Joseph Sirosh's group at Fair Isaac, Inc.
Homepage:
http://research.google.com/pubs/ShaileshKumar.html
Publications
Confidence Based Dual Reinforcement Q-Routing: An Adaptive On-Line Routing Algorithm 1999
Shailesh Kumar and Risto Miikkulainen, In 16th International Joint Conference on Artificial Intelligence (IJCAI-99), pp. 758--763, Stockholm, Sweden 1999. San Francisco, CA: Kaufmann.
Confidence Based Dual Reinforcement Q-Routing: An On-Line Adaptive Network Routing Algorithm 1998
Shailesh Kumar, Masters Thesis, Department of Computer Sciences, the University of Texas at Austin.. 108. Technical Report AI-98-267.
Confidence Based Q-Routing: An On-Line Adaptive Network Routing Algorithm 1998
Shailesh Kumar and Risto Miikkulainen, Smart Engineering Systems: Neural Networks, Fuzzy Logic, Data Mining, and Evolutionary ProgrammingC. H. Dagli and M. Akay and O. Ersoy and B. R. Fernandez and A. Smith (Eds.), Vol. 8 (1998).
Dual Reinforcement Q-Routing: An On-Line Adaptive Routing Algorithm 1997
Shailesh Kumar and Risto Miikkulainen, Smart Engineering Systems: Neural Networks, Fuzzy Logic, Data Mining, and Evolutionary ProgrammingC. H. Dagli, M. Akay, O. Ersoy, B. R. Fernandez and A. Smith (Eds.), Vol. 7 (1997).
On-Line Adaptation Of A Signal Predistorter Through Dual Reinforcement Learning 1996
Patrick Goetz, Shailesh Kumar and Risto Miikkulainen, In Machine Learning: Proceedings of the 13th Annual Conference (Bari, Italy), Lorenza Saitta (Eds.), pp. 175-181 1996. San Francisco, CA: Morgan Kaufmann.
Projects
Adaptive Packet Routing: The Confidence-Based Dual Reinforcement Q-Learning Algorithm 1998 - 2000
On-Line Adaptation of a Signal Predistorter through Dual Reinforcement Learning 1995 - 1997
Areas of Interest
Applications Reinforcement Learning
Labs
Formerly affiliated with Neural Networks