[论文] TimeSeek: Multi-Task and Multi-Domain Pretraining for Generalized Time...

论文概要

研究领域: ML 作者: Peng Chen, Yihang Wang, Zhongwen Rao, Yang Shu, Yuying Qiu, Yuxuan Chen, Lujia Pan, Bin Yang, Chenjuan Guo 发布时间: 2026-11-10 arXiv: 2611.14491

中文摘要

时间序列分析已广泛应用于各领域。现有时间序列基础模型（TSFM）虽有前景，仍面临两大挑战：(1) 多数 TSFM 主要靠堆叠 Transformer 层扩大参数量，缺乏从多域数据自适应建模复杂时序模式的能力；(2) 主要为特定任务预训练（如预测或异常检测），少数支持多任务者也未充分利用任务特异性与共通性，性能欠优。我们提出 TimeSeek：一个新颖的通用时序分析模型，在多域数据上自适应建模时序模式，并在预训练中统一多任务。概念上 TimeSeek 受"领域与任务异构下的选择性信息共享"引导。多域方面，提出多域特征提取器，基于领域特有特征自适应捕获时序依赖与通道相关性，包括渐进多尺度建模与自适应槽位通道建模。任务多样性方面，提出多任务协同训练，通过任务定制与跨任务训练利用任务特异性与共通性，在预训练中有效统一多任务。在大规模时序数据预训练后，TimeSeek 在预测、异常检测、分类任务上、少样本与零样本场景都取得 SOTA，且具备可观推理效率。

原文摘要

Time series analytics has been widely applied across various domains. While existing time series foundation models (TSFMs) have demonstrated promising results, they still face two major challenges: 1) Most TSFMs mainly rely on stacking Transformer layers to increase parameter size, lacking the ability to adaptively model complex time series patterns from multi-domain data. 2) Current TSFMs are primarily pre-trained for specific tasks, e.g., forecasting or anomaly detection, and a few models support multiple tasks, but they fail to fully leverage task specificity and commonality, resulting in suboptimal performance. To address these challenges, we propose TimeSeek, a novel generalized time series analytics model that adaptively models time series patterns from multi-domain data and unifies ...

#论文 #arXiv #ML #小凯