具身开发者社区

外观

论文arXiv 2610.1441518 天前

通过质量多样性环境生成对视觉-语言-动作模型进行红队测试 ​

针对VLA在轻微测试分布偏移下易失效，Q-DEG生成对抗环境并用成功轨迹微调，在仿真和sim-to-real中验证鲁棒性提升。

AI 整理
  • 仿真与训练
  • 数据与采集
  • 具身模型

原文链接：Red-Teaming of Vision-Language-Action Models via Quality Diversity Environment Generation（arXiv 2610.14415 · 发布日期：2026-10-10）

Red-Teaming of Vision-Language-Action Models via Quality Diversity Environment Generation

针对VLA在轻微测试分布偏移下易失效，Q-DEG生成对抗环境并用成功轨迹微调，在仿真和sim-to-real中验证鲁棒性提升。

核心要点 ​

  • VLA可通过大规模机器人示范学习，但常在物体位置、光照、纹理或相机姿态变化下失败。
  • Q-DEG用QD优化寻找高失败环境，并利用从VLA提取的策略嵌入促进环境多样性。
  • 方法通过对VLA自身成功rollout进行自模仿微调，缓解对抗环境暴露出的脆弱性。
  • 仿真中跨三个代表性VLA模型，相比基础模型和两种红队基线微调模型提升OOD鲁棒性。

验证范围与局限：验证包括仿真中的三个代表性VLA模型，并展示了对sim-to-real迁移的促进作用。

英文摘要 (Abstract) ​

A long-standing goal in robotics is to develop generalist robots that can perform tasks across a wide range of settings, flexibly specified with images and text. Vision-Language-Action (VLA) models offer a promising path toward this goal by training on large collections of robot demonstrations. However, while VLAs perform well in-distribution, they often fail under minor test-time shifts, such as changes in object positions, lighting, textures, and camera poses. To address this, we propose Quality Diversity for Diverse Environment Generation (Q-DEG), a method for improving VLA robustness by generating adversarial environments and fine-tuning the VLA on its own successful rollouts. Q-DEG uses Quality Diversity (QD) optimization to find a set of high-failure environments, while promoting diversity using policy embeddings extracted from the VLA. These environments expose vulnerabilities that can be mitigated through self-imitation fine-tuning, thereby improving the VLA's robustness. In simulation, we conduct experiments across three representative VLA models, and our results indicate that Q-DEG improves robustness to out-of-distribution (OOD) adversarial environments compared to the base VLA and VLAs fine-tuned using data collected from two red-teaming baselines. Furthermore, we demonstrate that improving VLA robustness through Q-DEG facilitates sim-to-real transfer. Our anonymous project website is at: https://qd-env.github.io/qdeg/.

作者：Shihan Zhao, Henry Chen, Sophie Hsu, Minjune Hwang, Siddharth Srikanth, Freddie Liang, Akanksha Saran, Aaquib Tabrez, Stefanos Z. Nikolaidis, Daniel Seita

来源

本页事实对应的公开页面，以原文为准。

  1. arXiv 2610.14415：Red-Teaming of Vision-Language-Action Models via Quality Diversity Environment Generation访问 2026-10-21