Saved in:
Bibliographic Details
Main Authors: Zhang, Kai, Sun, Siming, Fan, Zhengyu, Yang, Qinmin, Jiang, Xuejun
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.17845
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914050972057600
author Zhang, Kai
Sun, Siming
Fan, Zhengyu
Yang, Qinmin
Jiang, Xuejun
author_facet Zhang, Kai
Sun, Siming
Fan, Zhengyu
Yang, Qinmin
Jiang, Xuejun
contents Time series analysis faces significant challenges in handling variable-length data and achieving robust generalization. While Transformer-based models have advanced time series tasks, they often struggle with feature redundancy and limited generalization capabilities. Drawing inspiration from classical CNN architectures' pyramidal structure, we propose a Multi-Scale Representation Learning Framework based on a Conv-like ScaleFusion Transformer. Our approach introduces a temporal convolution-like structure that combines patching operations with multi-head attention, enabling progressive temporal dimension compression and feature channel expansion. We further develop a novel cross-scale attention mechanism for effective feature fusion across different temporal scales, along with a log-space normalization method for variable-length sequences. Extensive experiments demonstrate that our framework achieves superior feature independence, reduced redundancy, and better performance in forecasting and classification tasks compared to state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2509_17845
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Conv-like Scale-Fusion Time Series Transformer: A Multi-Scale Representation for Variable-Length Long Time Series
Zhang, Kai
Sun, Siming
Fan, Zhengyu
Yang, Qinmin
Jiang, Xuejun
Machine Learning
Time series analysis faces significant challenges in handling variable-length data and achieving robust generalization. While Transformer-based models have advanced time series tasks, they often struggle with feature redundancy and limited generalization capabilities. Drawing inspiration from classical CNN architectures' pyramidal structure, we propose a Multi-Scale Representation Learning Framework based on a Conv-like ScaleFusion Transformer. Our approach introduces a temporal convolution-like structure that combines patching operations with multi-head attention, enabling progressive temporal dimension compression and feature channel expansion. We further develop a novel cross-scale attention mechanism for effective feature fusion across different temporal scales, along with a log-space normalization method for variable-length sequences. Extensive experiments demonstrate that our framework achieves superior feature independence, reduced redundancy, and better performance in forecasting and classification tasks compared to state-of-the-art methods.
title Conv-like Scale-Fusion Time Series Transformer: A Multi-Scale Representation for Variable-Length Long Time Series
topic Machine Learning
url https://arxiv.org/abs/2509.17845