Seminars

View all Seminars  |  Download ICal for this event

Characterizations of threshold classifiers

Series: Bangalore Theory Seminars

Speaker: Mohit Sharma IIIT Delhi

Date/Time: Nov 29 11:00:00

Location: CSA Lecture Hall (Room No. 112, Ground Floor)

Abstract:
In this talk, we will first survey characterizations of Bayes optimal fair classifiers for commonly used fairness constraints. These can be typically written down as group-dependent or instance-dependent threshold classifiers on the posterior class probability. These characterizations help in investigating fairness-accuracy trade-offs and provably recovery from data bias. Finally, we state some advantages and disadvantages of working with threshold classifiers in practice. Much of the talk will be a survey of existing results in this area and recent work with Amit Deshpande (Microsoft Research India).


Microsoft teams link:

Link


We are grateful to the Kirani family (Link and the Walmart Center for Tech Excellence (Link for generously supporting this seminar series


Hosts: Rameesh Paul, KVN Sreenivas, Rahul Madhavan, Debajyoti Kar