AAAI 2026 | 时间序列(Time Series) 论文总结[上] (Oral+预测)

时空探索之旅

发布于 2026-03-10 14:39:58

2270

文章被收录于专栏：时空探索之旅时空探索之旅

AAAI 2026将在2026年1月20日到1月27日于新加坡（Singapore）举行。AAAI 2026会议主会共有23, 680篇论文投稿，其中4, 167 篇被接收，接收率为17.6%。

本文总结了2026 AAAI 上有关时间序列（Time Series）相关论文。如有疏漏，欢迎大家补充。

时间序列Topic：时间序列预测，异常检测，分类，表示学习，基础模型，大语言模型等内容。总计67篇，本文涉及35篇。

注：由于论文数目较多，分为上下篇，此为上篇，主要涵盖所有Oral和时间序列预测的文章，其中前5篇为Oral。下篇主要涵盖异常检测，分类，生成，表示学习，基础模型等。

Oral1. ReCast: Reliability-aware Codebook-assisted Lightweight Time Series Forecasting2. GraFT: Infusing Pre-trained Transformers with Relational Structure for Time Series Forecasting3. TimeCAP: A Channel-Aware Pre-Training Framework for Multivariate Time Series Forecasting4. Beyond Missing Data Imputation: Information- Theoretic Coupling of Missingness and Class Imbalance for Optimal Irregular Time Series Classification5. CometNet: Contextual Motif-guided Long-term Time Series ForecastingPoster6. IdealTSF: Can Non-Ideal Data Contribute to Enhancing the Performance of Time Series Forecasting Models?7. Time Series Forecasting via Direct Per-Step Probability Distribution Modeling8. Revitalizing Canonical Pre-Alignment for Irregular Multivariate Time Series Forecasting9. Goal-Oriented Time-Series Forecasting: FoundationFramework Design10. STrans:SpontaneousArchitecture Evolution for AdaptiveTime Series Forecasting11. MDMLP-EIA: Multi-domain Dynamic MLPs with EnergyInvariant Attention for Time Series Forecasting12. Sonnet: Spectral Operator Neural Network for MultivariableTime Series Forecasting13. FedSkeleton: Secure Multi-Party Graph SkeletonConstruction for Privacy-Preserving Federated Time-SeriesForecasting14. T3Time: Tri-Modal Time Series Forecasting via AdaptiveMulti-Head Alignment and Residual Fusion15. FreDN: Spectral Disentanglement for Time SeriesForecasting via Learnable Frequency Decomposition16. Sparse-Scale Transformer with Bidirectional Awareness forTime Series Forecasting17. OccamVTS: Distilling Vision Models to 1% Parameters forTime Series Forecasting18. Multi-Horizon Time Series Forecasting of non-parametricCDFs with Deep Lattice Networks19. Towards Non-Stationary Time Series Forecasting withTemporal Stabilization and Frequency Differencing20. TimeMosaic: Temporal Heterogeneity Guided Time Series Forecasting via Adaptive Granularity Patch and Segment-wise Decoding21. FeTS: A Feature-Aware Framework for Time Series Forecasting22. XLinear: A Lightweight and Accurate MLP-Based Model for Long-Term Time Series Forecasting with Exogenous Inputs23. Beyond Immediate Activation: Temporally Decoupled Backdoor Attacks on Time Series Forecasting24. DeepBooTS: Dual-Stream Residual Boosting for Drift-Resilient Time-Series Forecasting25. EMAformer: Enhancing Transformer Through Embedding Armor for Time Series Forecasting26. Optimal Look-back Horizon for Time Series Forecasting in Federated Learning27. M2FMoE: Multi-Resolution Multi-View Frequency Mixture-of-Experts for Extreme-Adaptive Time Series Forecasting28. Rethinking Irregular Time Series Forecasting: A Simple Yet Effective Baseline29. HN-MVTS: HyperNetwork-based Multivariate Time Series Forecasting30. APT: Affine Prototype-Timestamp for Time Series Forecasting Under Distribution Shift31. M3Time: LLM-Enhanced Multi-Modal, Multi-Scale, andMulti-Frequency Multivariate Time Series Forecasting32. Harmonic Dataset Distillation for Time Series Forecasting33. TimePMG: LLM-Based Time Series Forecasting with Period-Aware Multi-Scale Decomposition and Group-Wise Alignment34. TimePMG: LLM-Based Time Series Forecasting withPeriod-Aware Multi-Scale Decomposition and Group-WiseAlignment35. ProCAST: A Projection Framework for Coupled Aggregation Constrained Multivariate Time Series Forecasting

Oral

1 ReCast: Reliability-aware Codebook-assisted Lightweight Time Series Forecasting

链接：https://arxiv.org/abs/2511.11991

作者：Xiang Ma; Taihua Chen; Pengcheng Wang; Xuemei Li; Caiming Zhang

关键词：预测，轻量化

2 GraFT: Infusing Pre-trained Transformers with Relational Structure for Time Series Forecasting

作者：Yuqi Yuan; Xiong Luo; Qiaojuan Peng; Wenbing Zhao

关键词：预测，预训练Transformer，关系结构化

3 TimeCAP: A Channel-Aware Pre-Training Framework for Multivariate Time Series Forecasting

作者：Chuanru Ren; Yao Lu; Tianjin Huang; Haowen Zheng; Hengde Zhu; Yunyin Li; Hengxiao Li; Lu Liu

关键词：预测，通道感知，预训练

4 Beyond Missing Data Imputation: Information- Theoretic Coupling of Missingness and Class Imbalance for Optimal Irregular Time Series Classification

作者：Xin Qin; Mengna Liu; Wenjie Wang; Shuxin Li; Tianjiao Li; Xiufeng Liu; Xu Cheng

关键词：缺失值下的分类

5 CometNet: Contextual Motif-guided Long-term Time Series Forecasting

链接：https://arxiv.org/abs/2511.08049

作者：Weixu Wang; Xiaobo Zhou; Xin Qiao; Lei Wang; Tie Qiu

Poster

6 IdealTSF: Can Non-Ideal Data Contribute to Enhancing the Performance of Time Series Forecasting Models?

作者：Hua Wang; Jinghao Lu; Fan Zhang

关键词：预测，非理想数据

7 Time Series Forecasting via Direct Per-Step Probability Distribution Modeling

作者：Linghao Kong; Xiaopeng Hong

关键词：（概率）预测，分布建模

8 Revitalizing Canonical Pre-Alignment for Irregular Multivariate Time Series Forecasting

链接：https://www.arxiv.org/abs/2508.01971

作者：Ziyu Zhou; Yiming Huang; Yanyun Wang; Yuankai Wu; James Kwok; Yuxuan Liang

关键词：预测，不规则时间序列，对齐

9 Goal-Oriented Time-Series Forecasting: Foundation Framework Design

链接：https://arxiv.org/abs/2504.17493

作者：Luca-Andrei Fechete; Fadhel Ayed; Mohamed Sana; Nicola Piovesan; Wenjie Li; Antonio De Domenico; Tareq Si Salem

关键词：预测，目标导向

10 STrans:SpontaneousArchitecture Evolution for Adaptive Time Series Forecasting

作者：Haoyi Jia

关键词：预测，自发框架演化

11 MDMLP-EIA: Multi-domain Dynamic MLPs with Energy Invariant Attention for Time Series Forecasting

链接：https://arxiv.org/abs/2511.09924

作者：Hu Zhang; Zhien Dai; Zhaohui Tang; Yongfang Xie

关键词：预测，MLP，能量不变

12 Sonnet: Spectral Operator Neural Network for Multivariable Time Series Forecasting

链接：https://arxiv.org/abs/2505.15312

作者：Yuxuan Shu; Vasileios Lampos

关键词：预测，谱算子

13 FedSkeleton: Secure Multi-Party Graph Skeleton Construction for Privacy-Preserving Federated Time-Series

Forecasting

作者：Henggang Deng; Yuchao Tang; Wenjie Fu; Huandong Wang; Kun Chen; Tao Jiang

关键词：预测，安全多方，联邦学习，隐私保护

14 T3Time: Tri-Modal Time Series Forecasting via Adaptive Multi-Head Alignment and Residual Fusion

链接：https://arxiv.org/abs/2508.04251

作者：Abdul Monaf Chowdhury; Rabeya Akter; Safaeid Hossain Arib

关键词：预测，对齐，残差，多(3)模态

15 FreDN: Spectral Disentanglement for Time Series Forecasting via Learnable Frequency Decomposition

链接：https://www.arxiv.org/abs/2511.11817

作者：Zhongde An; Jinhong You; Jiyanglin Li; Yiming Tang; Wen Li; Heming Du; Shouguo Du

关键词：预测，谱域（频域），解耦

16 Sparse-Scale Transformer with Bidirectional Awareness for Time Series Forecasting

作者：Ying Liu; Bo Liu; Sheng Huang; Gang Luo; Wenbo Hu; Meng Wang; Richang Hong

关键词：预测，双向感知，稀疏化

17 OccamVTS: Distilling Vision Models to 1% Parameters for Time Series Forecasting

链接：https://arxiv.org/abs/2508.01727

作者：Sisuo Lyu; Siru Zhong; Weilin Ruan; Qingxiang Liu; Qingsong Wen; Hui Xiong; Yuxuan Liang

关键词：预测，视觉模型，蒸馏

18 Multi-Horizon Time Series Forecasting of non-parametric CDFs with Deep Lattice Networks

链接：https://arxiv.org/abs/2511.13756

作者：Niklas Erdmann; Lars Bentsen; Roy Stenbro; Heine Nygard Riise; Narada Dilp Warakagoda; Paal E. Engelstad

关键词：预测，累积分布函数（CDF）

19 Towards Non-Stationary Time Series Forecasting with Temporal Stabilization and Frequency Differencing

链接：https://arxiv.org/abs/2511.08229

作者：Junkai Lu; Peng Chen; Chenjuan Guo; Yang Shu; Meng Wang; Bin Yang

关键词：预测，非平稳时序，频域差分

20 TimeMosaic: Temporal Heterogeneity Guided Time Series Forecasting via Adaptive Granularity Patch and Segment-wise Decoding

链接：https://arxiv.org/abs/2509.19406

作者：Kuiye Ding; Fanda Fan; Chunyi Hou; Zheya Wang; Lei Wang; Zhengxin Yang; Jianfeng Zhan

关键词：预测，异质性

21 FeTS: A Feature-Aware Framework for Time Series Forecasting

作者：Le Wang; Jianyong Chen; Songbai Liu

关键词：预测，特征感知

22 XLinear: A Lightweight and Accurate MLP-Based Model for Long-Term Time Series Forecasting with Exogenous Inputs

作者：Xinyang Chen; Huidong Jin; Yu Huang; Zaiwen Feng

关键词：预测，MLP，外部变量

23 Beyond Immediate Activation: Temporally Decoupled Backdoor Attacks on Time Series Forecasting

作者：Zhixin Liu; Xuanlin Liu; Sihan Xu; Yaqiong Qiao; Ying Zhang; Xiangrui Cai

关键词：预测，后门攻击

24 DeepBooTS: Dual-Stream Residual Boosting for Drift-Resilient Time-Series Forecasting

链接：https://arxiv.org/abs/2511.06893

作者：Daojun Liang; Jing Chen; Xiao Wang; Yinglong Wang; Shuo Li

关键词：预测，概念漂移

25 EMAformer: Enhancing Transformer Through Embedding Armor for Time Series Forecasting

链接：https://arxiv.org/abs/2511.08396

作者：Zhiwei Zhang; Xinyi Du; Xuanchi Guo; Weihao Wang; Wenjuan Han

关键词：预测，Transformer

26 Optimal Look-back Horizon for Time Series Forecasting in Federated Learning

链接：https://arxiv.org/abs/2511.12791

作者：Dahao Tang; NAN YANG; YanliLi; Zhiyu Zhu; Zhibo Jin; Dong Yuan

关键词：预测，联邦学习，回溯窗口优化

27 M2FMoE: Multi-Resolution Multi-View Frequency Mixture-of-Experts for Extreme-Adaptive Time Series Forecasting

作者：Yaohui Huang; Runmin Zou; Yun Wang; Laeeq Aslam; Ruipeng Dong

关键词：预测，多视图，频域，混合专家系统

28 Rethinking Irregular Time Series Forecasting: A Simple Yet Effective Baseline

链接：https://arxiv.org/abs/2505.11250

作者：Xvyuan Liu; Xiangfei Qiu; Xingjian Wu; Zhengyu Li; Chenjuan Guo; Jilin Hu; Bin Yang

关键词：预测，不规则时序

29 HN-MVTS: HyperNetwork-based Multivariate Time Series Forecasting

链接：https://arxiv.org/abs/2511.08340

作者：Andrey Savchenko; Oleg Kachan

关键词：预测，超网络

30 APT: Affine Prototype-Timestamp for Time Series Forecasting Under Distribution Shift

链接：https://www.arxiv.org/abs/2511.12945

作者：Yujie Li; Zezhi Shao; Chengqing Yu; Yisong Fu; Tao Sun; Yongjun Xu; Fei Wang

关键词：预测，分布偏移

31 M3Time: LLM-Enhanced Multi-Modal, Multi-Scale, and

Multi-Frequency Multivariate Time Series Forecasting

作者：Shuning Jia; Baijun Song; Canming Ye; Chun Yuan

关键词：预测，多模态，多尺度，多频域，LLM

32 Harmonic Dataset Distillation for Time Series Forecasting

作者：Seungha Hong; Sanghwan Jang; Wonbin Kweon; Suyeon Kim; Gyuseok Lee; Hwanjo Yu

关键词：预测，谐波数据，蒸馏

33 TimePMG: LLM-Based Time Series Forecasting with Period-Aware Multi-Scale Decomposition and Group-Wise Alignment

作者：Feng Xiong; Zongxia Xie; Haoyu Wang; Yanru Sun; Jianhong Lin

关键词：预测，时空依赖性，熵谱

34 TimePMG: LLM-Based Time Series Forecasting with

Period-Aware Multi-Scale Decomposition and Group-Wise Alignment

作者：Zongjiang Shang; Binqing Wu; Dongliang Cui; Ling Chen

关键词：预测，分解，对齐，LLM

35 ProCAST: A Projection Framework for Coupled Aggregation Constrained Multivariate Time Series Forecasting

作者：Jiaqi Xue; Hongji Dong; Yucen Gao; Xiaofeng Gao; Guihai Chen

关键词：预测，耦合