No AI summary available for this article.
Why It Matters
Vision-language models (VLMs) are fragile under image corruption.
Provenance
Discovered via ArXiv and published by ArXiv.
Key Claims
Original description
Vision-language models (VLMs) are fragile under image corruption. We find that the wording of the question affects VLMs in two opposite ways. Verbose questions make VLMs substantially more robust---e.g., rephrasing "Is there a cat?" into "Please look carefully and answer: is there a cat?". Conversely, VLMs become more fragile under corruption when the question is semantically complex or finer-grained, e.g., "what colour is the cup left of the chair?" instead of "is there a cup?". Both effects stem from question-conditioned cross-modal attention, which induces a spectral filter over image patch...
Discovered via ArXiv
Research papers and preprints from arXiv.
Publisher: arxiv.org
ID: http://arxiv.org/abs/2609.20139v1 · Indexed about 1 hour ago