Study of Constrained Network Structures for WGANs on Numeric Data Generation
This addresses data augmentation for imbalanced numeric datasets, but it is incremental as it builds on existing WGAN methods with structural modifications.
The paper tackled the overfitting and ill-conditioning problems of conventional GANs in generating numeric data by studying constrained network structures (isomorphic, mirror, self-symmetric) in WGANs, resulting in improved performance in 17 out of 20 experiments, with Isomorphic WGAN being best in 15 out of 20 cases.
Some recent studies have suggested using GANs for numeric data generation such as to generate data for completing the imbalanced numeric data. Considering the significant difference between the dimensions of the numeric data and images, as well as the strong correlations between features of numeric data, the conventional GANs normally face an overfitting problem, consequently leads to an ill-conditioning problem in generating numeric and structured data. This paper studies the constrained network structures between generator G and discriminator D in WGAN, designs several structures including isomorphic, mirror and self-symmetric structures. We evaluates the performances of the constrained WGANs in data augmentations, taking the non-constrained GANs and WGANs as the baselines. Experiments prove the constrained structures have been improved in 17/20 groups of experiments. In twenty experiments on four UCI Machine Learning Repository datasets, Australian Credit Approval data, German Credit data, Pima Indians Diabetes data and SPECT heart data facing five conventional classifiers. Especially, Isomorphic WGAN is the best in 15/20 experiments. Finally, we theoretically proves that the effectiveness of constrained structures by the directed graphic model (DGM) analysis.