Network interference is a fundamental driving force behind phenomena such as intervention spillover, behavioral contagion, and information diffusion in interconnected systems. In this project, the investigator aims to develop statistical methods to understand and quantify whether, and to what extent, an individual’s behavior or status is influenced by others through network interactions. This project focuses on two central challenges in estimating network interference from observational data: individual heterogeneity and network confounding. To overcome the challenges, the investigator will develop an expressive and interpretable statistical model that adaptively learn how the effect of interference depends on individual-level contextual and network information. This model is expected to provide insights and an analytics tool towards the mechanisms of network interference. In addition, the investigator will design novel estimators to measure the causal effects of network interference, leveraging advanced machine learning techniques to address complex confounding among connected individuals and to make efficient use of limited experimental data. The methodologies developed in this project will advance the fields of causal inference and graph-based machine learning. The tools developed will have broad applicability to network data in social science, public health, political science, economics, and business, and will support new theoretical developments in areas such as social i