A0619
Title: Challenges in causality: Sensitivity analysis and automation
Authors: Carlos Cinelli - University of Washington (United States) [presenting]
Abstract: Causal inference has emerged as a central research area at the intersection of statistics and computer science, with numerous applications in empirical fields such as biomedicine and the social sciences. However, despite significant progress over the past three decades, many important problems remain open. The aim is to highlight two critical research directions: (i) drawing valid conclusions when key assumptions, such as the absence of unobserved confounding, are violated, and (ii) automating the causal inference pipeline. Key open challenges are outlined in the areas of sensitivity analysis and partial identification, as well as the automation of such tasks. Addressing these challenges is essential for making robust causal inferences under realistic settings.
