This is an online event. Registration is required.
Title: Bayesian Group Testing
Speaker: Prof. Marco Cuturi (Google Brain / ENSAE) https://marcocuturi.net/
Abstract: In this talk I will present recent work carried out to improve the performance of group testing using Bayesian analysis. The idea of group testing is that one can do better that testing people individually for a rare disease by pooling samples, and detecting whether one or possibly more people in a group are positive with a single test. We show in this work that a combination of sequential monte carlo with greedy combinatorial optimizers yields result that beat state of the art approaches.
Additional material : https://ai.googleblog.com/2020/07/exploring-faster-screening-with-fewer.html?m=1
|Date||September 16, 2020 (Wed) 16:30 - 17:30|