← Back to all news
arXiv AI By Guangyu Zhao, Yongchang Zhang, Haoxuan Ru, Kaichen He, Haoran Wen, Kun Zhan, Zeyu Zhang, Yin Li, Yiwu Zhong

STAR: Steering Action Representations during Inference

Read the original on arXiv AI →

The paper introduces STAR (Steering Action Representations), a training‑free inference technique for vision‑language‑action models that addresses a late‑layer grounding gap. STAR applies an exponential moving average over late‑layer residual updates, enabling later action representations to inherit object‑aware update directions from earlier layers without changing the model. By incorporating recent proprioceptive states to adjust mixing strength, STAR improves object awareness in late layers and boosts average task success rates by 4.0 percentage points on robot manipulation benchmarks.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

Related stories

arXiv Computer Vision
Oct 16

VLA-$\Delta$: Learning Physical State Transition for Vision-Language-Action Models

VLA‑Δ introduces a framework that explicitly learns physical state transitions for vision‑language‑action models, using changes between current and future states as learning targets. The authors build a large‑scale pretraining dataset with physical state‑transition annotations that cover task‑relevant semantics and visual motion changes, providing direct supervision for continuous action generation. The resulting model performs competitively on LIBERO and RoboTwin benchmarks and outperforms π₀.₅ in real‑robot experiments, while requiring no extra inference overhead.

By Qi Lu, Jiarun Li, Ce Zhang, Miao Pan, Yi Pan, Xu Wang, Wenqi Zhang, Xuhong Zhang
roboticsmultimodalbenchmarks
More like this →
arXiv AI
Jun 16

Action with Visual Primitives

arXiv:2605. 22183v3 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models have emerged as a promising paradigm for generalist robotic manipulation.

By Weilong Guo, Yuchen Wang, Renping Zhou, Yunfeng Zhang, Rui Fang, Yuyang Pang, Wenda Xu, Gao Huang
roboticsmultimodal
More like this →
arXiv Computer Vision
Aug 27

V-Link: Recovering Lost Visual Representations in Action DiT for Vision-Language-Action Models

V-Link is a method designed to enhance Vision‑Language‑Action (VLA) models by recovering visual representations during the transfer from vision‑language (VL) features to action (A) features. It introduces complementary Spatial and Semantic Query representations that are injected into Action DiT through asymmetric pathways, providing both semantic augmentation and dedicated geometric conditioning for action generation. Experiments on LIBERO, LIBERO‑Plus, RoboTwin 2.0, and real‑world AGIBOT A3 Ultra tasks show significant performance gains over the base GR00T N1.6 model.

By Yehao Lu, Jiarui Yang, Yuning Su, Yufeng Xie, Yu Zhong, Yazhou Zhang, Haiyu Lan, Kaixiang Lu, Peiwen Lin, Chuang Wang, Zequn Qin, Enyu Li, Xi Li
roboticsmultimodal
More like this →
arXiv AI
Aug 18

NebulaVLA: A Dual-Frequency Vision-Language-Action Model With Guide Action for Robotic Manipulation

arXiv:2608. 16503v1 Announce Type: cross Abstract: Real-world deployment of Vision-Language-Action (VLA) models is often bottlenecked by efficiency-performance trade-offs, cross-embodiment generalization, and execution smoothness.

By Cong Zhao, Shuai Tian, Xu Zhang, Baocheng Ni, Xinguo Song, Xueying Sun, Shu Jiang, Shouchang Yang, Bo Tang, Jin Deng, Ge Zhu, YongCheng Wang, Jin Xu, Ri Yang
roboticsmultimodal
More like this →