PAINT: Paying Attention to INformed Tokens to Mitigate Hallucination in Large Vision-Language Model

TL;DR

PAINT is a framework that reduces hallucinations in large vision-language models by selectively boosting attention to key visual tokens, improving caption accuracy without sacrificing performance.

Contribution

The paper introduces PAINT, a novel plug-and-play method that selectively enhances attention to important visual tokens to mitigate hallucinations in LVLMs.

Findings

Hallucinations arise from weakened attention to visual tokens in deeper layers.

Selective attention boosting to local and summary tokens significantly reduces hallucination rates.

PAINT achieves up to 62.3% reduction in hallucinations on MSCOCO dataset.

Abstract

Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities in understanding and describing visual content, achieving state-of-the-art performance across various vision-language tasks. However, these models often generate descriptions containing objects or details that are absent in the input image, a phenomenon commonly known as hallucination. Our work investigates the key reasons behind this issue by analyzing the pattern of self-attention in transformer layers. We find that hallucinations often arise from the progressive weakening of attention weight to visual tokens in the deeper layers of the LLM. Some previous works naively boost the attention of all visual tokens to mitigate this issue, resulting in suboptimal hallucination reduction. To address this, we identify two critical sets of visual tokens that facilitate the transfer of visual information from the vision encoder to the LLM. Local tokens encode grounded information about objects present in an image, while summary tokens capture the overall aggregated representation of the image. Importantly, these two sets of tokens require different levels of weight enhancement. To this end, we propose \textbf{PAINT} (\textbf{P}aying \textbf{A}ttention to \textbf{IN}formed \textbf{T}okens), a plug-and-play framework that intervenes in the self-attention mechanism of the LLM, selectively boosting the attention weights of local and summary tokens with experimentally learned margins. Evaluation on the MSCOCO image captioning dataset demonstrate that our approach reduces hallucination rates by up to 62.3\% compared to baseline models while maintaining accuracy. Code is available at \href{https://github.com/hasanar1f/PAINT}{https://github.com/hasanar1f/PAINT}