Murali Haran, Pennsylvania State University

Title: Sample quality for intractable normalizing function algorithms

Date: Friday, May 22nd, 2026
Time: 1:30PM (PDT)
Location: ASB 10900

Abstract: Models with intractable normalizing functions—and thus partially intractable likelihoods—are central to many fields, including network analysis, spatial modeling, and genomics. Despite their prevalence, the most practical algorithms for these models often lack rigorous theoretical justification, leaving researchers without clear guidelines for tuning or quality assessment. In this talk, I will overview current algorithmic approaches and introduce new diagnostics designed to evaluate sample quality. Beyond assessment, these tools offer practical insights for parameter tuning and provide a deeper understanding of the performance of popular intractable likelihood methods.