Education

  • Ph.D. in Statistics

    Rutgers University, 2018

  • B.Sc. in Computer Science and Economics

    Tsinghua University, Peking University

Experience

  • NYU, Stern, TOPS

    Assistant Professor, 09/2021 - present

  • UC Berkeley, EECS

    Postdoc Fellow, 09/2019 - 08/2021

  • Princeton University, ORFE

    Postdoc Fellow, 06/2018 - 08/2019

  • OpenAI, Robotics

    Scholar, 02/2019 - 05/2019

Interests

  • Tensor learning
  • Reinforcement learning
  • Transfer learning
  • Data Fusion and Diversity

Research

*
All Tensor Learning Reinforcement Learning Transfer Learning Business Applications

Developing efficient algorithms for tensor data is my primary focus, as tensors offer a more natural and comprehensive representation of the multi-dimensional physical world. My aim is to enhance the understanding and analysis of complex data structures through innovative approaches.

My research focuses on the design and development of reinforcement learning (RL) algorithms tailored for social applications in various domains, including business, education, and healthcare. By leveraging RL techniques, my collaborators and I aim to address specific challenges and opportunities that arise in these social contexts. Our goal is to create novel algorithms that can effectively optimize decision-making processes, improve outcomes, and contribute to the advancement of social applications.

My research centers around the concept of transferring knowledge from one or more source tasks to enhance learning performance in a target task. In particular, I focus on the application of transfer learning in the domains of reinforcement learning and tensor learning. By leveraging knowledge gained from related tasks, my collaborators and I aim to improve the efficiency and effectiveness of learning processes.

Featured Publications

[Click here for all publications] * [Google Scholar]

(2024). Data-Driven Knowledge Transfer in Batch $Q^*$ Learning. submitted.

PDF Code Dataset

(2024). STEFA: Semi-parametric tensor factor analysis by iteratively projected singular value decomposition. Journal of the Royal Statistical Society: Series B.

DOI PDF Code Dataset

(2024). Reinforcement Learning in Latent Heterogeneous Environments. Journal of the American Statistical Association.

DOI PDF Code Dataset

(2024). Tensor-view Topological Graph Neural Network. AISTATS, 2024.

DOI PDF Code Dataset

(2024). Time-Varying Matrix Factor Models. submitted.

DOI PDF Code Dataset

(2022). Community Network Auto-Regression for High-Dimensional Time Series. Journal of Econometrics.

PDF Code Dataset

(2021). Statistical Inference for High-Dimensional Matrix-Variate Factor Models. Journal of the American Statistical Association.

PDF Code Dataset

(2020). Constrained Factor Models for High-Dimensional Matrix-Variate Time Series. Journal of the American Statistical Association.

PDF Code Dataset

(2020). On Projection Robust Optimal Transport: Sample Complexity and Model Misspecification. AISTATS, 2021.

PDF Code Dataset

(2020). Estimation and Inference of High-Dimensional Matrix-Valued Threshold Factor Models. Scandinavian Journal of Statistics.

PDF Code Dataset

(2020). Transferred $Q$ Learning. Manuscript.

PDF Code Dataset

Contact

  • KMC 8-60, 44 West 4th Street, New York, NY 10012
  • Tuesday 12:00 to 14:00, Thursday 12:00 to 14:00
  • Book an appointment