PRIME: Deep Imbalanced Regression with Proxies

Jongin Lim, Sucheol Lee, Daeho Um, Sung-Un Park, Jinwoo Shin
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:37705-37729, 2025.

Abstract

Data imbalance remains a fundamental challenge in real-world machine learning. However, most existing work has focused on classification, leaving imbalanced regression underexplored despite its importance in many applications. To address this gap, we propose PRIME, a framework that leverages learnable proxies to construct a balanced and well-ordered feature space for imbalanced regression. At its core, PRIME arranges proxies to be uniformly distributed in the feature space while preserving the ordinal structure of regression targets, and then aligns each sample feature to its corresponding proxy. By using proxies as reference points, PRIME induces the desired structure of learned representations, promoting better generalization, especially in underrepresented target regions. Moreover, since proxy-based alignment resembles classification, PRIME enables the seamless application of class imbalance techniques to regression, facilitating more balanced feature learning. Extensive experiments demonstrate the effectiveness and broad applicability of PRIME, achieving state-of-the-art performance on four real-world regression benchmark datasets across diverse target domains.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-lim25a, title = {{PRIME}: Deep Imbalanced Regression with Proxies}, author = {Lim, Jongin and Lee, Sucheol and Um, Daeho and Park, Sung-Un and Shin, Jinwoo}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {37705--37729}, 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/lim25a/lim25a.pdf}, url = {https://proceedings.mlr.press/v267/lim25a.html}, abstract = {Data imbalance remains a fundamental challenge in real-world machine learning. However, most existing work has focused on classification, leaving imbalanced regression underexplored despite its importance in many applications. To address this gap, we propose PRIME, a framework that leverages learnable proxies to construct a balanced and well-ordered feature space for imbalanced regression. At its core, PRIME arranges proxies to be uniformly distributed in the feature space while preserving the ordinal structure of regression targets, and then aligns each sample feature to its corresponding proxy. By using proxies as reference points, PRIME induces the desired structure of learned representations, promoting better generalization, especially in underrepresented target regions. Moreover, since proxy-based alignment resembles classification, PRIME enables the seamless application of class imbalance techniques to regression, facilitating more balanced feature learning. Extensive experiments demonstrate the effectiveness and broad applicability of PRIME, achieving state-of-the-art performance on four real-world regression benchmark datasets across diverse target domains.} }
Endnote
%0 Conference Paper %T PRIME: Deep Imbalanced Regression with Proxies %A Jongin Lim %A Sucheol Lee %A Daeho Um %A Sung-Un Park %A Jinwoo Shin %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-lim25a %I PMLR %P 37705--37729 %U https://proceedings.mlr.press/v267/lim25a.html %V 267 %X Data imbalance remains a fundamental challenge in real-world machine learning. However, most existing work has focused on classification, leaving imbalanced regression underexplored despite its importance in many applications. To address this gap, we propose PRIME, a framework that leverages learnable proxies to construct a balanced and well-ordered feature space for imbalanced regression. At its core, PRIME arranges proxies to be uniformly distributed in the feature space while preserving the ordinal structure of regression targets, and then aligns each sample feature to its corresponding proxy. By using proxies as reference points, PRIME induces the desired structure of learned representations, promoting better generalization, especially in underrepresented target regions. Moreover, since proxy-based alignment resembles classification, PRIME enables the seamless application of class imbalance techniques to regression, facilitating more balanced feature learning. Extensive experiments demonstrate the effectiveness and broad applicability of PRIME, achieving state-of-the-art performance on four real-world regression benchmark datasets across diverse target domains.
APA
Lim, J., Lee, S., Um, D., Park, S. & Shin, J.. (2025). PRIME: Deep Imbalanced Regression with Proxies. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:37705-37729 Available from https://proceedings.mlr.press/v267/lim25a.html.

Related Material