MAGANet: Achieving Combinatorial Generalization by Modeling a Group Action

Geonho Hwang, Jaewoong Choi, Hyunsoo Cho, Myungjoo Kang
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:14237-14248, 2023.

Abstract

Combinatorial generalization refers to the ability to collect and assemble various attributes from diverse data to generate novel unexperienced data. This ability is considered a necessary passing point for achieving human-level intelligence. To achieve this ability, previous unsupervised approaches mainly focused on learning the disentangled representation, such as the variational autoencoder. However, recent studies discovered that the disentangled representation is insufficient for combinatorial generalization and is not even correlated. In this regard, we propose a novel framework for data generation that can robustly generalize under these distribution shift situations. Instead of representing each data, our model discovers the fundamental transformation between a pair of data by simulating a group action. To test the combinatorial generalizability, we evaluated our model in two settings: Recombination-to-Element and Recombination-to-Range. The experiments demonstrated that our method has quantitatively and qualitatively superior generalizability and generates better images than traditional models.

Cite this Paper


BibTeX
@InProceedings{pmlr-v202-hwang23b, title = {{MAGAN}et: Achieving Combinatorial Generalization by Modeling a Group Action}, author = {Hwang, Geonho and Choi, Jaewoong and Cho, Hyunsoo and Kang, Myungjoo}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {14237--14248}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/hwang23b/hwang23b.pdf}, url = {https://proceedings.mlr.press/v202/hwang23b.html}, abstract = {Combinatorial generalization refers to the ability to collect and assemble various attributes from diverse data to generate novel unexperienced data. This ability is considered a necessary passing point for achieving human-level intelligence. To achieve this ability, previous unsupervised approaches mainly focused on learning the disentangled representation, such as the variational autoencoder. However, recent studies discovered that the disentangled representation is insufficient for combinatorial generalization and is not even correlated. In this regard, we propose a novel framework for data generation that can robustly generalize under these distribution shift situations. Instead of representing each data, our model discovers the fundamental transformation between a pair of data by simulating a group action. To test the combinatorial generalizability, we evaluated our model in two settings: Recombination-to-Element and Recombination-to-Range. The experiments demonstrated that our method has quantitatively and qualitatively superior generalizability and generates better images than traditional models.} }
Endnote
%0 Conference Paper %T MAGANet: Achieving Combinatorial Generalization by Modeling a Group Action %A Geonho Hwang %A Jaewoong Choi %A Hyunsoo Cho %A Myungjoo Kang %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-hwang23b %I PMLR %P 14237--14248 %U https://proceedings.mlr.press/v202/hwang23b.html %V 202 %X Combinatorial generalization refers to the ability to collect and assemble various attributes from diverse data to generate novel unexperienced data. This ability is considered a necessary passing point for achieving human-level intelligence. To achieve this ability, previous unsupervised approaches mainly focused on learning the disentangled representation, such as the variational autoencoder. However, recent studies discovered that the disentangled representation is insufficient for combinatorial generalization and is not even correlated. In this regard, we propose a novel framework for data generation that can robustly generalize under these distribution shift situations. Instead of representing each data, our model discovers the fundamental transformation between a pair of data by simulating a group action. To test the combinatorial generalizability, we evaluated our model in two settings: Recombination-to-Element and Recombination-to-Range. The experiments demonstrated that our method has quantitatively and qualitatively superior generalizability and generates better images than traditional models.
APA
Hwang, G., Choi, J., Cho, H. & Kang, M.. (2023). MAGANet: Achieving Combinatorial Generalization by Modeling a Group Action. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:14237-14248 Available from https://proceedings.mlr.press/v202/hwang23b.html.

Related Material