History-Guided Iterative Visual Reasoning with Self-Correction
This addresses the problem of unreliable cross-modal reasoning for users of multimodal AI systems, representing a novel method rather than an incremental improvement.
The paper tackles the problem of limited reasoning reliability in multimodal large language models by proposing the H-GIVR framework, which uses historical reasoning information to enable dynamic error correction during iterative visual reasoning, achieving a 107% accuracy improvement over baseline on the ScienceQA dataset with Llama3.2-vision:11b.
Self-consistency methods are the core technique for improving the reasoning reliability of multimodal large language models (MLLMs). By generating multiple reasoning results through repeated sampling and selecting the best answer via voting, they play an important role in cross-modal tasks. However, most existing self-consistency methods are limited to a fixed ``repeated sampling and voting'' paradigm and do not reuse historical reasoning information. As a result, models struggle to actively correct visual understanding errors and dynamically adjust their reasoning during iteration. Inspired by the human reasoning behavior of repeated verification and dynamic error correction, we propose the H-GIVR framework. During iterative reasoning, the MLLM observes the image multiple times and uses previously generated answers as references for subsequent steps, enabling dynamic correction of errors and improving answer accuracy. We conduct comprehensive experiments on five datasets and three models. The results show that the H-GIVR framework can significantly improve cross-modal reasoning accuracy while maintaining low computational cost. For instance, using \texttt{Llama3.2-vision:11b} on the ScienceQA dataset, the model requires an average of 2.57 responses per question to achieve an accuracy of 78.90\%, representing a 107\% improvement over the baseline.