KDD 2025 | (8月轮)时间序列论文总结

KDD 2025将在2025年8月3号到7号在加拿大多伦多举行，本文总结了KDD 2025(August Cycle)有关时间序列（Time Series）相关文章，共计11篇，其中1-10为Research Track，11为ADS Track。如有疏漏，欢迎补充！

时间序列Topic：预测，异常检测，测试时适应等。

Research

1 IN-Flow: Instance Normalization Flow for Non-stationary Time Series Forecasting.

链接：https://dl.acm.org/doi/10.1145/3690624.3709260

作者：Wei Fan, Shun Zheng, Pengyang Wang, Rui Xie, Kun Yi, Qi Zhang, Jiang Bian, Yanjie Fu

关键词：预测，分布偏移，归一化流

IN-Flow

IN-Flow

2 Quantum Time-index Models with Reservoir for Time Series Forecasting

链接：https://dl.acm.org/doi/10.1145/3690624.3709228

作者：Wenbo Qiao, Jiaming Zhao, Peng Zhang

关键词：预测，时间索引，量子模型

3 DUET: Dual Clustering Enhanced Multivariate Time Series Forecasting

链接：https://dl.acm.org/doi/10.1145/3690624.3709325

代码：https://github.com/decisionintelligence/DUET

作者： Xiangfei Qiu, Xingjian Wu, Yan Lin, Chenjuan Guo, Jilin Hu, Bin Yang

关键词：预测，双向聚类

新智元：时序预测再出新范式！华东师大提出DUET：「双向聚类」新设计，性能刷新SOTA！| KDD 2025

DUET

DUET

4 ST-MTM: Masked Time Series Modeling with Seasonal-Trend Decomposition for Time Series Forecasting

链接：https://dl.acm.org/doi/10.1145/3690624.3709254

作者：Hyunwoo Seo, Chiehyeon Lim

关键词：预测，掩码时间序列建模，季节性 - 趋势分解的

5 Probabilistic Hypergraph Recurrent Neural Networks for Time-series Forecasting

链接：https://dl.acm.org/doi/10.1145/3690624.3709202

作者：Hongjie Chen, Ryan A. Rossi, Sungchul Kim, Kanak Mahadik, Hoda Eldardir

关键词：概率预测，超图，循环神经网络

6 Proactive Model Adaptation Against Concept Drift for Online Time Series Forecasting

链接：https://dl.acm.org/doi/10.1145/3690624.3709210

代码：https://github.com/SJTU-DMTai/OnlineTSF

作者： Lifan Zhao, Yanyan Shen

关键词：预测，在线学习，概念漂移

7 Learning Universal Multi-level Market Irrationality Factors to Improve Stock Return Forecasting

链接：https://dl.acm.org/doi/10.1145/3690624.3709328

代码：https://github.com/lIcIIl/UMI

作者：Chen Yang, Jingyuan Wang, Xiaohan Jiang, Junjie Wu

关键词：股票收益预测，市场非理性，深度学习，自监督学习

UMI

UMI

8 TSINR: Capturing Temporal Continuity via Implicit Neural Representations for Time Series Anomaly Detection

链接：https://dl.acm.org/doi/10.1145/3690624.3709266

代码：https://github.com/Leanna97/TSINR

作者：Mengxuan Li, Ke Liu, Hongyang Chen, Jiajun Bu, Hongwei Wang, Haishuai Wang

关键词：异常检测，隐藏神经表示，无监督学习

9 Noise-Resilient Point-wise Anomaly Detection in Time Series Using Weak Segment Labels

链接：https://dl.acm.org/doi/10.1145/3690624.3709257

代码：https://github.com/UCSC-REAL/Nrdetector

作者：Yaxuan Wang, Hao Cheng, Jing Xiong, Qingsong Wen, Han Jia, Ruixuan Song, Liyuan Zhang, Zhaowei Zhu, Yang Liu

关键词：异常检测，正样本和无标签学习

Nrdetector

Nrdetector

10 Augmented Contrastive Clustering with Uncertainty-Aware Prototyping for Time Series Test Time Adaptation

链接：https://dl.acm.org/doi/10.1145/3690624.3709239

代码：https://github.com/Tokenmw/ACCUP-main

作者：Peiliang Gong, Mohamed Ragab, Min Wu, Zhenghua Chen, Yongyi Su, Xiaoli Li, Daoqiang Zhang

关键词：测试时适应，原型学习

ACCUP

ACCUP

ADS

11 Multi-period Learning for Financial Time Series Forecasting

链接：https://dl.acm.org/doi/10.1145/3690624.3709422

代码：https://github.com/Meteor-Stars/MLF

作者：u Zhang, Zhengang Huang, Yunzhi Wu, Xun Lu, Erpeng Qi, Yunkai Chen, Zhongya Xue, Qitong Wang, Peng Wang, Wei Wang

关键词：金融时间序列预测，多周期学习

MLF