Task-Aware Mixture-of-Experts for Time Series Analysis

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wu, Xingjian, Li, Zhengyu, Cheng, Hanyin, Qiu, Xiangfei, Hu, Jilin, Guo, Chenjuan, Yang, Bin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918296285085696
author Wu, Xingjian
Li, Zhengyu
Cheng, Hanyin
Qiu, Xiangfei
Hu, Jilin
Guo, Chenjuan
Yang, Bin
author_facet Wu, Xingjian
Li, Zhengyu
Cheng, Hanyin
Qiu, Xiangfei
Hu, Jilin
Guo, Chenjuan
Yang, Bin
contents Time Series Analysis is widely used in various real-world applications such as weather forecasting, financial fraud detection, imputation for missing data in IoT systems, and classification for action recognization. Mixture-of-Experts (MoE), as a powerful architecture, though demonstrating effectiveness in NLP, still falls short in adapting to versatile tasks in time series analytics due to its task-agnostic router and the lack of capability in modeling channel correlations. In this study, we propose a novel, general MoE-based time series framework called PatchMoE to support the intricate ``knowledge'' utilization for distinct tasks, thus task-aware. Based on the observation that hierarchical representations often vary across tasks, e.g., forecasting vs. classification, we propose a Recurrent Noisy Gating to utilize the hierarchical information in routing, thus obtaining task-sepcific capability. And the routing strategy is operated on time series tokens in both temporal and channel dimensions, and encouraged by a meticulously designed Temporal \& Channel Load Balancing Loss to model the intricate temporal and channel correlations. Comprehensive experiments on five downstream tasks demonstrate the state-of-the-art performance of PatchMoE.
format Preprint
id arxiv_https___arxiv_org_abs_2509_22279
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Task-Aware Mixture-of-Experts for Time Series Analysis
Wu, Xingjian
Li, Zhengyu
Cheng, Hanyin
Qiu, Xiangfei
Hu, Jilin
Guo, Chenjuan
Yang, Bin
Machine Learning
Time Series Analysis is widely used in various real-world applications such as weather forecasting, financial fraud detection, imputation for missing data in IoT systems, and classification for action recognization. Mixture-of-Experts (MoE), as a powerful architecture, though demonstrating effectiveness in NLP, still falls short in adapting to versatile tasks in time series analytics due to its task-agnostic router and the lack of capability in modeling channel correlations. In this study, we propose a novel, general MoE-based time series framework called PatchMoE to support the intricate ``knowledge'' utilization for distinct tasks, thus task-aware. Based on the observation that hierarchical representations often vary across tasks, e.g., forecasting vs. classification, we propose a Recurrent Noisy Gating to utilize the hierarchical information in routing, thus obtaining task-sepcific capability. And the routing strategy is operated on time series tokens in both temporal and channel dimensions, and encouraged by a meticulously designed Temporal \& Channel Load Balancing Loss to model the intricate temporal and channel correlations. Comprehensive experiments on five downstream tasks demonstrate the state-of-the-art performance of PatchMoE.
title Task-Aware Mixture-of-Experts for Time Series Analysis
topic Machine Learning
url https://arxiv.org/abs/2509.22279