Acting Through Registers

Hugging Face Daily Papers Papers
robotics world-models action-learning registers flow-matching manipulation

Summary

The paper identifies persistent high-norm spatial outliers in World Action Model representations and introduces RegisterWAM, which adds learnable non-spatial registers paired with observation flow-matching loss supervision to reorganize action-relevant information, achieving 81.64% success on LIBERO-Plus and 93.0% on RoboTwin 2.0 without extra embodied pretraining data.

World Action Models (WAMs) use generative visual representations to condition robot actions, yet how these representations organize action-relevant information remains poorly understood. We identify persistent high-norm spatial outliers in WAM representations and find that their action relevance persists even after normalization and K/V projection attenuate their extreme magnitude. Motivated by this observation, we introduce RegisterWAM, which adds learnable non-spatial registers to the world model and uses them directly for action conditioning. Registers alone do not fully eliminate spatial outliers; pairing them with observation flow-matching (FM) loss supervision suppresses spatial outliers while concentrating high-norm responses in registers. Beyond this reorganization, observation FM loss supervision also improves action learning without registers, motivating an additional noised-observation FM objective that learns current-scene recovery from corrupted inputs. The resulting model achieves 81.64\% success on LIBERO-Plus without additional embodied pretraining data or foundation models, reaches 93.0\% average success on RoboTwin 2.0, and improves real-world manipulation performance.
Original Article
View Cached Full Text

Cached at: 10/15/26, 02:55 AM

Paper page - Acting Through Registers

Source: https://huggingface.co/papers/2610.17638

Abstract

WorldActionModels(WAMs)usegenerativevisualrepresentationstoconditionrobotactions,yethowtheserepresentationsorganizeaction-relevantinformationremainspoorlyunderstood.Weidentifypersistenthigh-normspatialoutliersinWAMrepresentationsandfindthattheiractionrelevancepersistsevenafternormalizationandK/Vprojectionattenuatetheirextrememagnitude.Motivatedbythisobservation,weintroduceRegisterWAM,whichaddslearnablenon-spatialregisterstotheworldmodelandusesthemdirectlyforactionconditioning.Registersalonedonotfullyeliminatespatialoutliers;pairingthemwithobservationflow-matching(FM)losssupervisionsuppressesspatialoutlierswhileconcentratinghigh-normresponsesinregisters.Beyondthisreorganization,observationFMlosssupervisionalsoimprovesactionlearningwithoutregisters,motivatinganadditionalnoised-observationFMobjectivethatlearnscurrent-scenerecoveryfromcorruptedinputs.Theresultingmodelachieves81.64\%successonLIBERO-Pluswithoutadditionalembodiedpretrainingdataorfoundationmodels,reaches93.0\%averagesuccessonRoboTwin2.0,andimprovesreal-worldmanipulationperformance.

View arXiv pageView PDFGitHubAdd to collection

Get this paper in your agent:

hf papers read 2610\.17638

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.17638 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.17638 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.17638 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.

Rolling-WAM: World Action Models with Rolling Imagination

Hugging Face Daily Papers

Rolling-WAM introduces a method to distribute denoising across replanning cycles in world action models for robotic manipulation, achieving a 4.5x speedup in replanning while maintaining competitive performance.

World Action Models: The Next Frontier in Embodied AI

Hugging Face Daily Papers

This survey paper introduces World Action Models (WAMs), a unified framework for embodied AI that integrates predictive state modeling with action generation. It provides a taxonomy of existing methods, analyzes the data ecosystem, and outlines evaluation protocols for this emerging paradigm.