Residual Decoding: Mitigating Hallucinations in Large Vision-Language Models via History-Aware Residual Guidance
This addresses a critical issue for users of LVLMs by mitigating hallucinations, though it is an incremental improvement as it builds on existing decoding mechanisms.
The paper tackles the problem of hallucinations in Large Vision-Language Models (LVLMs), where generated content is coherent but irrelevant to visual input, by proposing Residual Decoding (ResDec), a training-free method that uses historical information to correct biases, resulting in significant improvements in visual grounding and reduced object hallucinations.
Large Vision-Language Models (LVLMs) can reason effectively from image-text inputs and perform well in various multimodal tasks. Despite this success, they are affected by language priors and often produce hallucinations. Hallucinations denote generated content that is grammatically and syntactically coherent, yet bears no match or direct relevance to actual visual input. To address this problem, we propose Residual Decoding (ResDec). It is a novel training-free method that uses historical information to aid decoding. The method relies on the internal implicit reasoning mechanism and token logits evolution mechanism of LVLMs to correct biases. Extensive experiments demonstrate that ResDec effectively suppresses hallucinations induced by language priors, significantly improves visual grounding, and reduces object hallucinations. In addition to mitigating hallucinations, ResDec also performs exceptionally well on comprehensive LVLM benchmarks, highlighting its broad applicability.