In many settings such as education, healthcare, drug design, robotics, transportation, and achieving better-than-human performance in strategic games, it is important to make decisions sequentially. This poses two interconnected algorithmic and statistical challenges: effectively exploring to learn information about the underlying dynamics and effectively planning using this information. Reinforcement Learning (RL) is the main paradigm tackling both of these challenges simultaneously which is essential in the aforementioned applications. Over the last years, reinforcement learning has seen enormous progress both in solidifying our understanding on its theoretical underpinnings and in applying these methods in practice.
This workshop aims to highlight recent theoretical contributions, with an emphasis on addressing significant challenges on the road ahead. Such theoretical understanding is important in order to design algorithms that have robust and compelling performance in real-world applications. As part of the ICML 2020 conference, this workshop will be held virtually. It will feature keynote talks from six reinforcement learning experts tackling different significant facets of RL. It will also offer the opportunity for contributed material (see below the call for papers and our outstanding program committee). The authors of each accepted paper will prerecord a 10-minute presentation and will also appear in a poster session. Finally, the workshop will have a panel discussing important challenges in the road ahead.
University of Washington
Microsoft Research NYC
Research Assistant Professor
Universitat Pompeu Fabra
University of Alberta / DeepMind
University of Alberta
We invite submissions tackling hurdles in our theoretical understanding of reinforcement learning. Relevant submissions include (but are not limited to) classical topics such as sample-efficient exploration, off-policy learning, policy gradient methods, representation learning, transfer learning in RL. We are particularly interested in submissions which aim to broaden the range of problem settings and environments under which we have theoretical understanding:
Finally, we strongly encourage submissions that explore interdisciplinary connections of RL to other areas such as:
The papers should be 4 pages in ICML 2020 format excluding references (there can be unlimited appendix but reviewers are only required to read the first 4 pages). The submissions will be single blind. The papers will be evaluated with respect to four criteria and we encourage authors to make sure that their submissions make these contributions clear. First, the paper should be within the (broadly defined) scope of the workshop. Second, the paper should explicitly motivate the question it poses. Third, the paper should adequately contrast to prior work and explain the fundamental limitations preventing previous techniques from solving that question. Finally, the paper should provide in a crisp way the key theoretical idea that allows to address the limitation and make progress in our understanding of the question.
Papers accepted to ICML 2020 will not be considered. However, we encourage submission of recent papers accepted in other conferences, especially those drawing interdisciplinary connections.
We thank Hoang M. Le from providing the website template.