Wenwu Gong龚文武

Department of Mathematics, School of Mathematics and Computer Science, Nanchang University

I work on interpretable learning driven by mathematical priors: low-rank tensor representation, model-guided neural methods, and spatiotemporal modeling for computational imaging and recommendation.

Interpretable AI
Low-rank tensor representation
Spatiotemporal modeling
Computational imaging

Previously a SUSTech Presidential Postdoctoral Fellow in the Department of Statistics and Data Science (Sep. 2024 – Sep. 2026). Ph.D. in Mathematics, SUSTech, 2024.

News

2026 Joined the Department of Mathematics, School of Mathematics and Computer Science, Nanchang University.
2026 3DeepRep was published in Neurocomputing.
Mar. 2026 Talk on CATNN at IPDL, Chongqing.
2026 Three conference papers: CATNN at ICIP, and AWD-DPT and STTFN at ICIC.
2025 First prize, Innovative Application Award, Disaster Prevention Science and Technology Award.
2025 POI-LLM received the Best Paper Award at ICCBD+AI. TuckerAPP appeared in Signal Processing.

Selected papers

3DeepRep
Yunshan Li, Wenwu Gong, Chao Wang, Lili Yang. Neurocomputing, 678:133160, 2026. JCR Q1DOI
TuckerAPP
Wenwu Gong, Zhejun Huang, Jiaxin Lu, Lili Yang. Signal Processing, 240:110383, 2026. JCR Q2PDF Code
ARTD
Wenwu Gong, Zhejun Huang, Lili Yang. Applied Mathematical Modelling, 123:75–86, 2023. JCR Q1PDF Code

Full publication list

Research

Interpretable artificial intelligence

Hybrid models that keep a mathematical object inside the learner: neural differential equations, supported by the China Postdoctoral Science Foundation, and causal representation learning for imaging and recommendation (ProME, CLEAR).

Low-rank tensor representation

Model-driven Tucker methods (LRTuckerRep, GradientTD, ARTD) and learning-guided self-supervised models (3DeepRep, CATNN) for multi-dimensional recovery.

Spatiotemporal modeling

Prior-driven tensor recovery (TuckerAPP, LSPTD, LBNTD, BACP) and decoupled predictors (POI-LLM, STTFN, AWD-DPT) that separate long-term structure from short-term events.

Projects and funding

Appointment and education

2026 – Department of Mathematics, School of Mathematics and Computer Science, Nanchang University.
2024 – 2026 SUSTech Presidential Postdoctoral Fellow, Department of Statistics and Data Science.
2020 – 2024 Ph.D. in Mathematics, Department of Statistics and Data Science, SUSTech. Low-rank tensor representation.
2018 – 2020 M.Sc. in Mathematics, School of Mathematics, Harbin Institute of Technology. Probability and mathematical statistics.
2014 – 2018 B.Sc. in Mathematics and Applied Mathematics, Nanchang University.

Teaching

Probability Theory and Mathematical Statistics, teaching assistant, Sep. 2019 – May 2021. Lectures and exercise sessions for 120 students.
Tensor Decomposition and Its Applications, teaching assistant, Mar. 2023 – Jun. 2023. Tensor learning for 30 undergraduates.

Mentorship

Xinyu Shi (施忻雨), M.Sc., Sep. 2023 – Jun. 2026. Multimodal large models for spatiotemporal data. Next: doctoral program, Pengcheng Laboratory.
Xiongpeng Xia (夏雄鹏), M.Sc., Sep. 2023 – Jun. 2026. Multi-scale Transformers for long-horizon forecasting.
Jiaxin Lu (陆嘉欣), M.Sc., Sep. 2022 – Jun. 2024. SUSTech Top Ten Graduate. Tensor learning for urban traffic. Next: Peking University, Shenzhen.
Rongping Huang (黄荣平), M.Sc., Sep. 2021 – Jun. 2023. Bayesian tensor recovery for traffic imputation.

Talks

Mar. 2026 CATNN. IPDL, Chongqing.
Nov. 2025 LRTuckerRep. DMIP, Tokyo.
Jan. 2025 LBNTD. TRB Annual Meeting, Washington, D.C.
Mar. 2024 TuckerAPP. Graduate Academic Forum, SUSTech. First prize.