多尺度Transformer金字塔网络用于多变量时间序列预测
Multi-scale Transformer Pyramid Networks for Multivariate Time Series Forecasting
AI总结：
本文提出多尺度Transformer金字塔网络（MTPNet），通过维度不变嵌入和多个无约束尺度捕获时间依赖，在九个基准数据集上超越现有最先进方法。
摘要：
多变量时间序列（MTS）预测涉及对历史记录中的时间依赖关系进行建模。Transformer因其捕获长期依赖的能力，在MTS预测中表现出卓越的性能。然而，先前的工作仅限于在固定尺度或指数增长（多数以2为底）的多个尺度上建模时间依赖。这一局限性阻碍了其捕获多样化季节性（如小时和日模式）的有效性。在本文中，我们引入了一种维度不变嵌入技术，该技术捕获短期时间依赖并将MTS数据投影到更高维空间，同时保留MTS数据中时间步和变量的维度。此外，我们提出了一种新颖的多尺度Transformer金字塔网络（MTPNet），专门设计用于在多个无约束尺度上有效捕获时间依赖。预测是从不同尺度Transformer获得的多尺度潜在表示中推断出来的。在九个基准数据集上的广泛实验表明，所提出的MTPNet优于最近的最先进方法。
英文摘要：
Multivariate Time Series (MTS) forecasting involves modeling temporal dependencies within historical records. Transformers have demonstrated remarkable performance in MTS forecasting due to their capability to capture long-term dependencies. However, prior work has been confined to modeling temporal dependencies at either a fixed scale or multiple scales that exponentially increase (most with base 2). This limitation hinders their effectiveness in capturing diverse seasonalities, such as hourly and daily patterns. In this paper, we introduce a dimension invariant embedding technique that captures short-term temporal dependencies and projects MTS data into a higher-dimensional space, while preserving the dimensions of time steps and variables in MTS data. Furthermore, we present a novel Multi-scale Transformer Pyramid Network (MTPNet), specifically designed to effectively capture temporal dependencies at multiple unconstrained scales. The predictions are inferred from multi-scale latent representations obtained from transformers at various scales. Extensive experiments on nine benchmark datasets demonstrate that the proposed MTPNet outperforms recent state-of-the-art methods.
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。
