paper-with-me

Papers

DREAMSTEER: Latent World Models Can Steer VLA Policies During Deployment Without Any Finetuning

2026-07-03 · Hanchen Cui, Sergio Arnaud, Arjun Majumdar, Daniel Dugas, Elie Aljalbout, Karthik Desingh, Krishna Murthy Jatavallabhula, Franziska Meier arxiv

Pretrained vision-language-action (VLA) policies show promising zero-shot generalization, but often fail under deployment-time distribution shift, leading to decreased robustness and inconsistent instruction following. While prior work commonly tackles this by finetuning on in-distribution data, it assumes demonstrations collected on tasks in the target environment. In this work, we propose DREAMSTEER, a deployment-time steering framework for pretrained VLAs without any finetuning or parameter modifications. The key insight in DREAMSTEER is to leverage a latent world model and a value model to steer pretrained VLA policies. During deployment, DREAMSTEER samples candidate action chunks from a VLA policy and predefined motion primitives, imagines their outcomes using an action-conditioned latent world model, and ranks the imagined trajectories with a language-conditioned value model. Across four real-world manipulation benchmarks with unseen objects, DREAMSTEER improves task success rate from 23.75% to 66.25% and instruction-following accuracy from 38.75% to 56.25% over the base VLA policy.

📄 PDF Abstract BibTeX arXiv:2607.02865

Code (0)

등록된 구현이 없습니다.

Tasks

Zero-shot GeneralizationInstruction Following

Similar Papers 제목 키워드 기반

DreamSteerer: Enhancing Source Image Conditioned Editability using Personalized Diffusion Models

2024-10-15 · Zhengyang Yu, Zhaoyuan Yang, Jing Zhang

Recent text-to-image personalization methods have shown great promise in teaching a diffusion model user-specified concepts given a few images for reusing the acquired concepts in a novel context. With massive efforts be…

Denoising

Latent Policy Steering through One-Step Flow Policies

2026-03-05 · Hokyun Im, Andrey Kolobov, Jianlong Fu, Youngwoon Lee arxiv

Offline reinforcement learning (RL) allows robots to learn from offline datasets without risky exploration. Yet, offline RL's performance often hinges on a brittle trade-off between (1) return maximization, which can pus…

Reinforcement LearningOffline RL

Beyond Action Residuals: Real-World Robot Policy Steering via Bottleneck Latent Reinforcement Learning

2026-05-19 · Dongjie Yu, Kun Lei, Zhennan Jiang, Jia Pan 외 arxiv

Pretrained imitation policies have become a strong foundation for robot manipulation, but they often require online improvement to overcome execution errors, limited dataset coverage, and deployment mismatch. A central q…

Reinforcement LearningRobot Manipulation

Steering Generative Reinforcement Learning into Stable Robotic Controller

2026-06-15 · Yixuan Wang, Shutong Ding, Ke Hu, Tianxiang Gui 외 arxiv

Diffusion and flow-based generative policies provide a powerful policy class for reinforcement learning by inducing rich stochastic exploration through iterative action generation. However, the stochasticity of diffusion…

Reinforcement Learning

Inference-time Policy Steering via Vision and Touch

2026-06-12 · Yilin Wu, Zilin Si, Zeynep Temel, Oliver Kroemer 외 arxiv

Inference-time steering adapts pre-trained generative robot policies during deployment by verifying candidate actions before execution. While prior methods typically perform this verification only with visual observation…