Research

Publications

* denotes equal contribution. Click a figure to enlarge it.

SDFlow: Similarity-Driven Flow Matching for Time Series Generation

Li, W.*, Feng, S.*, Wu, P., Gao, X., Wu, M., & Zhao, P.

Conference on Neural Information Processing Systems (NeurIPS 2026)

CCF-AFull PaperPosterFirst Author

SDFlow introduces a similarity-driven, non-autoregressive flow-matching framework for time-series generation in high-dimensional discrete latent spaces. By learning a low-rank subspace and initializing generation with similarity-guided manifold anchors, it aligns the generative process with the geometry of real temporal data and supports parallel synthesis.

arXiv

@misc{li2026sdflow,
  title={SDFlow: Similarity-Driven Flow Matching for Time Series Generation}, 
  author={Li, Wei and Feng, Shibo and Wu, Pengcheng and Gao, Xingyu and Wu, Min and Zhao, Peilin},
  year={2026},
  eprint={2605.05736},
  archivePrefix={arXiv},
  primaryClass={cs.AI},
  url={https://arxiv.org/abs/2605.05736}, 
}

Adaptive Prototypical Contrastive Learning for Time Series Clustering

Li, W.

ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2026)

CCF-AFull PaperPosterSole Author

Addresses time-series clustering with an unknown number of clusters by combining hierarchical prototypes, contrastive learning, and the MDL principle to jointly learn representations and cluster cardinality.

ACM DL Code

@inproceedings{li2026apcl,
  title={Adaptive Prototypical Contrastive Learning for Time Series Clustering},
  author={Li, Wei},
  booktitle={ACM SIGKDD International Conference on Knowledge Discovery and Data Mining},
  year={2026},
  doi={10.1145/3770855.3817773},
  url={https://dl.acm.org/doi/10.1145/3770855.3817773}
}

ClusterPatchTST: Uncertainty-Aware Causal Clustering for Heterogeneous Time Series Forecasting

Li, W.

International Conference on Database Systems for Advanced Applications (DASFAA 2026)

CCF-BFull PaperSole Author

Uses causal clustering to capture cross-series structural heterogeneity and uncertainty modeling to improve robust and reliable forecasting.

DOI

@inproceedings{li2026clusterpatchtst,
  title={ClusterPatchTST: Uncertainty-Aware Causal Clustering for Heterogeneous Time Series Forecasting},
  author={Li, Wei},
  booktitle={International Conference on Database Systems for Advanced Applications},
  year={2026}
}

EnergyPatchTST: Multi-scale Time Series Transformers with Uncertainty Estimation for Energy Forecasting

Li, W., Wang, Z., Sun, Q., Gao, Q., & Yang, F.

International Conference on Intelligent Computing (ICIC 2025)

CCF-COralFirst Author

Combines multiscale Patch Transformers with uncertainty estimation for energy forecasting; the platform received a software copyright and 300+ GitHub stars.

DOI Code

@inproceedings{li2025energypatchtst,
  title={EnergyPatchTST: Multi-scale Time Series Transformers with Uncertainty Estimation for Energy Forecasting},
  author={Li, Wei and Wang, Zixin and Sun, Qizheng and Gao, Qixiang and Yang, Fenglei},
  booktitle={International Conference on Intelligent Computing},
  pages={319--330},
  year={2025},
  organization={Springer}
}

ScatterFusion: A Hierarchical Scattering Transform Framework for Enhanced Time Series Forecasting

Li, W.

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2026)

CCF-BSole Author

Combines hierarchical scattering transforms with multiscale fusion to extract stable time-frequency representations under noise and scale variation.

DOI

@inproceedings{li2026scatterfusion,
  title={ScatterFusion: A Hierarchical Scattering Transform Framework for Enhanced Time Series Forecasting},
  author={Li, Wei},
  booktitle={IEEE International Conference on Acoustics, Speech, and Signal Processing},
  year={2026}
}

AWGFormer: Adaptive Wavelet-Guided Transformer for Multi-resolution Time Series Forecasting

Li, W.

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2026)

CCF-BSole Author

This work integrates adaptive wavelet guidance with Transformers for multi-resolution time series forecasting.

DOI

@inproceedings{li2026awgformer,
  title={AWGFormer: Adaptive Wavelet-Guided Transformer for Multi-resolution Time Series Forecasting},
  author={Li, Wei},
  booktitle={IEEE International Conference on Acoustics, Speech, and Signal Processing},
  year={2026}
}

SWIFT: State-space Wavelet Integrated Forecasting Technology for Enhanced Time Series Prediction

Li, W.

International Conference on Artificial Neural Networks (ICANN 2025)

CCF-COralSole Author

This work combines state-space modeling and wavelet-based multi-scale modules for enhanced time series forecasting.

DOI Code

@inproceedings{li2025swift,
  title={SWIFT: State-Space Wavelet Integrated Forecasting Technology for Enhanced Time Series Prediction},
  author={Li, Wei},
  booktitle={International Conference on Artificial Neural Networks},
  pages={29--40},
  year={2025},
  organization={Springer}
}

TimeFlowDiffuser: A Hierarchical Diffusion Framework with Adaptive Context Sampling for Multi-Horizon Time Series Forecasting

Li, W.

International Conference on Artificial Neural Networks (ICANN 2025)

CCF-COralSole Author

This work introduces a hierarchical diffusion framework with adaptive context sampling for multi-horizon time series forecasting.

DOI

@inproceedings{li2025timeflowdiffuser,
  title={TimeFlowDiffuser: A Hierarchical Diffusion Framework with Adaptive Context Sampling for Multi-horizon Time Series Forecasting},
  author={Li, Wei},
  booktitle={International Conference on Artificial Neural Networks},
  pages={241--252},
  year={2025},
  organization={Springer}
}

LWSpace: Multi-Scale State Space Framework for Time Series Forecasting

Li, W.

International Conference on Intelligent Computing (ICIC 2025)

CCF-COralSole Author

This work proposes a multi-scale state space framework for efficient time series forecasting.

DOI

@inproceedings{li2025lwspace,
  title={LWSpace: A Multi-scale State Space Framework for Enhanced Time Series Forecasting},
  author={Li, Wei},
  booktitle={International Conference on Intelligent Computing},
  pages={295--306},
  year={2025},
  organization={Springer}
}

Olympic Medal Prediction via Adaptive Triple-Fusion: Combining AAC Model with ARIMA-State Space Dynamics

Li, W., Wang, Z., Gao, Q., & Yang, F.

International Conference on Applied Mathematics, Modelling and Intelligent Computing (CAMMIC 2025)

EI IndexPosterFirst Author

This work combines adaptive triple-fusion modeling with ARIMA-state space dynamics for Olympic medal prediction.

DOI

@inproceedings{li2025olympic,
  title={Olympic Medal Prediction via Adaptive Triple-Fusion: Combining AAC Model with ARIMA-State Space Dynamics},
  author={Li, Wei and Wang, Zixin and Gao, Qixiang and Yang, Fenglei},
  booktitle={Proceedings of the 2025 5th International Conference on Applied Mathematics, Modelling and Intelligent Computing},
  pages={699--703},
  year={2025}
}