Tony Liu
Incoming Assistant Prof. @ Mount Holyoke College. CIS PhD @ UPenn, advised by Lyle Ungar and Konrad Kording. Scientist/PM @ Roblox.
Hello! My research sits at the intersection of causal inference and machine learning, with a focus on observational methods: how do we build an understanding of the world when we cannot run a randomized experiment?
I develop machine learning methods that improve traditional observational causal methodology in order to answer causal questions in complex data domains. Alongside my awesome collaborators, I use these methods across data science applications in public health, social communication, mental wellness, and medicine.
In the classroom, I emphasize process-based learning through completion-graded assignments and continual student feedback throughout the lifecycle of a course. I believe in understanding students’ individual starting points and engaging with them from there, being particularly mindful of those coming from diverse educational backgrounds outside of the sciences and engineering.
selected publications
- Measuring Causal Effects of Civil Communication Without RandomizationTo appear in International AAAI Conference on Web and Social Media (ICWSM), 2024
- Automated Detection of Causal Inference Opportunities: Regression Discontinuity Subgroup DiscoveryTo appear in Transactions of Machine Learning Research (TMLR), 2023
- Data-driven exclusion criteria for instrumental variable studiesConference on causal learning and reasoning (CLeaR), 2022
- The relationship between text message sentiment and self-reported depressionJournal of affective disorders, 2022
- Quantifying causality in data science with quasi-experimentsNature computational science, 2021
- Machine learning for phone-based relationship estimation: the need to consider population heterogeneityProceedings of the ACM on interactive, mobile, wearable and ubiquitous technologies, 2019