Track the frontier, from condensed matter to cosmology.
Search papers, labs, and topics across Lattice.
EVO-WAM enables robot world action models (WAMs) to iteratively adapt to unseen manipulation tasks purely through self-generated video-action rollouts, eliminating the need for expert demonstrations or active environment interaction. To prevent training on hallucinated or physically invalid data, the framework pairs state-augmented autoregressive rollout generation with a dual-stage filter that uses vision-language models to select task-completing trajectory prefixes and inverse dynamics models to enforce video-action consistency. Across unseen benchmarks, this zero-interaction self-training paradigm improves Cosmos3's success rate from 26.9% to 68.0% on RoboTwin 2.0 and from 20.0% to 76.7% on real-world long-horizon composite tasks.
Robots can adapt to unseen tasks purely through simulated imagination, boosting real-world manipulation success from 20% to 77% without requiring new human demonstrations or external environment rollouts.
Improving robot policies on new tasks without collecting additional expert demonstrations remains a central challenge in robot learning. World action models (WAMs) use broad video priors to jointly predict future videos and actions, offering a potential source of supervision for adapting to new tasks. However, generated videos may fail to depict task completion, and even visually successful videos may be paired with inconsistent actions that lead to execution failure. We propose EVO-WAM, a framework that adapts WAMs to unseen tasks by learning from their own generated video-action trajectories, without executing candidate actions in an external environment. First, we augment WAM training with state prediction and anchored multi-frame context to enable complete autoregressive rollouts without external execution feedback. Second, we identify reliable training experience by selecting task-completing prefixes with a vision-language model and verifying their video-action consistency with an inverse dynamics model. Third, we iteratively train the WAM on verified prefixes and generate new rollouts with the updated model. On seven unseen RoboTwin 2.0 tasks, EVO-WAM increases average success rates from 26.9% to 68.0% for Cosmos3 and from 28.5% to 46.4% for DreamZero, reaching approximately 2.5times and 1.6times their initial success rates. On three unseen long-horizon composite tasks in the real world, it improves Cosmos3's average success rate from 20.0% to 76.7%, a gain of 56.7 percentage points. Project Page: https://evo-wam.github.io/.