7 papers
TrapVLA: Trapping Vision-Language-Action Models in Configured Failure Modes
Jun-Hui Liu, Kun-Yu Lin, Yi-Lin Wei +9
This work introduces Configured Failure Trapping, a novel backdoor attack task against Vision-Language-Action (VLA) models, which aims to activate attacks through stealthy textual…
A Closed-Loop Multi-Agent Framework for Robust Multi-Robot Manipulation
Yi-Xiang He, Lan Wei, Haoming Cen +6
Multi-robot systems provide the parallelism and redundancy necessary for long-horizon tasks, while Large Language Models (LLMs) offer the reasoning capabilities to decompose these…
RoboFlow4D: A Lightweight Flow World Model Toward Real-Time Flow-Guided Robotic Manipulation
Sixu Lin, Junliang Chen, Huaiyuan Xu +8
Planning and acting in 3D environments is a fundamental capability for robotic manipulation in the real world. Although prior work has explored predictive flow planners to guide 3D…
UT-ACA: Uncertainty-Triggered Adaptive Context Allocation for Long-Context Inference
Lang Zhou, Shuxuan Li, Zhuohao Li +3
Long-context inference remains challenging for large language models due to attention dilution and out-of-distribution degradation. Context selection mitigates this limitation by a…
Decomposed Object Manipulation via Dual-Actor Policy
Bin Fan, Jian-Jian Jiang, Zhuohao Li +5
Object manipulation, which focuses on learning to perform tasks on similar parts across different types of objects, can be divided into an approaching stage and a manipulation stag…
Hybrid Reward Normalization for Process-supervised Non-verifiable Agentic Tasks
Peiran Xu, Zhuohao Li, Xiaoying Xing +3
Large Language Models (LLMs) increasingly rely on external tools such as search engines to solve complex agentic tasks that require reasoning and external knowledge retrieval. Rece…