How to Fill the Optimum Set? Population Gradient Descent with Harmless Diversity

Chengyue Gong, Lemeng Wu, Qiang Liu
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:7650-7664, 2022.

Abstract

Although traditional optimization methods focus on finding a single optimal solution, most objective functions in modern machine learning problems, especially those in deep learning, often have multiple or infinite number of optimal points. Therefore, it is useful to consider the problem of finding a set of diverse points in the optimum set of an objective function. In this work, we frame this problem as a bi-level optimization problem of maximizing a diversity score inside the optimum set of the main loss function, and solve it with a simple population gradient descent framework that iteratively updates the points to maximize the diversity score in a fashion that does not hurt the optimization of the main loss. We demonstrate that our method can efficiently generate diverse solutions on multiple applications, e.g. text-to-image generation, text-to-mesh generation, molecular conformation generation and ensemble neural network training.

Cite this Paper


BibTeX
@InProceedings{pmlr-v162-gong22b, title = {How to Fill the Optimum Set? {P}opulation Gradient Descent with Harmless Diversity}, author = {Gong, Chengyue and Wu, Lemeng and Liu, Qiang}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {7650--7664}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/gong22b/gong22b.pdf}, url = {https://proceedings.mlr.press/v162/gong22b.html}, abstract = {Although traditional optimization methods focus on finding a single optimal solution, most objective functions in modern machine learning problems, especially those in deep learning, often have multiple or infinite number of optimal points. Therefore, it is useful to consider the problem of finding a set of diverse points in the optimum set of an objective function. In this work, we frame this problem as a bi-level optimization problem of maximizing a diversity score inside the optimum set of the main loss function, and solve it with a simple population gradient descent framework that iteratively updates the points to maximize the diversity score in a fashion that does not hurt the optimization of the main loss. We demonstrate that our method can efficiently generate diverse solutions on multiple applications, e.g. text-to-image generation, text-to-mesh generation, molecular conformation generation and ensemble neural network training.} }
Endnote
%0 Conference Paper %T How to Fill the Optimum Set? Population Gradient Descent with Harmless Diversity %A Chengyue Gong %A Lemeng Wu %A Qiang Liu %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-gong22b %I PMLR %P 7650--7664 %U https://proceedings.mlr.press/v162/gong22b.html %V 162 %X Although traditional optimization methods focus on finding a single optimal solution, most objective functions in modern machine learning problems, especially those in deep learning, often have multiple or infinite number of optimal points. Therefore, it is useful to consider the problem of finding a set of diverse points in the optimum set of an objective function. In this work, we frame this problem as a bi-level optimization problem of maximizing a diversity score inside the optimum set of the main loss function, and solve it with a simple population gradient descent framework that iteratively updates the points to maximize the diversity score in a fashion that does not hurt the optimization of the main loss. We demonstrate that our method can efficiently generate diverse solutions on multiple applications, e.g. text-to-image generation, text-to-mesh generation, molecular conformation generation and ensemble neural network training.
APA
Gong, C., Wu, L. & Liu, Q.. (2022). How to Fill the Optimum Set? Population Gradient Descent with Harmless Diversity. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:7650-7664 Available from https://proceedings.mlr.press/v162/gong22b.html.

Related Material