CAPS通过分拆时间序列的三类结构，实现更精准的预测。
CAPS: Unifying Attention, Recurrence, and Alignment in Transformer-based Time Series Forecasting
- 用旋转对齐相位，分路处理趋势、冲击与周期三类时间模式。
- 在多个基准上超越普通注意力与线性注意力，复杂度仍为线性。
- 适合需要长短期建模的时序预测任务，尤其关注模式分离的场景。
本文提出CAPS（Clock-weighted Aggregation with Prefix-products and Softmax），一种面向时间序列预测的结构化注意力机制，可解耦全局趋势、局部冲击与季节性模式三类时间结构。标准softmax注意力因全局归一化而混淆这些结构，近期递归模型则牺牲了长期、无序选择能力以换取有序因果结构。CAPS在单个注意力层中结合SO(2)旋转实现相位对齐，并引入三种加性门控路径：黎曼softmax、前缀积门控与时钟基线。我们设计了时钟机制，通过共享的时间重要性学习权重，动态调节各路径。在长短期预测基准上的实验表明，CAPS优于原始softmax与线性注意力，且在七种强基线上表现竞争力，同时保持线性复杂度。代码已公开于https://github.com/vireshpati/CAPS-Attention。
原文摘要 · Abstract (English)
This paper presents $\textbf{CAPS}$ (Clock-weighted Aggregation with Prefix-products and Softmax), a structured attention mechanism for time series forecasting that decouples three distinct temporal structures: global trends, local shocks, and seasonal patterns. Standard softmax attention entangles these through global normalization, while recent recurrent models sacrifice long-term, order-independent selection for order-dependent causal structure. CAPS combines SO(2) rotations for phase alignment with three additive gating paths -- Riemann softmax, prefix-product gates, and a Clock baseline -- within a single attention layer. We introduce the Clock mechanism, a learned temporal weighting that modulates these paths through a shared notion of temporal importance. Experiments on long- and short-term forecasting benchmarks surpass vanilla softmax and linear attention mechanisms and demonstrate competitive performance against seven strong baselines with linear complexity. Our code implementation is available at https://github.com/vireshpati/CAPS-Attention.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。