用可学习的小波替换注意力，让时间序列预测更快更省内存
DB2-TransF: All You Need Is Learnable Daubechies Wavelets for Time Series Forecasting
- 用可学习的达布勒小波替代自注意力，捕捉多尺度时序模式
- 在13个基准上精度媲美传统Transformer，内存消耗显著降低
- 适合需要高效推理的工业级时间序列预测场景
时间序列预测需有效捕捉复杂时序依赖，尤其在大规模高维场景下。尽管基于Transformer的架构擅长建模长程依赖，但其二次计算复杂度限制了可扩展性和适应性。为此，我们提出DB2-TransF，一种受Transformer启发的新架构，将自注意力机制替换为可学习的达布勒小波系数层。该小波模块能高效捕捉多尺度局部与全局模式，并增强多时间序列间的相关性建模能力。在13个标准预测基准上的大量实验表明，DB2-TransF在预测精度上达到或超过传统Transformer水平，同时显著降低内存使用。实验结果表明，DB2-TransF是一种可扩展且资源高效的先进时间序列预测框架。代码已开源：https://github.com/SteadySurfdom/DB2-TransF。
原文摘要 · Abstract (English)
Time series forecasting requires models that can efficiently capture complex temporal dependencies, especially in large-scale and high-dimensional settings. While Transformer-based architectures excel at modeling long-range dependencies, their quadratic computational complexity poses limitations on scalability and adaptability. To overcome these challenges, we introduce DB2-TransF, a novel Transformer-inspired architecture that replaces the self-attention mechanism with a learnable Daubechies wavelet coefficient layer. This wavelet-based module efficiently captures multi-scale local and global patterns and enhances the modeling of correlations across multiple time series for the time series forecasting task. Extensive experiments on 13 standard forecasting benchmarks demonstrate that DB2-TransF achieves comparable or superior predictive accuracy to conventional Transformers, while substantially reducing memory usage for the time series forecasting task. The obtained experimental results position DB2-TransF as a scalable and resource-efficient framework for advanced time series forecasting. Our code is available at https://github.com/SteadySurfdom/DB2-TransF
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。