AG 1, AG 2, AG 3, INET, AG 4, AG 5, D6, SWS, RG1, MMCI   Info

AG Audience

English

30 Minutes

As part of my research on Out-of-Distribution generalization, I reviewed the paper "Distributionally Robust Losses for Latent Covariate Mixtures". The premise of this paper is guaranteeing accuracy within a neighborhood of distributions centered at the training set distribution. The problem of optimizing worst-case loss in a set of distributions is extremely difficult. However, in this paper, an equivalent loss is introduced and then a tractable and convergent estimate of it is presented. In this presentation, I explain the problem setting, the underlying assumptions and at last, the theoretical results of this paper.

Ina Geisler

+49 681 9325 1802

--email hidden

Zoom

passcode not visible

logged in users only