Scaling Tasks, Not Samples: Mastering Humanoid Control through Multi-Task Model-Based Reinforcement Learning
This addresses the problem of scalable online learning for robotics, offering a more efficient approach than current methods focused on parameter or data scaling, though it is incremental in applying multi-task learning to model-based reinforcement learning.
The paper tackles the challenge of developing generalist robots by proposing to scale the number of tasks rather than samples per task, using multi-task model-based reinforcement learning, and achieves state-of-the-art performance on the HumanoidBench benchmark with higher sample efficiency than baselines.
Developing generalist robots capable of mastering diverse skills remains a central challenge in embodied AI. While recent progress emphasizes scaling model parameters and offline datasets, such approaches are limited in robotics, where learning requires active interaction. We argue that effective online learning should scale the \emph{number of tasks}, rather than the number of samples per task. This regime reveals a structural advantage of model-based reinforcement learning (MBRL). Because physical dynamics are invariant across tasks, a shared world model can aggregate multi-task experience to learn robust, task-agnostic representations. In contrast, model-free methods suffer from gradient interference when tasks demand conflicting actions in similar states. Task diversity therefore acts as a regularizer for MBRL, improving dynamics learning and sample efficiency. We instantiate this idea with \textbf{EfficientZero-Multitask (EZ-M)}, a sample-efficient multi-task MBRL algorithm for online learning. Evaluated on \textbf{HumanoidBench}, a challenging whole-body control benchmark, EZ-M achieves state-of-the-art performance with significantly higher sample efficiency than strong baselines, without extreme parameter scaling. These results establish task scaling as a critical axis for scalable robotic learning. The project website is available \href{https://yewr.github.io/ez_m/}{here}.