MSLA is the first physically deployable attack that uses adversarial lighting to break semantic alignment in VLMs such as CLIP, LLaVA, and BLIP, causing classification failures and hallucinations in real scenes.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3verdicts
UNVERDICTED 3representative citing papers
KindSleep estimates AHI from oximetry signals with R²=0.917 by first learning interpretable clinical concepts then fusing them with multimodal data, outperforming prior methods on 9,815 patients.
SAIL integrates anatomical priors at the representation level with semantic features via fusion to produce more anatomically aligned attribution maps in OCT without altering existing explainability techniques.
citing papers explorer
-
Challenging Vision-Language Models with Physically Deployable Multimodal Semantic Lighting Attacks
MSLA is the first physically deployable attack that uses adversarial lighting to break semantic alignment in VLMs such as CLIP, LLaVA, and BLIP, causing classification failures and hallucinations in real scenes.
-
KindSleep: Knowledge-Informed Diagnosis of Obstructive Sleep Apnea from Oximetry
KindSleep estimates AHI from oximetry signals with R²=0.917 by first learning interpretable clinical concepts then fusing them with multimodal data, outperforming prior methods on 9,815 patients.
-
SAIL: Structure-Aware Interpretable Learning for Anatomy-Aligned Post-hoc Explanations in OCT
SAIL integrates anatomical priors at the representation level with semantic features via fusion to produce more anatomically aligned attribution maps in OCT without altering existing explainability techniques.