Semi-generative modelling: Covariate-shift adaptation with cause and effect features

Julius von Kügelgen, Alexander Mey, Marco Loog

Research output: Contribution to journalConference articleScientificpeer-review

5 Citations (Scopus)
22 Downloads (Pure)


Current methods for covariate-shift adaptation use unlabelled data to compute importance weights or domain-invariant features, while the final model is trained on labelled data only. Here, we consider a particular case of covariate shift which allows us also to learn from unlabelled data, that is, combining adaptation with semi-supervised learning. Using ideas from causality, we argue that this requires learning with both causes, XC, and effects, XE, of a target variable, Y, and show how this setting leads to what we call a semi-generative model, P(Y,XE|XC,θ). Our approach is robust to domain shifts in the distribution of causal features and leverages unlabelled data by learning a direct map from causes to effects. Experiments on synthetic data demonstrate significant improvements in classification over purely-supervised and importance-weighting baselines.

Original languageEnglish
Number of pages9
JournalProceedings of Machine Learning Research
Publication statusPublished - 2020
Event22nd International Conference on Artificial Intelligence and Statistics, AISTATS 2019 - Naha, Japan
Duration: 16 Apr 201918 Apr 2019


Dive into the research topics of 'Semi-generative modelling: Covariate-shift adaptation with cause and effect features'. Together they form a unique fingerprint.

Cite this