Retrieve, Don't Retrain: Extending Vision Language Action Models to New Tasks at Test Time

Hugging Face Daily Papers Papers

Summary

This paper introduces a retrieval-augmented vision-language-action policy that eliminates per-task fine-tuning by using pre-trained models with indexed demonstrations, enabling efficient cross-embodiment generalization and task adaptation at test time.

Extending a vision-language-action (VLA) policy to a new task typically requires task-specific teleoperated demonstrations and per-task fine-tuning, making adaptation costly in both data collection and compute. In this paper, we show that this target-side per-task adaptation cost can be replaced by retrieval. Our retrieval-augmented policy is trained once on paired demonstrations from the target embodiment (query) and a cheaper embodiment (pool, e.g., human-hand video), then frozen. New tasks are added at deployment by appending pool-side demonstrations to a retrieval pool. The frozen policy conditions on retrieved trajectories at every control step, so new tasks are absorbed by indexing data rather than updating parameters. Fine-tuning is needed only to take on a new, unseen embodiment, not for each new task. We show that retrieval improves policies beyond a specific backbone, including standard VLA policies, but its effect is especially pronounced in Cosmos Policy, a video-generation-based world-action model (WAM). In this setting, retrieval supplies coarse task progression, while the WAM's future-image objective provides an additional visual consistency signal that strengthens the retrieval-conditioned actions. On PushT, we study how retrieval provides a reusable high-level motion prior for cross-embodiment generalization to unseen goal angles, while on RoboTwin 2.0 our method outperforms cross-embodiment baselines on unseen tasks, and we additionally demonstrate the method on a real robot.
Original Article
View Cached Full Text

Cached at: 06/16/26, 11:34 AM

Paper page - Retrieve, Don’t Retrain: Extending Vision Language Action Models to New Tasks at Test Time

Source: https://huggingface.co/papers/2606.15631 Published on Jun 14

·

Submitted byhttps://huggingface.co/Jeongeun

Parkon Jun 16

Abstract

Retrieval-augmented vision-language-action policies eliminate per-task fine-tuning costs by using pre-trained models with indexed demonstrations, enabling efficient cross-embodiment generalization and task adaptation.

Extending a vision-language-action (VLA) policy to a new task typically requires task-specificteleoperated demonstrationsandper-task fine-tuning, making adaptation costly in both data collection and compute. In this paper, we show that this target-side per-task adaptation cost can be replaced by retrieval. Ourretrieval-augmented policyis trained once on paired demonstrations from the target embodiment (query) and a cheaper embodiment (pool, e.g., human-hand video), then frozen. New tasks are added at deployment by appending pool-side demonstrations to aretrieval pool. Thefrozen policyconditions on retrieved trajectories at every control step, so new tasks are absorbed by indexing data rather than updating parameters. Fine-tuning is needed only to take on a new, unseen embodiment, not for each new task. We show that retrieval improves policies beyond a specific backbone, including standard VLA policies, but its effect is especially pronounced inCosmos Policy, avideo-generation-based world-action model(WAM). In this setting, retrieval supplies coarse task progression, while the WAM’sfuture-image objectiveprovides an additional visual consistency signal that strengthens the retrieval-conditioned actions. On PushT, we study how retrieval provides a reusable high-level motion prior forcross-embodiment generalizationto unseen goal angles, while on RoboTwin 2.0 our method outperforms cross-embodiment baselines on unseen tasks, and we additionally demonstrate the method on a real robot.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2606\.15631

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/2606.15631 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.15631 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.15631 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

Boosting Visual Instruction Tuning with Self-Supervised Guidance

Hugging Face Daily Papers

This paper proposes augmenting visual instruction tuning in multimodal language models with self-supervised tasks expressed as natural language instructions, improving vision-centric reasoning without additional architecture or annotations. By reformulating classical self-supervised pretext tasks as image-instruction-response triplets, the method achieves consistent performance improvements across multiple benchmarks by injecting only 3-10% visually grounded instructions into the training data.