具有显著性意识的多路由思考：重新审视视觉-语言推理
Saliency-Aware Multi-Route Thinking: Revisiting Vision-Language Reasoning
AI总结：
本文提出Saliency-Aware Principle（SAP）方法，通过高层次推理原则提升视觉-语言推理的稳定性与效率，减少幻觉并提升响应速度。
摘要：
视觉-语言模型（VLMs）旨在通过联合利用视觉和文本模态进行推理。虽然为大型语言模型（LLMs）分配额外的推理时间计算已被证明是有效的，但在VLMs中实现类似的扩展仍然具有挑战性。关键障碍在于视觉输入通常只在生成开始时提供一次，而文本推理（例如早期视觉摘要）是自回归生成的，导致推理越来越偏向文本，并允许早期视觉基础错误的积累。此外，普通的指导在推理期间进行视觉基础时往往粗略且嘈杂，使得在长文本上引导推理变得困难。为了解决这些挑战，我们提出了\emph{显著性意识原则}（SAP）选择。SAP在高层次推理原则上操作，而不是在token级轨迹上，这使得在嘈杂反馈下能够稳定地控制离散生成，同时允许后续推理步骤在需要重新咨询视觉证据时进行重新咨询。此外，SAP支持多路由推理，能够并行探索多样化的推理行为。SAP是模型无关且数据无关的，不需要额外训练。实验结果表明，在可比的token生成预算下，SAP实现了有竞争力的性能，特别是在减少对象幻觉方面，同时比CoT式长序列推理产生了更稳定的推理和更低的响应延迟。
英文摘要：
Vision-language models (VLMs) aim to reason by jointly leveraging visual and textual modalities. While allocating additional inference-time computation has proven effective for large language models (LLMs), achieving similar scaling in VLMs remains challenging. A key obstacle is that visual inputs are typically provided only once at the start of generation, while textual reasoning (e.g., early visual summaries) is generated autoregressively, causing reasoning to become increasingly text-dominated and allowing early visual grounding errors to accumulate. Moreover, vanilla guidance for visual grounding during inference is often coarse and noisy, making it difficult to steer reasoning over long texts. To address these challenges, we propose \emph{Saliency-Aware Principle} (SAP) selection. SAP operates on high-level reasoning principles rather than token-level trajectories, which enable stable control over discrete generation under noisy feedback while allowing later reasoning steps to re-consult visual evidence when renewed grounding is required. In addition, SAP supports multi-route inference, enabling parallel exploration of diverse reasoning behaviors. SAP is model-agnostic and data-free, requiring no additional training. Empirical results show that SAP achieves competitive performance, especially in reducing object hallucination, under comparable token-generation budgets while yielding more stable reasoning and lower response latency than CoT-style long sequential reasoning.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
