Generative Intervention Models for Causal Perturbation Modeling

Nora Schneider, Lars Lorch, Niki Kilbertus, Bernhard Schölkopf, Andreas Krause
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:53388-53412, 2025.

Abstract

We consider the problem of predicting perturbation effects via causal models. In many applications, it is a priori unknown which mechanisms of a system are modified by an external perturbation, even though the features of the perturbation are available. For example, in genomics, some properties of a drug may be known, but not their causal effects on the regulatory pathways of cells. We propose a generative intervention model (GIM) that learns to map these perturbation features to distributions over atomic interventions in a jointly-estimated causal model. Contrary to prior approaches, this enables us to predict the distribution shifts of unseen perturbation features while gaining insights about their mechanistic effects in the underlying data-generating process. On synthetic data and scRNA-seq drug perturbation data, GIMs achieve robust out-of-distribution predictions on par with unstructured approaches, while effectively inferring the underlying perturbation mechanisms, often better than other causal inference methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-schneider25a, title = {Generative Intervention Models for Causal Perturbation Modeling}, author = {Schneider, Nora and Lorch, Lars and Kilbertus, Niki and Sch\"{o}lkopf, Bernhard and Krause, Andreas}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {53388--53412}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/schneider25a/schneider25a.pdf}, url = {https://proceedings.mlr.press/v267/schneider25a.html}, abstract = {We consider the problem of predicting perturbation effects via causal models. In many applications, it is a priori unknown which mechanisms of a system are modified by an external perturbation, even though the features of the perturbation are available. For example, in genomics, some properties of a drug may be known, but not their causal effects on the regulatory pathways of cells. We propose a generative intervention model (GIM) that learns to map these perturbation features to distributions over atomic interventions in a jointly-estimated causal model. Contrary to prior approaches, this enables us to predict the distribution shifts of unseen perturbation features while gaining insights about their mechanistic effects in the underlying data-generating process. On synthetic data and scRNA-seq drug perturbation data, GIMs achieve robust out-of-distribution predictions on par with unstructured approaches, while effectively inferring the underlying perturbation mechanisms, often better than other causal inference methods.} }
Endnote
%0 Conference Paper %T Generative Intervention Models for Causal Perturbation Modeling %A Nora Schneider %A Lars Lorch %A Niki Kilbertus %A Bernhard Schölkopf %A Andreas Krause %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-schneider25a %I PMLR %P 53388--53412 %U https://proceedings.mlr.press/v267/schneider25a.html %V 267 %X We consider the problem of predicting perturbation effects via causal models. In many applications, it is a priori unknown which mechanisms of a system are modified by an external perturbation, even though the features of the perturbation are available. For example, in genomics, some properties of a drug may be known, but not their causal effects on the regulatory pathways of cells. We propose a generative intervention model (GIM) that learns to map these perturbation features to distributions over atomic interventions in a jointly-estimated causal model. Contrary to prior approaches, this enables us to predict the distribution shifts of unseen perturbation features while gaining insights about their mechanistic effects in the underlying data-generating process. On synthetic data and scRNA-seq drug perturbation data, GIMs achieve robust out-of-distribution predictions on par with unstructured approaches, while effectively inferring the underlying perturbation mechanisms, often better than other causal inference methods.
APA
Schneider, N., Lorch, L., Kilbertus, N., Schölkopf, B. & Krause, A.. (2025). Generative Intervention Models for Causal Perturbation Modeling. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:53388-53412 Available from https://proceedings.mlr.press/v267/schneider25a.html.

Related Material