Is Graph Mixup Beneficial? Investigating Interpolation And Empirical Performance of Graph Mixup Methods

Simon Forbat, Rainer Gemulla
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:31377-31410, 2026.

Abstract

Mixup is a widely used data augmentation technique that constructs new training examples by interpolating between existing ones. While simple and effective in domains like vision and language, applying mixup to graph data is non-trivial and there is no independent empirical evidence for its effectiveness. To fill this gap, we conducted an extensive evaluation study following a unified, established evaluation protocol for graph classification. In contrast to prior results, we found that none of the state-of-the-art mixup methods yielded statistically significant improvements over the no-mixup baseline. To obtain further insights, we analyzed the graphs generated from these mixup methods from an interpolation perspective. We found that (i) many mixup methods failed to interpolate well, (ii) high interpolation error led to performance degradation, and (iii) even good interpolation properties did not lead to performance improvements. Our findings question the efficacy of existing graph mixup methods and highlight the need for a more rigorous exploration and evaluation.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-forbat26a, title = {Is Graph Mixup Beneficial? {I}nvestigating Interpolation And Empirical Performance of Graph Mixup Methods}, author = {Forbat, Simon and Gemulla, Rainer}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {31377--31410}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/forbat26a/forbat26a.pdf}, url = {https://proceedings.mlr.press/v306/forbat26a.html}, abstract = {Mixup is a widely used data augmentation technique that constructs new training examples by interpolating between existing ones. While simple and effective in domains like vision and language, applying mixup to graph data is non-trivial and there is no independent empirical evidence for its effectiveness. To fill this gap, we conducted an extensive evaluation study following a unified, established evaluation protocol for graph classification. In contrast to prior results, we found that none of the state-of-the-art mixup methods yielded statistically significant improvements over the no-mixup baseline. To obtain further insights, we analyzed the graphs generated from these mixup methods from an interpolation perspective. We found that (i) many mixup methods failed to interpolate well, (ii) high interpolation error led to performance degradation, and (iii) even good interpolation properties did not lead to performance improvements. Our findings question the efficacy of existing graph mixup methods and highlight the need for a more rigorous exploration and evaluation.} }
Endnote
%0 Conference Paper %T Is Graph Mixup Beneficial? Investigating Interpolation And Empirical Performance of Graph Mixup Methods %A Simon Forbat %A Rainer Gemulla %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-forbat26a %I PMLR %P 31377--31410 %U https://proceedings.mlr.press/v306/forbat26a.html %V 306 %X Mixup is a widely used data augmentation technique that constructs new training examples by interpolating between existing ones. While simple and effective in domains like vision and language, applying mixup to graph data is non-trivial and there is no independent empirical evidence for its effectiveness. To fill this gap, we conducted an extensive evaluation study following a unified, established evaluation protocol for graph classification. In contrast to prior results, we found that none of the state-of-the-art mixup methods yielded statistically significant improvements over the no-mixup baseline. To obtain further insights, we analyzed the graphs generated from these mixup methods from an interpolation perspective. We found that (i) many mixup methods failed to interpolate well, (ii) high interpolation error led to performance degradation, and (iii) even good interpolation properties did not lead to performance improvements. Our findings question the efficacy of existing graph mixup methods and highlight the need for a more rigorous exploration and evaluation.
APA
Forbat, S. & Gemulla, R.. (2026). Is Graph Mixup Beneficial? Investigating Interpolation And Empirical Performance of Graph Mixup Methods. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:31377-31410 Available from https://proceedings.mlr.press/v306/forbat26a.html.

Related Material