用B样条自适应分段时间序列，提升长时预测精度与内存效率。
BSAT: B-Spline Adaptive Tokenizer for Long-Term Time Series Forecasting
- 基于B样条动态划分时间序列，重点捕捉高曲率区域的特征。
- 在多个数据集上实现高压缩比下媲美主流模型的预测性能。
- 适合内存受限场景，如边缘设备部署或超长序列分析。
使用Transformer进行长期时间序列预测受到自注意力二次复杂度和均匀分块刚性的限制，可能与数据语义结构不匹配。本文提出一种无需参数的新型方法——B-Spline自适应分词器（BSAT），通过拟合时间序列的B样条实现自适应分段。该方法在高曲率区域自动放置标记，并将每个可变长度基函数表示为固定大小的标记，包含其系数和位置信息。此外，我们提出一种混合位置编码，结合加性可学习位置编码与层间可学习基底的旋转位置编码（L-RoPE），使每层能够关注不同的时间依赖关系。在多个公开基准测试中，所提模型在高压缩率下表现出色，性能优于或媲美强基线模型，特别适用于内存约束严格的应用场景。
原文摘要 · Abstract (English)
Long-term time series forecasting using transformers is hampered by the quadratic complexity of self-attention and the rigidity of uniform patching, which may be misaligned with the data's semantic structure. In this paper, we introduce the \textit{B-Spline Adaptive Tokenizer (BSAT)}, a novel, parameter-free method that adaptively segments a time series by fitting it with B-splines. BSAT algorithmically places tokens in high-curvature regions and represents each variable-length basis function as a fixed-size token, composed of its coefficient and position. Further, we propose a hybrid positional encoding that combines a additive learnable positional encoding with Rotary Positional Embedding featuring a layer-wise learnable base: L-RoPE. This allows each layer to attend to different temporal dependencies. Our experiments on several public benchmarks show that our model is competitive with strong performance at high compression rates. This makes it particularly well-suited for use cases with strong memory constraints.
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。