Leslie Pack Kaelbling is an American scientist who studies robots. She is the Panasonic Professor of Computer Science and Engineering at the Massachusetts Institute of Technology. She is well known for using a type of math problem called partially observable Markov decision processes, which were originally developed in operations research, to help improve artificial intelligence and robotics. In 1997, she won the IJCAI Computers and Thought Award for using a method called reinforcement learning to help robots learn how to control their movements and for creating tools that help robots navigate their environments. In 2000, she was chosen to be a Fellow of the Association for the Advancement of Artificial Intelligence.

Career

Kaelbling earned a bachelor's degree in Philosophy in 1983 and a doctorate in Computer Science in 1990, both from Stanford University. During this time, she worked with the Center for the Study of Language and Information. Later, she was employed by SRI International and a robotics company called Teleos Research, which was connected to SRI International. She then became a professor at Brown University. In 1999, she moved to MIT to join its faculty. Her research includes how machines make decisions when they do not have all the information, teaching machines to learn, and how machines sense their environment, all applied to robotics.

Journal of Machine Learning Research

In the spring of 2000, she and two-thirds of the editorial board of the Kluwer-owned journal Machine Learning resigned to show disagreement with the journal’s pay-to-access archives and its small payments to authors. Kaelbling helped start and was the first editor-in-chief of the Journal of Machine Learning Research, a peer-reviewed open access journal on similar topics. This journal allows researchers to publish articles for free, keeps the rights to their work, and makes all archives freely available online. After the resignations, Kluwer changed its policy to let authors upload their papers online after peer review. Kaelbling said this new policy was reasonable and would have avoided the need for an alternative journal. However, the editorial board had clearly stated they wanted this change, and the policy only changed after the resignations and the founding of JMLR.

Selected works

• Reinforcement Learning: A Survey (LP Kaelbling, ML Littman, AW Moore). Journal of Artificial Intelligence Research (JAIR) 4 (1996) 237-285. A widely referenced overview of reinforcement learning.
• Planning and acting in partially observable stochastic domains (LP Kaelbling, ML Littman, AR Cassandra). Artificial Intelligence 101 (1), 99-134.
• Acting under uncertainty: Discrete Bayesian models for mobile-robot navigation (AR Cassandra, LP Kaelbling, JA Kurien). Intelligent Robots and Systems (2) 963-972.
• The synthesis of digital machines with provable epistemic properties (SJ Rosenschein, LP Kaelbling). Proceedings of the 1986 Conference on Theoretical Aspects of Reasoning about Knowledge, 83-98.
• Practical reinforcement learning in continuous spaces (WD Smart, LP Kaelbling). 2000 International Conference on Machine Learning (ICML), 903-910.
• Hierarchical task and motion planning in the now (LP Kaelbling, T Lozano-Pérez). 2011 IEEE International Conference on Robotics and Automation (ICRA), 1470-1477.