Semi-Ensemble: A Simple Approach Over-parameterize Model Interpolation

Jiwoon Lee, Jaeho Lee
Proceedings of UniReps: the First Workshop on Unifying Representations in Neural Models, PMLR 243:182-193, 2024.

Abstract

We develop a unified framework for interpolating two models with various degrees of over-parameterization, having model merging and model ensemble as special cases. Instead of directly interpolating models in their original parameter space, the proposed Semi-Ensemble interpolates the over-parameterized versions of the models in a higher-dimensional joint parameter space. Here, the over-parameterizations recover each endpoint model when projected to some low-dimensional subspace spanned by a fraction of bases. By carefully constructing the joint parameter space, the interpolated model can achieve a smooth tradeoff between the total number of parameters and the model accuracy, outperforming existing baselines. Intriguingly, we show that Semi-ensembles can sometimes achieve a better performance than vanilla ensembles, even with a slightly smaller number of parameters.

Cite this Paper

BibTeX
@InProceedings{pmlr-v243-lee24a, title = {Semi-Ensemble: A Simple Approach Over-parameterize Model Interpolation}, author = {Lee, Jiwoon and Lee, Jaeho}, booktitle = {Proceedings of UniReps: the First Workshop on Unifying Representations in Neural Models}, pages = {182--193}, year = {2024}, editor = {Fumero, Marco and Rodolá, Emanuele and Domine, Clementine and Locatello, Francesco and Dziugaite, Karolina and Mathilde, Caron}, volume = {243}, series = {Proceedings of Machine Learning Research}, month = {15 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v243/lee24a/lee24a.pdf}, url = {https://proceedings.mlr.press/v243/lee24a.html}, abstract = {We develop a unified framework for interpolating two models with various degrees of over-parameterization, having model merging and model ensemble as special cases. Instead of directly interpolating models in their original parameter space, the proposed Semi-Ensemble interpolates the over-parameterized versions of the models in a higher-dimensional joint parameter space. Here, the over-parameterizations recover each endpoint model when projected to some low-dimensional subspace spanned by a fraction of bases. By carefully constructing the joint parameter space, the interpolated model can achieve a smooth tradeoff between the total number of parameters and the model accuracy, outperforming existing baselines. Intriguingly, we show that Semi-ensembles can sometimes achieve a better performance than vanilla ensembles, even with a slightly smaller number of parameters.} }
Endnote
%0 Conference Paper %T Semi-Ensemble: A Simple Approach Over-parameterize Model Interpolation %A Jiwoon Lee %A Jaeho Lee %B Proceedings of UniReps: the First Workshop on Unifying Representations in Neural Models %C Proceedings of Machine Learning Research %D 2024 %E Marco Fumero %E Emanuele Rodolá %E Clementine Domine %E Francesco Locatello %E Karolina Dziugaite %E Caron Mathilde %F pmlr-v243-lee24a %I PMLR %P 182--193 %U https://proceedings.mlr.press/v243/lee24a.html %V 243 %X We develop a unified framework for interpolating two models with various degrees of over-parameterization, having model merging and model ensemble as special cases. Instead of directly interpolating models in their original parameter space, the proposed Semi-Ensemble interpolates the over-parameterized versions of the models in a higher-dimensional joint parameter space. Here, the over-parameterizations recover each endpoint model when projected to some low-dimensional subspace spanned by a fraction of bases. By carefully constructing the joint parameter space, the interpolated model can achieve a smooth tradeoff between the total number of parameters and the model accuracy, outperforming existing baselines. Intriguingly, we show that Semi-ensembles can sometimes achieve a better performance than vanilla ensembles, even with a slightly smaller number of parameters.
APA
Lee, J. & Lee, J.. (2024). Semi-Ensemble: A Simple Approach Over-parameterize Model Interpolation. Proceedings of UniReps: the First Workshop on Unifying Representations in Neural Models, in Proceedings of Machine Learning Research 243:182-193 Available from https://proceedings.mlr.press/v243/lee24a.html.

Related Material