Instruction-Aligned Visual Attention for Mitigating Hallucinations in Large Vision-Language Models
This addresses hallucinations in LVLMs for applications requiring accurate image descriptions, though it is an incremental improvement over existing decoding techniques.
The paper tackles the problem of hallucinations in Large Vision-Language Models (LVLMs) by proposing an Instruction-Aligned Visual Attention (IAVA) approach, which reduces over-attention to irrelevant image tokens and achieves improved performance on benchmarks like MME, POPE, and TextVQA.
Despite the significant success of Large Vision-Language models(LVLMs), these models still suffer hallucinations when describing images, generating answers that include non-existent objects. It is reported that these models tend to over-focus on certain irrelevant image tokens that do not contain critical information for answering the question and distort the output. To address this, we propose an Instruction-Aligned Visual Attention(IAVA) approach, which identifies irrelevant tokens by comparing changes in attention weights under two different instructions. By applying contrastive decoding, we dynamically adjust the logits generated from original image tokens and irrelevant image tokens, reducing the model's over-attention to irrelevant information. The experimental results demonstrate that IAVA consistently outperforms existing decoding techniques on benchmarks such as MME, POPE, and TextVQA in mitigating object hallucinations. Our IAVA approach is available online at https://github.com/Lee-lab558/IAVA.