Real-world time-series data is highly non stationary and complex in dynamics that operate across multiple timescales, ranging from fast, short-term changes to slow, long-term trends. Most existing models rely on fixed-scale structural priors, such as patch-based tokenization, fixed frequency transformations, or frozen backbone architectures. This often leads to over-regularization of temporal dynamics, which limits their ability to adaptively model the full spectrum of temporal variations and impairs their performance on unpredictable, Sudden, high-magnitude events. To address this, we introduce the Multi-scale Temporal Network (MSTN), a novel deep learning architecture founded on a hierarchical multi-scale and sequence modeling principle. The MSTN framework integrates: (i) a multi-scale convolutional encoder that constructs a hierarchical feature pyramid for local patterns (ii) a sequence modeling component for long-range temporal dependencies. We empirically validate this with BiLSTM and Transformer variants, establishing a flexible foundation for future architectural advancements. and (iii) a gated fusion mechanism augmented with squeeze-and-excitation (SE) and multi-head temporal attention (MHTA) for dynamic, context-aware feature integration. This design enables MSTN to adaptively model temporal patterns from milliseconds to long-range dependencies within a unified framework. Extensive evaluations across time-series long-horizon forecasting, imputation, classification and generalizability study demonstrate that MSTN achieves competitive state-of-the-art (SOTA) performance, showing improvements over contemporary approaches including EMTSF, LLM4TS, HiMTM, TIME-LLM, MTST, SOFTS, iTransformer, TimesNet, and PatchTST. In total, MSTN establishes new SOTA performance on 24 of 32 benchmark datasets, demonstrating its consistent performance across diverse temporal tasks.
翻译:现实世界的时间序列数据具有高度非平稳性,其动态特性在多个时间尺度上运作,涵盖从快速的短期变化到缓慢的长期趋势。大多数现有模型依赖于固定尺度的结构先验,例如基于分块的标记化、固定频率变换或冻结的主干架构。这通常导致对时间动态的过度正则化,限制了其自适应建模全频谱时间变化的能力,并削弱了其对不可预测、突发、高幅度事件的性能。为解决这一问题,我们引入了多尺度时间网络(MSTN),这是一种基于分层多尺度与序列建模原理的新型深度学习架构。MSTN框架整合了:(i)一个多尺度卷积编码器,用于构建局部模式的分层特征金字塔;(ii)一个用于长程时间依赖性的序列建模组件,我们通过BiLSTM和Transformer变体进行了实证验证,为未来架构进展建立了灵活基础;以及(iii)一个结合了挤压-激励(SE)和多头时间注意力(MHTA)的门控融合机制,用于动态、上下文感知的特征集成。该设计使MSTN能够在统一框架内自适应地建模从毫秒级到长程依赖的时间模式。在时间序列长程预测、插补、分类和泛化性研究中的广泛评估表明,MSTN实现了具有竞争力的最先进(SOTA)性能,相较于包括EMTSF、LLM4TS、HiMTM、TIME-LLM、MTST、SOFTS、iTransformer、TimesNet和PatchTST在内的当代方法均显示出改进。总体而言,MSTN在32个基准数据集中的24个上确立了新的SOTA性能,证明了其在多样化时间任务中的一致表现。