`

Timezone: »

 
Poster
Online Adaptation to Label Distribution Shift
Ruihan Wu · Chuan Guo · Yi Su · Kilian Weinberger

Thu Dec 09 08:30 AM -- 10:00 AM (PST) @ None #None

Machine learning models often encounter distribution shifts when deployed in the real world. In this paper, we focus on adaptation to label distribution shift in the online setting, where the test-time label distribution is continually changing and the model must dynamically adapt to it without observing the true label. This setting is common in many real world scenarios such as medical diagnosis, where disease prevalences can vary substantially at different times of the year. Leveraging a novel analysis, we show that the lack of true label does not hinder estimation of the expected test loss, which enables the reduction of online label shift adaptation to conventional online learning. Informed by this observation, we propose adaptation algorithms inspired by classical online learning techniques such as Follow The Leader (FTL) and Online Gradient Descent (OGD) and derive their regret bounds. We empirically verify our findings under both simulated and real world label distribution shifts and show that OGD is particularly effective and robust to a variety of challenging label shift scenarios.

Author Information

Ruihan Wu (Cornell University)
Chuan Guo (Facebook AI Research)
Yi Su (Cornell University)
Kilian Weinberger (Cornell University)

More from the Same Authors