Biography
I am a Professor and Ph.D. Supervisor at the School of Data Science & Engineering, East China Normal University, and a recipient of the National Natural Science Fund for Excellent Young Scientists Fund Program (Overseas). I received my Ph.D. in Computer Science from the Department of Computer Science at the University of Manchester, UK, and previously served as a tenured Associate Professor in the Department of Computer Science at Aalborg University, Denmark.
My research interests lie in data management and data analytics, with applications in intelligent transportation, digital energy, and smart water resource management. I have published extensively in leading international conferences and journals, including Nature Communications, SIGMOD, VLDB, ICDE, AAAI, KDD, WWW, ICML, NeurIPS, ICLR, VLDB Journal, and TKDE. I serve as a Program Committee member and Area Chair for numerous prestigious conferences, including ICML, NeurIPS, ICLR, AAAI, ICDE, KDD, and PVLDB.
Research Interests
Data Management and Analytics, Machine Learning
Time Series Analytics, Spatio-Temporal Data Analytics, AI for Science
Time Series Foundation Models, AutoML, Multi-Agents
Selected Publications
Full list: DBLP, Google Scholar.
- 2026
- PATRA: Pattern-Aware Alignment and Balanced Reasoning for Time Series Question Answering. ICML
- KITE: Knowledge-Guided Probabilistic Modeling for Time Series Forecasting with Exogenous Variables. ICML
- TeamWork: Multivariate Time Series Anomaly Detection via Asymmetric Role-aware Channel Modeling. ICML
- Invariant Representation Learning for Source-Free Time Series Forecasting with LLM-Centric Proxy Denoising. ICML
- Aurora: Towards Universal Generative Multimodal Time Series Forecasting. ICLR
- SwiftTS: A Swift Selection Framework for Time Series Pre-trained Models via Multi-task Meta-Learning. ICLR
- GCGNet: Graph-Consistent Generative Network for Time Series Forecasting with Exogenous Variables. ICLR
- ARROW: An Adaptive Rollout and Routing Method for Global Weather Forecasting. ICLR
- Unlocking the Value of Text: Event-Driven Reasoning and Multi-Level Alignment for Time Series. ICLR
- ASTGI: Adaptive Spatio-Temporal Graph Interactions for Irregular Multivariate Time Series Forecasting. ICLR
- CoRA: Boosting Time Series Foundation Models for Multivariate Forecasting through Correlation-aware Adapter. ICLR
- Towards Multimodal Time Series Anomaly Detection with Semantic Alignment and Condensed Interaction. ICLR
- Towards Non-Stationary Time Series Forecasting with Temporal Stabilization and Frequency Differencing. AAAI
- MoST: A Foundation Model for Multi-modality Spatio-temporal Traffic Prediction. KDD
- CCD: Capturing Cross-Correlations with Deformable Convolutional Networks for Multivariate Time Series Forecasting. KDD
- TiWeaver: Unified Temporal Dynamics Modeling via Contextual Patching. KDD
- TimeBlocks: Versatile and Continual Time-Series Blockbase. KDD
- 2025
- Predicting microbial community structure and temporal dynamics by using graph neural network models. Nature Communications
- CrossAD: Time Series Anomaly Detection with Cross-scale Associations and Cross-window Modeling. NeurIPS
- Enhancing Time Series Forecasting through Selective Representation Spaces: A Patch Perspective. NeurIPS Spotlight
- Learning to Factorize Spatio-Temporal Foundation Models. NeurIPS Spotlight
- DBLoss: Decomposition-based Loss Function for Time Series Forecasting. NeurIPS
- Unsupervised Time Series Anomaly Prediction with Importance-based Generative Contrastive Learning. KDD
- LightGTS: A Lightweight General Time Series Forecasting Model. ICML
- Towards a General Time Series Forecasting Model with Unified Representation and Adaptive Transfer. ICML
- K^2VAE: A Koopman-Kalman Enhanced Variational AutoEncoder for Probabilistic Time Series Forecasting. ICML
- TSFM-Bench: A Comprehensive and Unified Benchmark of Foundation Models for Time Series Forecasting. KDD
- TAB: Unified Benchmarking of Time Series Anomaly Detection Methods. PVLDB
- Spatio-Temporal Prediction on Streaming Data: A Unified Federated Continuous Learning Framework. TKDE
- Gaussian Process Latent Variable Modeling for Few-Shot Time Series Forecasting. TKDE
- Assessing Pre-Trained Models for Transfer Learning Through Distribution of Spectral Components. AAAI
- Enhancing Diversity for Data-free Quantization. CVPR ORAL
- AimTS: Augmented Series and Image Contrastive Learning for Time Series Classification. ICDE
- AID-SQL: Adaptive In-Context Learning of Text-to-SQL with Difficulty-Aware Instruction and Retrieval-Augmented Generation. ICDE
- Data Driven Decision Making with Time Series and Spatio-Temporal Data. ICDE
- EasyTime: Time Series Forecasting Made Easy. ICDE
- Learning Generalizable Skills from Offline Multi-Task Data for Multi-Agent Cooperation. ICLR
- Towards a General Time Series Anomaly Detector with Adaptive Bottlenecks and Dual Adversarial Decoders. ICLR
- Air Quality Prediction with Physics-Guided Dual Neural ODEs in Open Systems. ICLR
- CATCH: Channel-Aware Multivariate Time Series Anomaly Detection via Frequency Patching. ICLR
- DUET: Dual Clustering Enhanced Multivariate Time Series Forecasting. KDD
- MM-Path: Multi-modal, Multi-granularity Path Representation Learning. KDD
- 2024
- Fully Automated Correlated Time Series Forecasting in Minutes. PVLDB
- A Memory Guided Transformer for Time Series Forecasting. PVLDB
- Noise Matters: Cross Contrastive Learning for Flink Anomaly Detection. PVLDB
- RCRank: Multimodal Ranking of Root Causes of Slow Queries in Cloud Database Systems. PVLDB
- TFB: Towards Comprehensive and Fair Benchmarking of Time Series Forecasting Methods. PVLDB
- QCore: Data-Efficient, On-Device Continual Calibration for Quantized Models. PVLDB
- Efficient Stochastic Routing in Path-Centric Uncertain Road Networks. PVLDB
- AutoCTS++: zero-shot joint neural architecture and hyperparameter search for correlated time series forecasting. VLDB Journal
- A Unified Replay-Based Continuous Learning Framework for Spatio-Temporal Prediction on Streaming Data. ICDE
- Routing with Massive Trajectory Data. ICDE
- Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecasting. ICLR
- 2023
- AutoCTS+: Joint Neural Architecture and Hyperparameter Search for Correlated Time Series Forecasting. Proc. ACM Manag. Data
- LightTS: Lightweight Time Series Classification with Adaptive Ensemble Distillation. Proc. ACM Manag. Data
- MagicScaler: Uncertainty-aware, Predictive Autoscaling. PVLDB
- LightPath: Lightweight and Scalable Path Representation Learning. KDD
- CGF: A Category Guidance Based PM2.5 Sequence Forecasting Training Framework. TKDE
- 2022
- RetroGraph: Retrosynthetic Planning with Graph Search. KDD
- Context-Aware Path Ranking in Road Networks. TKDE
- Anomaly Detection in Time Series with Robust Variational Quasi-Recurrent Autoencoders. ICDE
- Weakly-supervised Temporal Path Representation Learning with Contrastive Curriculum Learning. ICDE
- Towards Spatio-Temporal Aware Traffic Time Series Forecasting. ICDE
- Robust and Explainable Autoencoders for Unsupervised Time Series Outlier Detection. ICDE
- Triformer: Triangular, Variable-Specific Attentions for Long Sequence Multivariate Time Series Forecasting. IJCAI
- Outlier Detection for Streaming Task Assignment in Crowdsourcing. WWW
Ongoing Projects
- Agent-based planning and optimization for complex time series tasks, Huawei University-Enterprise Collaboration Project, 2026-2027.
- Multimodal industrial time series interpretation and semantic description, Huawei University-Enterprise Collaboration Project, 2025-2027.
- Intelligent risk inspection for cloud platforms, Alibaba University-Enterprise Collaboration Project, 2025-2026.
- Multi-dimensional covariate fusion representation and fine-tuning enhancement, Huawei University-Enterprise Collaboration Project, 2025-2026.
- Explainable automatic anomaly prediction for time series, National Natural Science Foundation of China, 2024-2027.