Entropic Projection Alignment: Estimating, Explaining, and Improving Model Performance Under Distribution Shift

Salim I. Amoukou, Emanuele Albini, Tom Bewley, Saumitra Mishra, Manuela Veloso
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:3079-3087, 2026.

Abstract

We propose a unified framework for addressing three key challenges of distribution shift: (1) estimating a model’s performance on an unlabeled target domain, (2) explaining the shift by identifying the features responsible, and (3) improving the target domain performance. Our method, Entropic Projection Alignment (EPA), aligns the source distribution to the target by matching carefully selected moments while simultaneously minimising the KL divergence from the source. This formulation yields a unique closed-form solution for importance weights, achieving robustness through implicit variance control. Drawing on domain adaptation theory, we establish that moment matching is sufficient for reliable estimation and adaptation, avoiding the need for full density ratio recovery. Extensive experiments, together with strong theoretical guarantees, demonstrate that EPA consistently outperforms state-of-the-art baselines while offering substantial computational efficiency.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-amoukou26a, title = { Entropic Projection Alignment: Estimating, Explaining, and Improving Model Performance Under Distribution Shift }, author = {Amoukou, Salim I. and Albini, Emanuele and Bewley, Tom and Mishra, Saumitra and Veloso, Manuela}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {3079--3087}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/amoukou26a/amoukou26a.pdf}, url = {https://proceedings.mlr.press/v300/amoukou26a.html}, abstract = { We propose a unified framework for addressing three key challenges of distribution shift: (1) estimating a model’s performance on an unlabeled target domain, (2) explaining the shift by identifying the features responsible, and (3) improving the target domain performance. Our method, Entropic Projection Alignment (EPA), aligns the source distribution to the target by matching carefully selected moments while simultaneously minimising the KL divergence from the source. This formulation yields a unique closed-form solution for importance weights, achieving robustness through implicit variance control. Drawing on domain adaptation theory, we establish that moment matching is sufficient for reliable estimation and adaptation, avoiding the need for full density ratio recovery. Extensive experiments, together with strong theoretical guarantees, demonstrate that EPA consistently outperforms state-of-the-art baselines while offering substantial computational efficiency. } }
Endnote
%0 Conference Paper %T Entropic Projection Alignment: Estimating, Explaining, and Improving Model Performance Under Distribution Shift %A Salim I. Amoukou %A Emanuele Albini %A Tom Bewley %A Saumitra Mishra %A Manuela Veloso %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-amoukou26a %I PMLR %P 3079--3087 %U https://proceedings.mlr.press/v300/amoukou26a.html %V 300 %X We propose a unified framework for addressing three key challenges of distribution shift: (1) estimating a model’s performance on an unlabeled target domain, (2) explaining the shift by identifying the features responsible, and (3) improving the target domain performance. Our method, Entropic Projection Alignment (EPA), aligns the source distribution to the target by matching carefully selected moments while simultaneously minimising the KL divergence from the source. This formulation yields a unique closed-form solution for importance weights, achieving robustness through implicit variance control. Drawing on domain adaptation theory, we establish that moment matching is sufficient for reliable estimation and adaptation, avoiding the need for full density ratio recovery. Extensive experiments, together with strong theoretical guarantees, demonstrate that EPA consistently outperforms state-of-the-art baselines while offering substantial computational efficiency.
APA
Amoukou, S.I., Albini, E., Bewley, T., Mishra, S. & Veloso, M.. (2026). Entropic Projection Alignment: Estimating, Explaining, and Improving Model Performance Under Distribution Shift . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:3079-3087 Available from https://proceedings.mlr.press/v300/amoukou26a.html.

Related Material