Imagen de Portada

VLA-OPD: Bridging Offline SFT and Online RL for Vision-Language-Action Models via On-Policy Distillation

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Zhong, Zhide, Yan, Haodong, Li, Junfeng, He, Junjie, Zhang, Tianran, Li, Haoang
Formato:	Preprint
Publicado:	2026
Materias:	Robotics
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

FlowVLA: Visual Chain of Thought-based Motion Reasoning for Vision-Language-Action Models
por: Zhong, Zhide, et al.
Publicado: (2025)

DualCoT-VLA: Visual-Linguistic Chain of Thought via Parallel Reasoning for Vision-Language-Action Models
por: Zhong, Zhide, et al.
Publicado: (2026)

PD-VLA: Accelerating Vision-Language-Action Model Integrated with Action Chunking via Parallel Decoding
por: Song, Wenxuan, et al.
Publicado: (2025)

OmniVLA-RL: A Vision-Language-Action Model with Spatial Understanding and Online RL
por: Jie, Haoxiang, et al.
Publicado: (2026)

S-VAM: Shortcut Video-Action Model by Self-Distilling Geometric and Semantic Foresight
por: Yan, Haodong, et al.
Publicado: (2026)

ReconVLA: Reconstructive Vision-Language-Action Model as Effective Robot Perceiver
por: Song, Wenxuan, et al.
Publicado: (2025)

CEED-VLA: Consistency Vision-Language-Action Model with Early-Exit Decoding
por: Song, Wenxuan, et al.
Publicado: (2025)

RationalVLA: A Rational Vision-Language-Action Model with Dual System
por: Song, Wenxuan, et al.
Publicado: (2025)

Unified Diffusion VLA: Vision-Language-Action Model via Joint Discrete Denoising Diffusion Process
por: Chen, Jiayi, et al.
Publicado: (2025)

DyGeoVLN: Infusing Dynamic Geometry Foundation Model into Vision-Language Navigation
por: Liu, Xiangchen, et al.
Publicado: (2026)

CoA-VLA: Improving Vision-Language-Action Models via Visual-Textual Chain-of-Affordance
por: Li, Jinming, et al.
Publicado: (2024)

StereoVLA: Enhancing Vision-Language-Action Models with Stereo Vision
por: Deng, Shengliang, et al.
Publicado: (2025)

ACoT-VLA: Action Chain-of-Thought for Vision-Language-Action Models
por: Zhong, Linqing, et al.
Publicado: (2026)

PointVLA: Injecting the 3D World into Vision-Language-Action Models
por: Li, Chengmeng, et al.
Publicado: (2025)

RL Token: Bootstrapping Online RL with Vision-Language-Action Models
por: Xu, Charles, et al.
Publicado: (2026)

PAPO-VLA: Planning-Aware Policy Optimization for Vision-Language-Action Models
por: Guo, Peizheng, et al.
Publicado: (2026)

Refined Policy Distillation: From VLA Generalists to RL Experts
por: Jülg, Tobias, et al.
Publicado: (2025)

DFM-VLA: Iterative Action Refinement for Robot Manipulation via Discrete Flow Matching
por: Chen, Jiayi, et al.
Publicado: (2026)

FD-VLA: Force-Distilled Vision-Language-Action Model for Contact-Rich Manipulation
por: Zhao, Ruiteng, et al.
Publicado: (2026)

NS-VLA: Towards Neuro-Symbolic Vision-Language-Action Models
por: Zhu, Ziyue, et al.
Publicado: (2026)

Decoupling KL and Trajectories: A Unified Perspective for SFT, DAgger, Offline RL, and OPD in LLM Distillation
por: Zhao, Anhao, et al.
Publicado: (2026)

RotVLA: Rotational Latent Action for Vision-Language-Action Model
por: Li, Qiwei, et al.
Publicado: (2026)

FutureVLA: Joint Visuomotor Prediction for Vision-Language-Action Model
por: Xu, Xiaoxu, et al.
Publicado: (2026)

IRL-VLA: Training an Vision-Language-Action Policy via Reward World Model
por: Jiang, Anqing, et al.
Publicado: (2025)

Agentic-VLA: Efficient Online Adaptation for Vision-Language-Action Models
por: Jin, Ruofan, et al.
Publicado: (2026)

VLA-Reasoner: Empowering Vision-Language-Action Models with Reasoning via Online Monte Carlo Tree Search
por: Guo, Wenkai, et al.
Publicado: (2025)

X-DiffVLA: X-Embodied Diffusion Action Heads for Vision-Language-Action Models
por: Li, Boyu, et al.
Publicado: (2026)

Rethinking the Practicality of Vision-language-action Model: A Comprehensive Benchmark and An Improved Baseline
por: Song, Wenxuan, et al.
Publicado: (2026)

AT-VLA: Adaptive Tactile Injection for Enhanced Feedback Reaction in Vision-Language-Action Models
por: Li, Xiaoqi, et al.
Publicado: (2026)

CollabVLA: Self-Reflective Vision-Language-Action Model Dreaming Together with Human
por: Sun, Nan, et al.
Publicado: (2025)

Counterfactual VLA: Self-Reflective Vision-Language-Action Model with Adaptive Reasoning
por: Peng, Zhenghao "Mark", et al.
Publicado: (2025)

RynnVLA-002: A Unified Vision-Language-Action and World Model
por: Cen, Jun, et al.
Publicado: (2025)

RePO-VLA: Recovery-Driven Policy Optimization for Vision-Language-Action Models
por: Liufu, Weijia, et al.
Publicado: (2026)

D-VLA: A High-Concurrency Distributed Asynchronous Reinforcement Learning Framework for Vision-Language-Action Models
por: Guo, Yucheng, et al.
Publicado: (2026)

CogVLA: Cognition-Aligned Vision-Language-Action Model via Instruction-Driven Routing & Sparsification
por: Li, Wei, et al.
Publicado: (2025)

UrbanVLA: A Vision-Language-Action Model for Urban Micromobility
por: Li, Anqi, et al.
Publicado: (2025)

SG-VLA: Learning Spatially-Grounded Vision-Language-Action Models for Mobile Manipulation
por: Tu, Ruisen, et al.
Publicado: (2026)

DAM-VLA: A Dynamic Action Model-Based Vision-Language-Action Framework for Robot Manipulation
por: Peng, Xiongfeng, et al.
Publicado: (2026)

HyperVLA: Efficient Inference in Vision-Language-Action Models via Hypernetworks
por: Xiong, Zheng, et al.
Publicado: (2025)

AIR-VLA: Vision-Language-Action Systems for Aerial Manipulation
por: Sun, Jianli, et al.
Publicado: (2026)