LapMerge：线性时间内的可微精确预算令牌合并
LapMerge: Differentiable Exact-Budget Token Merging in Linear Time
AI总结：
LapMerge是一种线性时间可微精确预算令牌合并算子，在视觉与视觉-语言任务中性能相当或更优，速度提升达两个数量级，兼具精确预算、可微性与线性成本优势。
摘要：
现代视觉与语言模型处理的令牌序列越来越长，减少序列长度是降低下游成本的主要手段。然而令牌压缩面临一个根本权衡：选择令牌会丢弃信息，而现有合并方法仅通过成对令牌-令牌匹配或MN次令牌-输出交互来保留信息。我们提出LapMerge，一种可微合并算子，针对有序序列解决该权衡问题：学习到的质量分布在拉普拉斯核密度下将序列划分为恰好N个等质量片段，通过前缀/后缀扫描聚合每个片段，因此前向和反向传播均在O(M+N)时间内运行。这种精确输出预算、端到端可微性与线性成本的组合，正是先前的压缩器相互权衡的特性。预算可在不改变算子参数的情况下调整，且学习到的质量可作为令牌级别的预算分配位置映射。在视觉和视觉-语言任务中，LapMerge在强大的令牌压缩器上达到相当或更优性能，同时在长序列长度下速度提升达两个数量级；当令牌顺序反映空间或时间局部性时，其效果尤为显著，既可作为即插即用的压缩模块，也可作为可训练的架构组件。
英文摘要：
Modern vision and language models process ever longer token sequences, and reducing them is the main lever on downstream cost. Token compression, however, faces a fundamental trade-off: selecting tokens discards information, while existing merging methods preserve it only through pairwise token-to-token matching or $MN$ token-to-output interactions. We introduce LapMerge, a differentiable merging operator that resolves this trade-off for ordered sequences: a learned mass distribution partitions the sequence into exactly $N$ equal-mass segments under a Laplace-kernel density, and prefix/suffix scans aggregate each segment, so both the forward and the backward pass run in $\mathcal{O}(M+N)$ time. This combination - an exact output budget, end-to-end differentiability, and linear cost - is precisely what previous reducers trade off against one another. The budget can be changed without changing operator parameters, and the learned mass doubles as a token-level map of where the budget is allocated. Across vision and vision-language tasks, LapMerge matches or improves on strong token reducers while being up to two orders of magnitude faster at large sequence lengths; it is particularly effective when token order reflects spatial or temporal locality, and works both as a drop-in reduction module and as a trainable architectural component.
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。
