CORTEXA
← Browse
arxivcs.ROcs.AIcs.CV2026-07-15

Semantic Anchoring for Robotic Action Representations

Yuan Xu, Youheng Shi, Chengyang Li, Wentao Zhu, Yizhou Wang

Vision-Language-Action (VLA) models inherit rich semantic representations from pretrained Vision-Language Models, yet fine-tuning on limited robot demonstrations degrades this structure and undermines generalization. A fundamental question therefore arises: what constitutes a good action representation? Inspired by the mirror neuron theory's insight that observation and execution share an intention-level encoding, we examine whether a robot's action representations preserve the semantic structure captured by pretrained encoders. Systematic probing confirms that this structure erodes during finetuning, and that its quality synchronizes with both task success and out-of-distribution generalization. We further introduce a plug-and-play method that anchors action representations to a semantic manifold while decomposing representations into a shared semantic channel and a private channel, all discarded at inference, leaving the deployed model unchanged. Validated on different VLA backbones across simulation and real-world benchmarks, our method yields up to +18.7% on real-world in-distribution tasks and +21.5% on out-of-distribution generalization.

View free PDFSource page

Related papers

arxivcs.AIcs.CVcs.RO2026-07-16

Action QFormer: Structured Representation Shaping under Action Supervision in Vision-Language-Action Models

Yufeng Ji, Wenhao Tang, Haoyi Niu, Koushil Sreenath, Yi Wu, Zhongyu Li

Action supervision in vision-language-action (VLA) models is often treated as a downstream objective for learning action prediction. In this paper, we study it instead as a force that shapes inherited multimodal representations. We show that this shaping has a dual effect: it is…

View free PDFSource page
arxivcs.CVcs.AIcs.RO2026-07-03

Token-Based Affordance Grounding with Large Vision-Language Models

Seung Il Lee, Qinqian Lei, Daguang Xu, Dong Yang, Robby T. Tan, Yixin Chen, et al.

Affordance grounding aims to localize image regions that support a specific action, serving as a core capability for physical intelligence and embodied perception. Previous studies have primarily relied on weakly supervised learning with action labels from exocentric images. Howe…

View free PDFSource page
arxivcs.ROcs.AIcs.CVcs.LG2026-07-12

Action Map Policy: Learning 3D Closed-loop Manipulation via Pixel Classification

Haojie Huang, Zhang Ye, Linfeng Zhao, Boce Hu, Mingxi Jia, Yu Qi, et al.

The action space poses a major challenge in robot learning, since it is often high-dimensional, can span long time horizons, and frequently admits multi-modal optimal solutions. A good choice of action representation and loss function can help to address these concerns, but there…

View free PDFSource page
arxivcs.CVcs.AIcs.LGcs.RO2026-07-06

From Fixed to Free Cameras: Calibration-Free View-Robust Vision-Language-Action Model

Wenhao Li, Xueying Jiang, Quanhao Qian, Deli Zhao, Shijian Lu, Gongjie Zhang, et al.

Real-world robot deployment rarely maintains the training-stage camera setup, where cameras often experience repositioning or remounting depending on actual scenarios. Existing view-robust Vision-Language-Action (VLA) policies tolerate such camera variations only when the camera…

View free PDFSource page
arxivcs.ROcs.AIcs.CV2026-07-10

TS-Mask VLA: 2D Temporal-Spatial Masking for Vision-Language-Action Model with Effective Bridging

Shengzhuo Yang, Ronghao Yu, Chuanjie Lv, Linpeng Peng, Hang Yu, Jie Ren, et al.

Vision-language-action (VLA) models aim to understand natural-language instructions and visual observations, and to generate and execute corresponding actions as embodied agents. Recently, autoregressive token-based action generation has driven the development of many representat…

View free PDFSource page
arxivcs.CVcs.AIcs.RO2026-07-02

VLAFlow: A Unified Training Framework for Vision-Language-Action Models via Co-training and Future Latent Alignment

Guoyang Xia, Fengfa Li, Hongjin Ji, Lei Ren, Fangxiang Feng, Kun Zhan, et al.

Vision-language-action models (VLAs) have recently advanced robotic manipulation, yet the effects of different robot-data pre-training paradigms remain difficult to compare because existing models often differ in architecture, data, action space, and evaluation protocol. We prese…

View free PDFSource page