GThinker：通过线索引导重思考迈向通用多模态推理
GThinker: Towards General Multimodal Reasoning via Cue-Guided Rethinking
提出「视觉惯性」问题并设计线索引导的重思考机制，可被视觉推理模型复用，用于缓解模型对初始视觉感知的过度自信，提升复杂视觉问答的可靠性。
原文依据 GThinker leverages Cue-Rethinking, a flexible reasoning pattern that not only grounds reasoning in visual cues but also strategically triggers a re-examination of these cues to resolve inconsistencies.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论尽管多模态推理近期取得了进展，多模态大语言模型（MLLMs）在初始视觉感知可能产生误导的复杂任务上仍然表现不佳。这一性能差距源于我们称为「视觉惯性」的关键推理缺陷：MLLMs 在文本语境中擅长迭代反思，却倾向于不加批判地固守其初始视觉解释，很少加以修正。为克服这一局限，我们提出 GThinker，一个具备新型自适应视觉重思考能力的 MLLM。GThinker 利用线索重思考（Cue-Rethinking），一种灵活的推理模式，不仅将推理建立在视觉线索之上，还策略性地触发对这些线索的重新审视以解决不一致。为灌输这一能力，我们引入一种新颖的两阶段训练框架。该框架以模式引导的冷启动开始，并通过评判引导的选择机制加以增强以从失败案例中学习，随后进行激励强化学习。我们进一步构建了 GThinker-11K 数据集，通过迭代式多模态标注流程为训练提供支持。大量实验表明，GThinker 显著缓解了推理过程中的视觉惯性，在富含此类挑战的 M3CoT 基准上取得领先的 81.5%，超越了强大的 o4-mini 模型。此外，GThinker 在一系列多模态推理基准上均取得一致提升，平均增益 2.1%，展示了赋予 MLLMs 重新思考其所见与所思能力的广泛益处。数据、模型和代码见 https://github.com/jefferyZhan/GThinker。
英文原文摘要 1654 字符
Despite recent advances in multimodal reasoning, Multimodal Large Language Models (MLLMs) still struggle on complex tasks where initial visual perceptions can be misleading. This performance gap stems from a critical reasoning flaw we term Visual Inertia: while MLLMs excel at iterative reflection in textual contexts, they tend to uncritically commit to their initial visual interpretations and rarely revise them. To overcome this limitation, we introduce GThinker, an MLLM equipped with a novel adaptive visual rethinking capability. GThinker leverages Cue-Rethinking, a flexible reasoning pattern that not only grounds reasoning in visual cues but also strategically triggers a re-examination of these cues to resolve inconsistencies. To instill this capability, we introduce a novel two-stage training framework. It begins with a pattern-guided cold start, enhanced by a judge-guided selective mechanism to learn from failure cases, followed by incentive reinforcement learning. We further curate the GThinker-11K dataset to power the training with an iterative multimodal annotation pipeline. Extensive experiments demonstrate that GThinker significantly mitigates visual inertia during reasoning, achieving a leading 81.5% on the M3CoT benchmark, which is rich in such challenges, surpassing the powerful o4-mini model. Furthermore, GThinker shows consistent improvements across a range of multimodal reasoning benchmarks with an average gain of 2.1%, showcasing the broad benefits of equipping MLLMs with the ability to rethink both what they see and how they think. Data, model, and code are available at https://github.com/jefferyZhan/GThinker.