HaloQuest: A Visual Hallucination Dataset for Advancing Multimodal Reasoning

TL;DR

HaloQuest is a new multimodal hallucination dataset that uses synthetic and real images to evaluate and improve vision-language models, revealing current models' struggles and proposing new evaluation methods.

Contribution

Introduces HaloQuest, a large-scale dataset with synthetic images for benchmarking and fine-tuning VLMs to reduce hallucination in multimodal reasoning.

Findings

Current VLMs achieve below 36% accuracy on HaloQuest.

Fine-tuning on HaloQuest reduces hallucination without harming standard reasoning.

Generated images correlate highly (r=0.97) with real images in benchmarking.

Abstract

Hallucination has been a major problem for large language models and remains a critical challenge when it comes to multimodality in which vision-language models (VLMs) have to deal with not just textual but also visual inputs. Despite rapid progress in VLMs, resources for evaluating and addressing multimodal hallucination are limited and mostly focused on evaluation. This work introduces HaloQuest, a novel visual question answering dataset that captures various aspects of multimodal hallucination such as false premises, insufficient contexts, and visual challenges. A novel idea from HaloQuest is to leverage synthetic images, apart from real ones, to enable dataset creation at scale. With over 7.7K examples spanning across a wide variety of categories, HaloQuest was designed to be both a challenging benchmark for VLMs and a fine-tuning dataset for advancing multimodal reasoning. Our experiments reveal that current models struggle with HaloQuest, with all open-source VLMs achieving below 36% accuracy. On the other hand, fine-tuning on HaloQuest significantly reduces hallucination rates while preserving performance on standard reasoning tasks. Our results discover that benchmarking with generated images is highly correlated (r=0.97) with real images. Last but not least, we propose a novel Auto-Eval mechanism that is highly correlated with human raters (r=0.99) for evaluating VLMs. In sum, this work makes concrete strides towards understanding, evaluating, and mitigating hallucination in VLMs, serving as an important step towards more reliable multimodal AI systems in the future.