位置重绑定缓存复用：交错多模态推理中无重放的视觉重访
Position Rebinding Cache Reuse: Replay-Free Visual Revisiting for Interleaved Multimodal Reasoning
AI总结：
提出位置重绑定缓存复用（PRCR）框架，通过重绑历史视觉KV缓存的位置坐标，实现无重放的高效视觉重访，在多个多模态推理基准上平均准确率提升5%，计算量减少数万倍。
摘要：
交错多模态推理通过在多步生成过程中重访视觉证据来改善视觉定位，但现有方法通常依赖令牌重放，即重复前馈选定的视觉令牌。一个自然的捷径是直接复用历史的视觉键值（KV）缓存。然而，我们识别出该策略的一个关键失败模式：缓存的视觉键已经绑定到其原始位置上下文。这种过时的位置绑定会在后续解码上下文中扭曲注意力，并可能触发严重的自回归解码崩溃。这一失败表明，有效的缓存复用需要在与当前解码状态兼容的位置下重建视觉证据，而不是直接复制位置绑定的历史缓存条目。为此，我们提出位置重绑定缓存复用（PRCR），一种用于无重放视觉重访的缓存级框架。PRCR将原始视觉KV缓存及其空间坐标一起存储，然后重新分配位置兼容的坐标以选择条目，并在将重建的缓存注入活动解码器缓存之前重绑它们的键。这种设计复用了历史视觉证据，同时保留了文本位置连续性和相对视觉结构。在多个多模态推理基准上的实验表明，PRCR实现了与重放相当或更好的性能，平均准确率提升5%，并将视觉重访计算量减少高达数万倍。
英文摘要：
Interleaved multimodal reasoning improves visual grounding by revisiting visual evidence during multi-step generation, yet existing methods typically rely on token replay, repeatedly forwarding selected visual tokens. A natural shortcut is to reuse the historical visual key-value (KV) cache directly. However, we identify a critical failure mode of this strategy: cached visual keys are already bound to their original positional context. Such stale positional binding distorts attention under later decoding contexts and can trigger severe autoregressive decoding collapse. This failure suggests that effective cache reuse requires reconstructing visual evidence under positions compatible with the current decoding state, rather than directly copying position-bound historical cache entries. To this end, we propose Position Rebinding Cache Reuse (PRCR), a cache-level framework for replay-free visual revisiting. PRCR stores raw visual KV cache together with their original spatial coordinates, then reassigns position-compatible coordinates to select entries and rebinds their keys before injecting the reconstructed cache into the active decoder cache. This design reuses historical visual evidence while preserving textual positional continuity and relative visual structure. Experiments across multiple multimodal reasoning benchmarks show that PRCR achieves replay-level or better performance, improving average accuracy by 5 percent and reducing visual-revisiting computation by up to tens of thousands of times.
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。
