Directly Forecasting Belief for Reinforcement Learning with Delays
This work addresses the problem of delayed sensory perceptions in RL for agents, offering a novel approach that improves learning efficiency and accuracy, though it is incremental in nature.
The paper tackles the challenge of compounding errors in reinforcement learning with delays by introducing DFBT, a method that directly forecasts states from observations, which reduces compounding errors and achieves superior performance on the MuJoCo benchmark compared to state-of-the-art baselines.
Reinforcement learning (RL) with delays is challenging as sensory perceptions lag behind the actual events: the RL agent needs to estimate the real state of its environment based on past observations. State-of-the-art (SOTA) methods typically employ recursive, step-by-step forecasting of states. This can cause the accumulation of compounding errors. To tackle this problem, our novel belief estimation method, named Directly Forecasting Belief Transformer (DFBT), directly forecasts states from observations without incrementally estimating intermediate states step-by-step. We theoretically demonstrate that DFBT greatly reduces compounding errors of existing recursively forecasting methods, yielding stronger performance guarantees. In experiments with D4RL offline datasets, DFBT reduces compounding errors with remarkable prediction accuracy. DFBT's capability to forecast state sequences also facilitates multi-step bootstrapping, thus greatly improving learning efficiency. On the MuJoCo benchmark, our DFBT-based method substantially outperforms SOTA baselines. Code is available at https://github.com/QingyuanWuNothing/DFBT.