Research

My research focuses on probabilistic models and neural networks for learning problems arising from real applications. My research topics include variational inference, representation learning, and attention networks. See my publication list.

Research highlights

• Graph Neural Networks for Inference in Gaussian processes and Spatial Data Modeling

  Idea: consider strong correlations only and run inference within neighborhoods with a shared Graph Neural Network

  Benefits: accurate & fast

  Publications: [Liu et al., AISTATS 2019], [Liu et al., ACML 2020], [Appleby et al., AAAI2020]

• Graph Generative Models

  1. the first rigorous probablistic treatment of autoregressive generative models for graphs

     Publications: [Chen et al., ICML 2021]

• A Probablistic Approach for Graph Matching

  1. optimizing a distribution of graph matchings

     Publications: [Liu et al., ICML 2021]

• Anomaly Detection

  1. Using GAN ensemble for anomaly detection: significantly improve the performance over single models

     Publications:[Han et al., AAAI 2021]

Teaching

• Fall 2021: CS 137 Deep Neural Networks
• Fall 2020: COMP 137 Deep Neural Networks
• Fall 2019: COMP 150 Machine Learning for Graph Analytics
• Spring 2019: COMP 150 Deep Neural Networks
• Fall 2018: COMP 135 Introduction to Machine Learning
• Spring 2018: COMP 135 Introduction to Machine Learning
• Fall 2017: COMP 150-01 Machine Learning for Ecology and Sustainability

Bio Sketch

I received my B.S. in computer science from Hebei University of Technology in 2006. After three years study in LAMDA group, I received my M.S. degree from Nanjing Univeristy in 2009. My advisor was Prof. Yuan Jiang and Zhi-Hua Zhou. Then I worked in Alibaba for one year and a half. After that, I went to Oregon State University and completed my PhD degree in 2016. My PhD advisor was Prof. Thomas Dietterich. Prior to joining Tufts, I worked as a postdoc researcher at Columbia University working with Prof. David Blei.