Rethinking Importance Weighting for Transfer Learning

by   Nan Lu, et al.

A key assumption in supervised learning is that training and test data follow the same probability distribution. However, this fundamental assumption is not always satisfied in practice, e.g., due to changing environments, sample selection bias, privacy concerns, or high labeling costs. Transfer learning (TL) relaxes this assumption and allows us to learn under distribution shift. Classical TL methods typically rely on importance-weighting – a predictor is trained based on the training losses weighted according to the importance (i.e., the test-over-training density ratio). However, as real-world machine learning tasks are becoming increasingly complex, high-dimensional, and dynamical, novel approaches are explored to cope with such challenges recently. In this article, after introducing the foundation of TL based on importance-weighting, we review recent advances based on joint and dynamic importance-predictor estimation. Furthermore, we introduce a method of causal mechanism transfer that incorporates causal structure in TL. Finally, we discuss future perspectives of TL research.


page 23

page 27


Transfer Adaptation Learning: A Decade Survey

The world we see is ever-changing and it always changes with people, thi...

Rethinking Importance Weighting for Deep Learning under Distribution Shift

Under distribution shift (DS) where the training data distribution diffe...

A One-step Approach to Covariate Shift Adaptation

A default assumption in many machine learning scenarios is that the trai...

Dimension Reduction for Robust Covariate Shift Correction

In the covariate shift learning scenario, the training and test covariat...

Effects of sampling skewness of the importance-weighted risk estimator on model selection

Importance-weighting is a popular and well-researched technique for deal...

Mandoline: Model Evaluation under Distribution Shift

Machine learning models are often deployed in different settings than th...

When is Importance Weighting Correction Needed for Covariate Shift Adaptation?

This paper investigates when the importance weighting (IW) correction is...

Please sign up or login with your details

Forgot password? Click here to reset