OpenTS-FM: Time Series Foundation Models

OpenTS-FM is a series of time series foundation models, offering strong zero-shot and few-shot abilities, across data domains and analytics tasks.  This initiative addresses the Generalization challenge (the 'G' in the AGREE principles).

Time Series Foundation Models for Anomaly Detection

ICLR 2025 TFB Publication Thumbnail

DADA: Towards A General Time Series Anomaly Detector with Adaptive Bottlenecks And Dual Adversarial Decoders

Qichao Shentu, Beibu Li, Kai Zhao, Yang Shu, Zhongwen Rao, Lujia Pan, Bin Yang, Chenjuan Guo

International Conference on Learning Representations (ICLR), 2025.

[Bib]

Time Series Foundation Models for Forecasting

ICML 2025 FACTS Publication Thumbnail

LightGTS: A Lightweight General Time Series Forecasting Model

Yihang Wang, Yuying Qiu, Peng Chen, Yang Shu, Zhongwen Rao, Lujia Pan, Bin Yang, Chenjuan Guo

International Conference on Machine Learning (ICML), 2025.

[Bib]| [Project]

ICML 2025 FACTS Publication Thumbnail

ROSE: Towards a General Time Series Forecasting Model with Unified Representation and Adaptive Transfer

Yihang Wang, Yuying Qiu, Peng Chen, Kai Zhao, Yang Shu, Zhongwen Rao, Lujia Pan, Bin Yang, Chenjuan Guo

International Conference on Machine Learning (ICML), 2025.

[Bib]

Time Series Foundation Models for Classification

ICDE 2025 FACTS Publication Thumbnail

AimTS: Augmented Series and Image Contrastive Learning for Time Series Classification

Yuxuan Chen, Shanshan Huang, Yunyao Cheng, Peng Chen, Zhongwen Rao, Yang Shu, Bin Yang, Lujia Pan, Chenjuan Guo

International Conference on Data Engineering (ICDE), 2025.

[Bib] | [Project]