About Me

I’m currently a Senior Researcher at Microsoft Research Asia. I received my Ph.D. in Computer Science in Summer 2024 from the University of Utah. Prior to that, I earned my B.S. in both Statistics and Computer Science (double degree) from the School of the Gifted Young (少年班学院) at the University of Science and Technology of China (USTC) in 2018, and my Master’s in Computer Science from Temple University in 2019.

Research Interests

• Bayesian Methods and Generative Models
• Scientific Machine Learning for Physical Signal
• Tensor Learning (flexible low-rank representation for high-order data)
• Advanced Time Series (Tensor-valued Time Series, Spatial-Temporal data)

Industry & Application

• AI4Science: I’ve been involved in several research and start-up projects on AI + medical/bio scenarios. I own two first-author patents in AIDD (AI for Drug Discovery), which contribute as core patents for a start-up that has raised over $10M+.

• AI4Finance: I have quite experiences of AI-based quantitative trading/pricing strategy development during the internship/competition in top hedge fund and investment bank (Morgan Stanley, World Quant and UBS etc.). During the Microsoft I’m a core contributor of (MarS)(github 1.4K star), an open-source financial market simulation engine powered by generative foundation Model.

• Others: I am a dedicated writer to introduce Bayesian machine learning (in Chinese) with 15,000 followers on Zhihu (知乎) — Chinese Quora. See my Zhihu page.

News

• May 2025: We release R&D-Agent, which is verified by OpenAI-MLE-Bench! as Best MLE Agent and got Star 5.8k+ stars on Github! !
• Spring 2025: Two Papers accepted by KDD 2025 and ICLR 2025!
• May 2024: I will join Microsoft Research Asia as a Senior Researcher in 2024 summer!

Selected Publications

BayOTIDE: Bayesian Online Multivariate Time series Imputation with functional decomposition
Shikai Fang, Qingsong Wen, Yingtao Luo, Shandian Zhe, Liang Sun
ICML 2024 Spotlight! (top 3%)
[arXiv] [code] [slides] [知乎]
TL;DR: [Time Series & Bayesian] Efficient online imputation for functional multivariate time series.

Solving High Frequency and Multi-Scale PDEs with Gaussian Processes
Shikai Fang*, Madison Cooley*, Da Long*, Shibo Li, Robert Kirby, Shandian Zhe
ICLR 2024 [arXiv] [code] [slides] [知乎]
TL;DR: [AI4Phy & Bayesian] Gausisan Process based surrogate model for High-freq. PDE

Functional Bayesian Tucker Decomposition for Continuous-indexed Tensor
Shikai Fang, Xin Yu, Zheng Wang, Shibo Li, Mike Kirby, Shandian Zhe,
ICLR 2024 [arXiv] [code] [slides]
TL;DR: [Tensor & AI4Phy] Functional Tucker Decomposition for Physical Signal reconstruction

Streaming Factor Trajectory Learning for Temporal Tensor Decomposition
Shikai Fang, Xin Yu, Shibo Li, Zheng Wang, Robert Kirby, Shandian Zhe
NeurIPS 2023
[arXiv] [code] [slides]
TL;DR: [Tensor & Time Series] Online Trajectory Learning for Tensor-Valued Time Series

Dynamic Tensor Decomposition via Neural Diffusion-Reaction Processes
Zheng Wang*, Shikai Fang*, Shibo Li, Shandian Zhe
NeurIPS 2023 Spotlight! (top 10%) [arXiv] [code]
TL;DR:[Tensor & Time Series] Temporal Tensor <=> Diffusion-Reaction Process on Hypergraph

Provably Convergent Schrodinger Bridge with Applications to Probabilistic Time Series Imputation
Yu Chen*, Wei Deng*,Shikai Fang*, Fengpei Li*, Tianjiao Nicole Yang, Yikai Zhang, Kashif Rasul, Shandian Zhe, Anderson Schneider, and Yuriy Nevmyvaka,
ICML 2023 [arXiv] [code]
TL;DR: [Generative & Time Series] Schrodinger Bridge for Time Series Imputation

Bayesian Continuous-Time Tucker Decomposition
Shikai Fang, Akil Narayan, Robert Kirby, Shandian Zhe
ICML 2022 Oral! (top 2%) [paper] [code] [slides]
TL;DR: [Tensor & Time Series] Continuous-Time Tucker Model for Tensor-Valued Time Series

Streaming Bayesian Deep Tensor Factorization
Shikai Fang, Zheng Wang, Zhimeng Pan, Ji Liu, and Shandian Zhe
ICML 2021 [paper] [code] [supp]
TL;DR: [Tensor & Online learning] Bayesian deep learning + Tensor

Bayesian Streaming Sparse Tucker Decomposition
Shikai Fang, Robert. M. Kirby, and Shandian Zhe
UAI 2021 [paper] [code] [supp]
TL;DR: [Tensor & Online learning] Sparse Tucker + Online Learning