HarmoWAM: Harmonizing Generalizable and Precise Manipulation via Adaptive World Action Models

TL;DR

HarmoWAM introduces an adaptive world model that unifies predictive and reactive control for robots, enabling both generalization and precise manipulation across unseen environments.

Contribution

The paper proposes HarmoWAM, a novel end-to-end world action model that adaptively combines predictive and reactive experts for improved robotic manipulation.

Findings

HarmoWAM achieves 33% and 29% improvements over prior models in zero-shot generalization.

It effectively handles variations in background, position, and object semantics in unseen environments.

The adaptive gating mechanism enables dynamic switching between control strategies.

Abstract

World Action Models (WAMs) have emerged as a promising paradigm for robot control by modeling physical dynamics. Current WAMs generally follow two paradigms: the "Imagine-then-Execute" approach, which uses video prediction to infer actions via inverse dynamics, and the "Joint Modeling" approach, which jointly models actions and video representations. Based on systematic experiments, we observe a fundamental trade-off between these paradigms: the former explicitly leverages world models for generalizable transit but lacks interaction precision, whereas the latter enables fine-grained, temporally coherent action generation but is constrained by the exploration space of the training distribution. Motivated by these findings, we propose HarmoWAM, an end-to-end WAM that fully leverages a world model to unify predictive and reactive control, enabling both generalizable transit and precise manipulation. Specifically, the world model provides spatio-temporal physical priors that condition two complementary action experts: a predictive expert that leverages latent dynamics for iterative action generation, and a reactive expert that directly infers actions from predicted visual evolution. To enable adaptive coordination, a Process-Adaptive Gating Mechanism is proposed to automatically determine the timing and location of switching between them. This allows the world model to drive the reactive expert to expand the exploration space and the predictive expert to perform precise interactions across different stages of a task. For evaluation, we construct three training-unseen test environments across six real-world robotic tasks, covering variations in background, position, and object semantics. Notably, HarmoWAM achieves strong zero-shot generalization across these scenarios, significantly outperforming prior state-of-the-art VLA models and WAMs by margins of 33% and 29%, respectively.