12
Jun
Colloquium
Generalized Data Thinning Using Sufficient Statistics
Jacob Bien, University of Southern California, LA
12.06.2023
3:00 pm - 4:30 pm
LMU Department of Statistics and via zoom
Sample splitting, a common tool in data science, faces limitations. A recent method, convolution-closed data thinning, offers an alternative when sample splitting isn't feasible. This talk explores sufficiency as a key principle, leading to a unified framework called generalized data thinning.
Related
©jittawit.21 - stock.adobe.com
AI Keynote Series • 20.11.2025 • Online via Zoom
Distilling Heterogeneous Treatment Effects: Stable Subgroup Estimation in Causal Inference
Join the lecture with Melody Huang from Political Science and Statistics & Data Science at Yale University.
©jittawit.21 - stock.adobe.com
AI Keynote Series • 13.11.2025 • Online via Zoom
Personalized Care Through Causal & Federated Learning: From Data to Decisions
Join the lecture with Julie Josse from French National Instiute for Research in Digital Science and Technology (Inria).