Call for Papers

While theorists and experimentalists share a common interest in advancing the field, their research objectives, methodologies, and challenges sometimes diverge significantly. This workshop aims to bridge this gap and to shed light on recent developments and synergies in both communities. Specifically, we aim to promote the following long-term desiderata.
  • Communicate existing results. As the field evolves rapidly, theorists and experimentalists often find themselves immersed in their domain, occasionally overlooking valuable insights and challenges encountered by the other. Participants will have the opportunity to present key findings, best practices, and lessons learned, emphasizing the importance of cross-disciplinary awareness. This proactive sharing of knowledge will help create a collaborative atmosphere that promotes a deeper appreciation for the existing works and encourages fruitful discussions on the current state and future directions in RL.
  • Identify new problem classes of practical interest. We aim to emphasize new structures and perspectives that have not been widely investigated yet. Experimentalists can present algorithms that work surprisingly well but lack theoretical understanding. Equally important are the cases where algorithms fail despite expectations. This collaboration will ensure that theoretical progress addresses the most compelling issues faced in practice and that advancements in empirical research will get the attention of theorists, creating a mutually beneficial exchange of ideas.
We invite submissions for the ICML 2024 workshop Aligning Reinforcement Learning Experimentalists and Theorists. Contributions can be theoretical and/or empirical, focusing on some of the topics mentioned below. Contributions from adjacent fields such as control are also welcome. The papers can present new work or summarize the recent work of the author(s). Papers submitted or accepted to other conferences or journals are also welcome. Submitted papers will be reviewed by the program committee. All accepted papers will have a poster presentation. Outstanding papers will also be considered for contributed talks of approximately 15 minutes each. Please note that at least one author of each accepted paper must be available for the poster and/or oral presentation. There will be no proceedings.

Important dates

Submission instructions

The page limit is 9 pages (excluding references and the appendix). Submissions may include supplementary material, but reviewers are only required to read the first 9 pages. Submissions should use the template provided by the adapted NeurIPS 2024 LaTeX style files. The reviewing process is double-blind. Parallel submissions (to a journal, conference, workshop, or preprint repository) and already published papers are allowed.


The workshop will cover a range of sub-topics including (but not limited to):

  • MDPs and Dynamic Programming
  • Temporal Difference Methods
  • Policy Optimization
  • Model-based RL and Planning
  • Exploration in RL
  • Offline RL
  • Unsupervised and Intrinsically Motivated RL
  • Representation Learning in RL
  • Lifelong and Non-stationary RL
  • Hierarchical RL
  • Partially Observable RL
  • Multi-Agent RL
  • Multi-Objective RL
  • Transfer and Meta RL
  • Control
  • Deep RL
  • Imitation Learning and Inverse RL
  • Risk-sensitive and Robust RL
  • Real-world applications of RL