Metric Learning-based Generative Adversarial Network

Abstract

Generative Adversarial Networks (GANs), as a framework for estimating generative models via an adversarial process, have attracted huge attention and have proven to be powerful in a variety of tasks. However, training GANs is well known for being delicate and unstable, partially caused by its sig- moid cross entropy loss function for the discriminator. To overcome such a problem, many researchers directed their attention on various ways to measure how close the model distribution and real distribution are and have applied dif- ferent metrics as their objective functions. In this paper, we propose a novel framework to train GANs based on distance metric learning and we call it Metric Learning-based Gener- ative Adversarial Network (MLGAN). The discriminator of MLGANs can dynamically learn an appropriate metric, rather than a static one, to measure the distance between generated samples and real samples. Afterwards, MLGANs update the generator under the newly learned metric. We evaluate our ap- proach on several representative datasets and the experimen- tal results demonstrate that MLGANs can achieve superior performance compared with several existing state-of-the-art approaches. We also empirically show that MLGANs could increase the stability of training GANs.

References in corpus (11)

• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Improved Training of Wasserstein GANs
• Improved Techniques for Training GANs
• Energy-based Generative Adversarial Network
• A Survey on Metric Learning for Feature Vectors and Structured Data
• Towards Principled Methods for Training Generative Adversarial Networks
• The Cramer Distance as a Solution to Biased Wasserstein Gradients
• Least Squares Generative Adversarial Networks
• Loss-Sensitive Generative Adversarial Networks on Lipschitz Densities
• Training Triplet Networks with GAN

Cited by in corpus (1)

• Label-Removed Generative Adversarial Networks Incorporating with K-Means