Jundong Li, University of Virginia assistant professor of electrical and computer engineering with joint appointments in computer science and the School of Data Science, shared advancements in graph mining at the 29th International Joint Conference on Artificial Intelligence, January 7-15.

Li and his co-authors probe causal inference from observational data that can benefit healthcare professionals, e-commerce entrepreneurs and online educators, among others. Their paper, IGNITE: A Minimax Game Toward Learning Individual Treatment Effects from Networked Observational Data, investigates whether graph information within observational data can be leveraged to control hidden confounding bias. They also discuss how to build more effective models that yield unbiased treatment effects estimation for each individual represented in the data.