Multi Time Scale World Models
This addresses a technical hurdle in AI for building agents that reason over complex, uncertain environments, though it appears incremental as it builds on existing probabilistic and state-space modeling approaches.
The paper tackles the problem of learning world models that operate at multiple temporal scales for accurate long-horizon predictions and uncertainty estimates, showing that their MTS3 model outperforms recent methods on system identification benchmarks.
Intelligent agents use internal world models to reason and make predictions about different courses of their actions at many scales. Devising learning paradigms and architectures that allow machines to learn world models that operate at multiple levels of temporal abstractions while dealing with complex uncertainty predictions is a major technical hurdle. In this work, we propose a probabilistic formalism to learn multi-time scale world models which we call the Multi Time Scale State Space (MTS3) model. Our model uses a computationally efficient inference scheme on multiple time scales for highly accurate long-horizon predictions and uncertainty estimates over several seconds into the future. Our experiments, which focus on action conditional long horizon future predictions, show that MTS3 outperforms recent methods on several system identification benchmarks including complex simulated and real-world dynamical systems. Code is available at this repository: https://github.com/ALRhub/MTS3.