Collapse by Conditioning: Training Class-conditional GANs with Limited Data

Class-conditioning offers a direct means of controlling a Generative Adversarial Network (GAN) based on a discrete input variable. While necessary in many applications, the additional information provided by the class labels could even be expected to benefit the training of the GAN itself. Contrary to this belief, we observe that class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generative ability. Motivated by this observation, we propose a training strategy for conditional GANs (cGANs) that effectively prevents the observed mode-collapse by leveraging unconditional learning. Our training strategy starts with an unconditional GAN and gradually injects conditional information into the generator and the objective function. The proposed method for training cGANs with limited data results not only in stable training but also in generating high-quality images, thanks to the early-stage exploitation of the shared information across classes. We analyze the aforementioned mode collapse problem in comprehensive experiments on four datasets. Our approach demonstrates outstanding results compared with state-of-the-art methods and established baselines. Since the introduction of generative adversarial networks (GANs) by Goodfellow et al. (2014), there has been a substantial progress in realistic image and video generations. The contents of such generations are often controlled by conditioning the process by means of conditional GANs (cGANS) (Mirza & Osindero, 2014). In practice, cGANs are of high interest, as they can generate and control a wide variety of outputs using a single model.