Mask World Model: Predicting What Matters for Robust Robot Policy Learning
For robot policy learning, MWM addresses the problem of overfitting to irrelevant visual factors by predicting masks, leading to more robust and generalizable control policies.
The Mask World Model (MWM) predicts semantic masks instead of RGB pixels, imposing a geometric information bottleneck that filters out visual noise. It significantly outperforms state-of-the-art RGB-based world models on LIBERO and RLBench benchmarks and shows robust generalization in real-world experiments.
World models derived from large-scale video generative pre-training have emerged as a promising paradigm for generalist robot policy learning. However, standard approaches often focus on high-fidelity RGB video prediction, this can result in overfitting to irrelevant factors, such as dynamic backgrounds and illumination changes. These distractions reduce the model's ability to generalize, ultimately leading to unreliable and fragile control policies. To address this, we introduce the Mask World Model (MWM), which leverages video diffusion architectures to predict the evolution of semantic masks instead of pixels. This shift imposes a geometric information bottleneck, forcing the model to capture essential physical dynamics and contact relations while filtering out visual noise. We seamlessly integrate this mask dynamics backbone with a diffusion-based policy head to enable robust end-to-end control. Extensive evaluations demonstrate the superiority of MWM on the LIBERO and RLBench simulation benchmarks, significantly outperforming the state-of-the-art RGB-based world models. Furthermore, real-world experiments and robustness evaluation (via random token pruning) reveal that MWM exhibits superior generalization capabilities and robust resilience to texture information loss.