Methodological Advances in Experimental Design

Drew Dimmery will discuss methodological advances in experimental design, emphasizing resolving issues before data collection.

They will cover three key areas: enhancing treatment assignment for better generalization of causal effects, improving the estimation of heterogeneous treatment effects by connecting design to the MAXCUT problem, and developing efficient online assignment procedures for sequential experiments. These advancements aim to optimize experimental outcomes and robustness, drawing on recent research and algorithmic innovations.

Organized by:

Institute of AI in Management LMU Munich