Improving Model Compatibility of Generative Adversarial Networks by Boundary Calibration
This addresses the problem of generating synthetic data that better trains classifiers for researchers and practitioners in machine learning, representing an incremental improvement focused on a specific domain.
The paper tackles the model compatibility problem in Generative Adversarial Networks (GANs), where existing GANs often generate 'easier' synthetic data far from classifier boundaries, by proposing Boundary-Calibration GANs (BCGANs) that incorporate a boundary-calibration loss to match posterior distributions, resulting in improved model compatibility while maintaining realistic image generation.
Generative Adversarial Networks (GANs) is a powerful family of models that learn an underlying distribution to generate synthetic data. Many existing studies of GANs focus on improving the realness of the generated image data for visual applications, and few of them concern about improving the quality of the generated data for training other classifiers -- a task known as the model compatibility problem. As a consequence, existing GANs often prefer generating `easier' synthetic data that are far from the boundaries of the classifiers, and refrain from generating near-boundary data, which are known to play an important roles in training the classifiers. To improve GAN in terms of model compatibility, we propose Boundary-Calibration GANs (BCGANs), which leverage the boundary information from a set of pre-trained classifiers using the original data. In particular, we introduce an auxiliary Boundary-Calibration loss (BC-loss) into the generator of GAN to match the statistics between the posterior distributions of original data and generated data with respect to the boundaries of the pre-trained classifiers. The BC-loss is provably unbiased and can be easily coupled with different GAN variants to improve their model compatibility. Experimental results demonstrate that BCGANs not only generate realistic images like original GANs but also achieves superior model compatibility than the original GANs.