視覚言語行動モデルのための安定した言語ガイダンス
Stable Language Guidance for Vision-Language-Action Models
VLAモデルが言語の言い回しの変化に弱い「モダリティ崩壊」を解決するため、意味意図と視覚的アフォーダンスを分離する確率的フレームワークRSSを提案し、敵対的な言語摂動下でも頑健な操作性能を実現した。
詳しい要約
1. どんなもの？
2. 先行研究と比べてどこがすごい？
3. 技術・手法の肝は？
4. どうやって有効だと検証した？
5. 議論はある？
6. 次に読むべき論文は？
※ AIが要旨から生成した要約です。正確性は原文をご確認ください。
著者: Zhihao Zhan, Yuhao Chen, Jiaying Zhou, Qinhan Lyu, Hao Liu, Keze Wang, Liang Lin, Guangrun Wang
分類: cs.RO, cs.CL
原文アブストラクト
Vision-Language-Action (VLA) models have demonstrated impressive capabilities in generalized robotic control; however, they remain notoriously brittle to linguistic perturbations. We identify a critical ``modality collapse'' phenomenon where strong visual priors overwhelm sparse linguistic signals, causing agents to overfit to specific instruction phrasings while ignoring the underlying semantic intent. To address this, we propose Residual Semantic Steering (RSS), a probabilistic framework that disentangles physical affordance from semantic execution. RSS introduces two theoretical innovations: (1) Monte Carlo Syntactic Integration, which approximates the true semantic posterior via dense, LLM-driven distributional expansion, and (2) Residual Affordance Steering, a dual-stream decoding mechanism that explicitly isolates the causal influence of language by subtracting the visual affordance prior. Theoretical analysis suggests that RSS effectively maximizes the mutual information between action and intent while suppressing visual distractors. Empirical results across diverse manipulation benchmarks demonstrate that RSS achieves state-of-the-art robustness, maintaining performance even under adversarial linguistic perturbations. We release our code at https://github.com/Doo-mon/RSS.
