"RL in Big Worlds" Workshop @RLC2026

[Théo VINCENT]

Hi everyone,

We are organizing a workshop at RLC this summer about learning methods that are designed to control agents in big worlds.

A big world is an environment that is more complex than the learning agent. This greater complexity can be materialized by having the environment use more resources than the agent. In big worlds, the agent only observes a small portion of the full state.

Those constraints are imposed to better represent the challenges encountered when learning in real-world scenarios, with the hope of one day, enabling test-time learning. More details can be found in our call for papers: https://rlinbigworlds.ca/call_for_papers.html.

The workshop will take place on August 15th in Montreal.

Submissions are open at https://openreview.net/group?id=rl-conference.cc/RLC/2026/Workshop/RL_in_Big_Worlds until May 15th. You can register as reviewers by filling this form: https://forms.gle/haecvNnAmjTAQztt8.

Best,
The organization team
Alex Levandowski, Carlo D'Eramo, Esraa Elelimy, Khurram Javed, Kris De Asis, Théo Vincent,
Jan Peters, and Richard Sutton