Schedule

   Segment 1

 9:00 - 9:30  Opening remarks

 9:30 - 10:00 Invited talk 1. Sherry Yang, Google DeepMind/UC Berkeley. 

Decision Making with Internet Scale Knowledge.

10:00 - 10:30 Invited talk 2. Ida Momennejad, Microsoft Research.

RL-inspired Evaluation & Agentic Architecture with LLMs. 

              -- Coffee Break --

   Segment 2 

11:00 - 11:30 Invited talk 3. Michael Dennis, Google DeepMind.

Genie: Generative Interactive Environments. 

11:30 - 12:00 Invited talk 4. Roberta Raileanu, Meta GenAI.

Teaching Large Language Models When, Where and How to Refine their Reasoning.

              -- Lunch Break  --

   Segment 3

 1:00 - 1:30  Lightning spotlight talks

Transformers Learn Temporal Difference Methods for In-Context Reinforcement Learning. [pdf]

Policy Learning with a Language Bottleneck. [pdf]

Concept-Based Interpretable Reinforcement Learning with Limited to No Human Labels. [pdf]

Dynamics Generalisation with Behaviour Foundation Models. [pdf]

 1:30 - 2:30  Panel discussion. 

Training agents with foundation models, state-of-the-art, future directions, and best practices.

Participants: 

• Ida Momennejad (Microsoft Research)

• Michael Dennis (Google DeepMind)

• Roberta Raileanu (MetaGenAI).

      Chair: Mauricio Tec (Harvard University)

              -- Coffee Break  --

   Segment 4

 3:00 - 3:15  Closing remarks

 3:15 - 4:00  Poster session