Zhaohan Feng

Mixture-of-World Models: Scaling Multi-Task Reinforcement Learning with Modular Latent Dynamics

Self-predictive Mamba: Improving Multi-agent Reinforcement Learning with Self-predictive Encoding