HarnessWAM: Bridging Prediction and Deliberation in World Action Models
2026-08-10 • Robotics
Robotics
AI summaryⓘ
The authors address a problem with World Action Models (WAMs), which can predict and control robot actions only over short times but struggle with longer, complex tasks. They introduce HarnessWAM, a new system that plans tasks using a vision-language model, keeps track of what the robot has done, and breaks big tasks into smaller, doable steps. HarnessWAM uses two feedback loops to monitor progress and adjust plans as needed, so the robot can recover from mistakes without losing its understanding of the environment. Their approach improves the robot's success rates on complex tasks compared to previous methods.
World Action Modelsembodied controlvision-language modeltask planningfeedback loopscene belieftask graphexecution recoveryatomic skillsrobotic manipulation
Authors
Zhaopeng Gu, Bingke Zhu, Tianxi Lin, Guibo Zhu, Yingying Chen, Kai Wang, Tingyu Yuan, Chaoyang Zhao, Zhaowen Li, Peng Su, Jinqiao Wang
Abstract
World Action Models (WAMs) jointly learn environmental dynamics and robot actions, introducing priors over physical evolution into embodied control. However, finite-horizon prediction and action generation are insufficient for complex embodied tasks that require global planning, cross-stage state maintenance, execution verification, and failure recovery. We refer to this mismatch as the prediction-deliberation gap of WAMs. To address this gap, we propose HarnessWAM, an agentic framework for WAMs. HarnessWAM employs a vision-language-model-based Task Manager to maintain an evidence-grounded scene belief and a structured task graph. A capability-conditioned executable-space projection further constrains open-ended semantic plans into sequences of atomic skills that satisfy task dependencies, embodiment-state constraints, and the capability boundary of the underlying WAM. During execution, HarnessWAM operates through an event-driven, dual-timescale feedback loop: a lightweight progress estimator continuously provides high-frequency execution evidence, while the Task Manager deliberates at salient milestones by jointly considering the current observation, task state, and interaction history to determine whether to advance the task, acquire additional observations, revise the plan, or initiate local recovery. This mechanism enables the robot to recover its state after a subtask failure and resume execution without discarding previously acquired scene knowledge. HarnessWAM achieves state-of-the-art full-task and subtask success rates of 59.6% and 69.9% on RoboMemArena, and an SR of 23.7% on RoboCerebra Ideal. These results demonstrate that model-external structured state maintenance and closed-loop agentic decision making can effectively extend the local control capabilities of WAMs into embodied task execution that is plannable, verifiable, and recoverable.