WAVE-0: Bridging VLA Policies and WAMs with Lightweight Latent Future Prediction

Hugging Face Daily Papers Papers
robotics vision-language-action world-models manipulation latent-planning imitation-learning

Summary

WAVE-0 提出一个轻量级潜在世界模型，将 VLA 策略与世界动作模型（WAM）结合，使预测的动作后果直接引导动作生成，在 RoboTwin2.0 和 LIBERO 系列基准上取得 91.47%/98.7% 的成功率，并将 π0.5 的真实任务成功率从 38.3% 提升至 73.3%。

Vision-language-action (VLA) policies excel at generalist robotic manipulation but generate actions without explicitly modeling their environmental consequences. World action models (WAMs) incorporate future prediction, yet most rely on costly video generation, while latent alternatives either use predicted futures only as auxiliary targets or infer latent actions from suboptimal visual transitions through costly large-scale video pretraining. We introduce WAVE-0 (World-model-Augmented VLA Execution), a framework that bridges VLAs and WAMs through a lightweight latent world model. By connecting the semantic understanding of pretrained VLAs with explicit future prediction, WAVE-0 allows anticipated action consequences to directly guide action generation. Built on a pretrained VLA, its 257M world model translates VLM-generated latent action plans into predicted future representations that condition the action expert, unifying latent planning, future prediction, and action generation. The world model is trained directly on ground-truth robot actions with a counterfactual ranking objective, requiring neither pixel reconstruction nor large-scale video pretraining. WAVE-0 achieves success rates of 91.47% on RoboTwin 2.0 and 98.7% on LIBERO, along with 88.0%, 61.4%, and 75.0% on LIBERO-Plus, LIBERO-Pro, and LIBERO-Para, respectively. On real-world tasks, it improves the success rate of π_{0.5} from 38.3% to 73.3%. These results demonstrate strong manipulation performance and generalization across diverse distribution shifts.
Original Article
View Cached Full Text

Cached at: 10/14/26, 06:50 AM

Paper page - WAVE-0: Bridging VLA Policies and WAMs with Lightweight Latent Future Prediction

Source: https://huggingface.co/papers/2610.18568

Abstract

Vision-language-action(VLA)policiesexcelatgeneralistroboticmanipulationbutgenerateactionswithoutexplicitlymodelingtheirenvironmentalconsequences.Worldactionmodels(WAMs)incorporatefutureprediction,yetmostrelyoncostlyvideogeneration,whilelatentalternativeseitherusepredictedfuturesonlyasauxiliarytargetsorinferlatentactionsfromsuboptimalvisualtransitionsthroughcostlylarge-scalevideopretraining.WeintroduceWAVE-0(World-model-AugmentedVLAExecution),aframeworkthatbridgesVLAsandWAMsthroughalightweightlatentworldmodel.ByconnectingthesemanticunderstandingofpretrainedVLAswithexplicitfutureprediction,WAVE-0allowsanticipatedactionconsequencestodirectlyguideactiongeneration.BuiltonapretrainedVLA,its257MworldmodeltranslatesVLM-generatedlatentactionplansintopredictedfuturerepresentationsthatconditiontheactionexpert,unifyinglatentplanning,futureprediction,andactiongeneration.Theworldmodelistraineddirectlyonground-truthrobotactionswithacounterfactualrankingobjective,requiringneitherpixelreconstructionnorlarge-scalevideopretraining.WAVE-0achievessuccessratesof91.47%onRoboTwin2.0and98.7%onLIBERO,alongwith88.0%,61.4%,and75.0%onLIBERO-Plus,LIBERO-Pro,andLIBERO-Para,respectively.Onreal-worldtasks,itimprovesthesuccessrateofπ_{0.5}from38.3%to73.3%.Theseresultsdemonstratestrongmanipulationperformanceandgeneralizationacrossdiversedistributionshifts.

View arXiv pageView PDFProject pageGitHub4Add to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.18568 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.18568 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.18568 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

World Action Modeling with Progressive Visual Planning

arXiv cs.AI

ProWAM is a progressive world action model that jointly predicts robot actions and an ordered sequence of sparse visual sub-goals, enabling efficient long-horizon control without dense video generation. It achieves state-of-the-art results on LIBERO-Plus and RoboTwin, plus 70% zero-shot real-world success.

RealtimeWAM: One-Step Asynchronous World Action Models

Hugging Face Daily Papers

RealtimeWAM proposes one-step asynchronous World Action Models using Teacher-Anchored Consistency Distillation (TACD) and Cross-Expert Wavefront Pipelining (CEWP), achieving near-lossless performance (<1% drop) on LIBERO, LIBERO-Plus, and RoboTwin with up to 25x end-to-end speedup on H100.