TR2026-049
Directional Embedding Smoothing for Robust Vision Language Models
-
- , "Directional Embedding Smoothing for Robust Vision Language Models", International Conference on Learning Representations (ICLR) Workshop on Agents in the Wild, April 2026.BibTeX TR2026-049 PDF Presentation
- @inproceedings{Wang2026apr4,
- author = {{Wang, Ye and Liu, Jing and Koike-Akino, Toshiaki}},
- title = {{Directional Embedding Smoothing for Robust Vision Language Models}},
- booktitle = {International Conference on Learning Representations (ICLR) Workshop on Agents in the Wild},
- year = 2026,
- month = apr,
- url = {https://www.merl.com/publications/TR2026-049}
- }
- , "Directional Embedding Smoothing for Robust Vision Language Models", International Conference on Learning Representations (ICLR) Workshop on Agents in the Wild, April 2026.
-
MERL Contacts:
-
Research Area:
Abstract:
The safety and reliability of vision-language models (VLMs) are a crucial part of deploying trustworthy agentic AI systems. However, VLMs remain vulnerable to jailbreaking attacks that undermine their safety alignment to yield harmful outputs. In this work, we extend the Randomized Embedding Smoothing and Token Aggregation (RESTA) defense to VLMs and evaluate its performance against the JailBreakV-28K benchmark of multi-modal jailbreaking attacks. We find that RESTA is effective in reducing attack success rate over this diverse corpus of attacks, in particular, when employing directional embedding noise, where the injected noise is aligned with the original token embedding vectors. Our results demonstrate that RESTA can contribute to securing VLMs within agentic systems, as a lightweight, inference-time defense layer of an overall security framework.