HexaGAN: Generative Adversarial Nets for Real World Classification

Uiwon Hwang, Dahuin Jung, Sungroh Yoon
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:2921-2930, 2019.

Abstract

Most deep learning classification studies assume clean data. However, when dealing with the real world data, we encounter three problems such as 1) missing data, 2) class imbalance, and 3) missing label problems. These problems undermine the performance of a classifier. Various preprocessing techniques have been proposed to mitigate one of these problems, but an algorithm that assumes and resolves all three problems together has not been proposed yet. In this paper, we propose HexaGAN, a generative adversarial network framework that shows promising classification performance for all three problems. We interpret the three problems from a single perspective to solve them jointly. To enable this, the framework consists of six components, which interact with each other. We also devise novel loss functions corresponding to the architecture. The designed loss functions allow us to achieve state-of-the-art imputation performance, with up to a 14% improvement, and to generate high-quality class-conditional data. We evaluate the classification performance (F1-score) of the proposed method with 20% missingness and confirm up to a 5% improvement in comparison with the performance of combinations of state-of-the-art methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v97-hwang19a, title = {{H}exa{GAN}: Generative Adversarial Nets for Real World Classification}, author = {Hwang, Uiwon and Jung, Dahuin and Yoon, Sungroh}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {2921--2930}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/hwang19a/hwang19a.pdf}, url = {https://proceedings.mlr.press/v97/hwang19a.html}, abstract = {Most deep learning classification studies assume clean data. However, when dealing with the real world data, we encounter three problems such as 1) missing data, 2) class imbalance, and 3) missing label problems. These problems undermine the performance of a classifier. Various preprocessing techniques have been proposed to mitigate one of these problems, but an algorithm that assumes and resolves all three problems together has not been proposed yet. In this paper, we propose HexaGAN, a generative adversarial network framework that shows promising classification performance for all three problems. We interpret the three problems from a single perspective to solve them jointly. To enable this, the framework consists of six components, which interact with each other. We also devise novel loss functions corresponding to the architecture. The designed loss functions allow us to achieve state-of-the-art imputation performance, with up to a 14% improvement, and to generate high-quality class-conditional data. We evaluate the classification performance (F1-score) of the proposed method with 20% missingness and confirm up to a 5% improvement in comparison with the performance of combinations of state-of-the-art methods.} }
Endnote
%0 Conference Paper %T HexaGAN: Generative Adversarial Nets for Real World Classification %A Uiwon Hwang %A Dahuin Jung %A Sungroh Yoon %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-hwang19a %I PMLR %P 2921--2930 %U https://proceedings.mlr.press/v97/hwang19a.html %V 97 %X Most deep learning classification studies assume clean data. However, when dealing with the real world data, we encounter three problems such as 1) missing data, 2) class imbalance, and 3) missing label problems. These problems undermine the performance of a classifier. Various preprocessing techniques have been proposed to mitigate one of these problems, but an algorithm that assumes and resolves all three problems together has not been proposed yet. In this paper, we propose HexaGAN, a generative adversarial network framework that shows promising classification performance for all three problems. We interpret the three problems from a single perspective to solve them jointly. To enable this, the framework consists of six components, which interact with each other. We also devise novel loss functions corresponding to the architecture. The designed loss functions allow us to achieve state-of-the-art imputation performance, with up to a 14% improvement, and to generate high-quality class-conditional data. We evaluate the classification performance (F1-score) of the proposed method with 20% missingness and confirm up to a 5% improvement in comparison with the performance of combinations of state-of-the-art methods.
APA
Hwang, U., Jung, D. & Yoon, S.. (2019). HexaGAN: Generative Adversarial Nets for Real World Classification. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:2921-2930 Available from https://proceedings.mlr.press/v97/hwang19a.html.

Related Material