Where is the Truth? The Risk of Getting Confounded in a Continual World

Florian Peter Busch, Roshni Ramanna Kamath, Rupert Mitchell, Wolfgang Stammer, Kristian Kersting, Martin Mundt
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:6046-6076, 2025.

Abstract

A dataset is confounded if it is most easily solved via a spurious correlation which fails to generalize to new data. In this work, we show that, in a continual learning setting where confounders may vary in time across tasks, the challenge of mitigating the effect of confounders far exceeds the standard forgetting problem normally considered. In particular, we provide a formal description of such continual confounders and identify that, in general, spurious correlations are easily ignored when training for all tasks jointly, but it is harder to avoid confounding when they are considered sequentially. These descriptions serve as a basis for constructing a novel CLEVR-based continually confounded dataset, which we term the ConCon dataset. Our evaluations demonstrate that standard continual learning methods fail to ignore the dataset’s confounders. Overall, our work highlights the challenges of confounding factors, particularly in continual learning settings, and demonstrates the need for developing continual learning methods to robustly tackle these.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-busch25a, title = {Where is the Truth? {T}he Risk of Getting Confounded in a Continual World}, author = {Busch, Florian Peter and Kamath, Roshni Ramanna and Mitchell, Rupert and Stammer, Wolfgang and Kersting, Kristian and Mundt, Martin}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {6046--6076}, 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/busch25a/busch25a.pdf}, url = {https://proceedings.mlr.press/v267/busch25a.html}, abstract = {A dataset is confounded if it is most easily solved via a spurious correlation which fails to generalize to new data. In this work, we show that, in a continual learning setting where confounders may vary in time across tasks, the challenge of mitigating the effect of confounders far exceeds the standard forgetting problem normally considered. In particular, we provide a formal description of such continual confounders and identify that, in general, spurious correlations are easily ignored when training for all tasks jointly, but it is harder to avoid confounding when they are considered sequentially. These descriptions serve as a basis for constructing a novel CLEVR-based continually confounded dataset, which we term the ConCon dataset. Our evaluations demonstrate that standard continual learning methods fail to ignore the dataset’s confounders. Overall, our work highlights the challenges of confounding factors, particularly in continual learning settings, and demonstrates the need for developing continual learning methods to robustly tackle these.} }
Endnote
%0 Conference Paper %T Where is the Truth? The Risk of Getting Confounded in a Continual World %A Florian Peter Busch %A Roshni Ramanna Kamath %A Rupert Mitchell %A Wolfgang Stammer %A Kristian Kersting %A Martin Mundt %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-busch25a %I PMLR %P 6046--6076 %U https://proceedings.mlr.press/v267/busch25a.html %V 267 %X A dataset is confounded if it is most easily solved via a spurious correlation which fails to generalize to new data. In this work, we show that, in a continual learning setting where confounders may vary in time across tasks, the challenge of mitigating the effect of confounders far exceeds the standard forgetting problem normally considered. In particular, we provide a formal description of such continual confounders and identify that, in general, spurious correlations are easily ignored when training for all tasks jointly, but it is harder to avoid confounding when they are considered sequentially. These descriptions serve as a basis for constructing a novel CLEVR-based continually confounded dataset, which we term the ConCon dataset. Our evaluations demonstrate that standard continual learning methods fail to ignore the dataset’s confounders. Overall, our work highlights the challenges of confounding factors, particularly in continual learning settings, and demonstrates the need for developing continual learning methods to robustly tackle these.
APA
Busch, F.P., Kamath, R.R., Mitchell, R., Stammer, W., Kersting, K. & Mundt, M.. (2025). Where is the Truth? The Risk of Getting Confounded in a Continual World. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:6046-6076 Available from https://proceedings.mlr.press/v267/busch25a.html.

Related Material