IRIS: Intent Resolution via Inference-time Saccades for Open-Ended VQA in Large Vision-Language Models
This addresses ambiguity resolution in visual question answering for users of large vision-language models, offering a novel training-free approach with significant accuracy gains.
The paper tackled the problem of ambiguity in open-ended visual question answering by using real-time eye-tracking data to resolve intent, resulting in more than doubling accuracy on ambiguous questions from 35.2% to 77.2% while maintaining performance on unambiguous queries.
We introduce IRIS (Intent Resolution via Inference-time Saccades), a novel training-free approach that uses eye-tracking data in real-time to resolve ambiguity in open-ended VQA. Through a comprehensive user study with 500 unique image-question pairs, we demonstrate that fixations closest to the time participants start verbally asking their questions are the most informative for disambiguation in Large VLMs, more than doubling the accuracy of responses on ambiguous questions (from 35.2% to 77.2%) while maintaining performance on unambiguous queries. We evaluate our approach across state-of-the-art VLMs, showing consistent improvements when gaze data is incorporated in ambiguous image-question pairs, regardless of architectural differences. We release a new benchmark dataset to use eye movement data for disambiguated VQA, a novel real-time interactive protocol, and an evaluation suite.