Pooling Image Datasets With Multiple Covariate Shift and Imbalance
This work addresses the challenge of handling multiple covariate shifts in overparameterized models for medical imaging and other domains, providing a more efficient method compared to existing invariant representation learning techniques.
The paper tackles the problem of pooling image datasets with multiple covariate shifts and imbalances, which is common in multi-institutional studies, by proposing a Category theory-based solution that avoids complex multi-stage training pipelines. The approach is validated through extensive experiments on real datasets and offers a unified perspective on various problem settings.
Small sample sizes are common in many disciplines, which necessitates pooling roughly similar datasets across multiple institutions to study weak but relevant associations between images and disease outcomes. Such data often manifest shift/imbalance in covariates (i.e., secondary non-imaging data). Controlling for such nuisance variables is common within standard statistical analysis, but the ideas do not directly apply to overparameterized models. Consequently, recent work has shown how strategies from invariant representation learning provides a meaningful starting point, but the current repertoire of methods is limited to accounting for shifts/imbalances in just a couple of covariates at a time. In this paper, we show how viewing this problem from the perspective of Category theory provides a simple and effective solution that completely avoids elaborate multi-stage training pipelines that would otherwise be needed. We show the effectiveness of this approach via extensive experiments on real datasets. Further, we discuss how this style of formulation offers a unified perspective on at least 5+ distinct problem settings, from self-supervised learning to matching problems in 3D reconstruction.