I am a PhD candidate at School of Computing and Information (SCI), University of Pittsburgh. I work at PITT Computational Social Dynamics Lab (PICSO) directed by Dr. Yu-Ru Lin, and PITT Initiative on the Computational Social Science (PittCSS). I worked at Facebook (2019), Sumsung Research America (2018).

Research

My research lies in the fields of machine learning, deep learning, data/network science, and computational social science. I am motivated to understand dynamics and heterogeneity in complex social systems, and to support data-driven decisions for social good. I use methods including network analysis and modeling, statistical analysis and causal inference, data mining and machine learning techniques. My major research projects include:

• tackling spurious associations in practical observational studies (with a focus on Simpson’s paradox), and supporting humans to achieve better data-driven decisions.
• battling COVID-19 infodemic, including characterizing user susceptibility to COVID-19 misinformation on social media, as well as assessment of offline risk of COVID-19 from online misinformation.
• modeling and understanding the dynamics of information and disease in networks, as well as detecting anomaly in dynamic networks.

Check out my publications and CV for more information.

Updates

• 06/2022 I attended ICWSM 2022 @Atlanta, US and gave oral presentation.
• 11/2021 I defended my dissertation proposal and officially become a PhD candidate.
• 08/2021 I gave a presentation of our work at PaCSS 2021.
• 07/2021 Our paper is accepted by ICWSM 2022. This study seeks to understand online susceptible users, i.e., those who are likely to be attracted by, believe and spread COVID-19 misinformation, ranging from social bots to humans with various misinformation engagement levels.
• 03/2021 Our paper is published at IEEE Journal of Biomedical and Health Informatics. We utilize deep generative learning to address the disease progression uncertainty in diease forecasting problem.
• 01/2021 Our paper is published at Nature Communication. We provide a framework to optimize the selection of surveillance site locations and show that accurate forecasting of respiratory diseases for locations without surveillance is feasible.
• 08/2020 I gave a presentation of our work at Politics and Computational Social Science (PaCSS) 2020.