<ul><li>Dyna-style off-policy model-based reinforcement learning (DMBRL) algorithms are facing a performance gap when applied across different benchmark environments.</li><li>While DMBRL algorithms perform well in OpenAI Gym, their performance drops significantly in DeepMind Control Suite (DMC) with proprioceptive observations.</li><li>Modern techniques designed to address issues in these settings do not consistently improve performance across all environments.</li><li>Adding synthetic rollouts to the training process, which is the backbone of Dyna-style algorithms, significantly degrades performance in most DMC environments.</li></ul>

Stealing That Free Lunch: Exposing the Limits of Dyna-Style Reinforcement Learning

Discover more