Mitigating Hallucinations in Video Large Language Models via Spatiotemporal-Semantic Contrastive Decoding
This addresses a critical issue for users relying on accurate video understanding in AI applications, but it is incremental as it builds on existing decoding methods.
The paper tackles the problem of hallucination in Video Large Language Models, where outputs are inconsistent with video content, by proposing a Spatiotemporal-Semantic Contrastive Decoding strategy that disrupts video features to suppress hallucinations, resulting in effective mitigation while preserving model capabilities.
Although Video Large Language Models perform remarkably well across tasks such as video understanding, question answering, and reasoning, they still suffer from the problem of hallucination, which refers to generating outputs that are inconsistent with explicit video content or factual evidence. However, existing decoding methods for mitigating video hallucinations, while considering the spatiotemporal characteristics of videos, mostly rely on heuristic designs. As a result, they fail to precisely capture the root causes of hallucinations and their fine-grained temporal and semantic correlations, leading to limited robustness and generalization in complex scenarios. To more effectively mitigate video hallucinations, we propose a novel decoding strategy termed Spatiotemporal-Semantic Contrastive Decoding. This strategy constructs negative features by deliberately disrupting the spatiotemporal consistency and semantic associations of video features, and suppresses video hallucinations through contrastive decoding against the original video features during inference. Extensive experiments demonstrate that our method not only effectively mitigates the occurrence of hallucinations, but also preserves the general video understanding and reasoning capabilities of the model.