Learning Loss Functions for Semi-supervised Learning via Discriminative Adversarial Networks
This work addresses the challenge of improving predictor performance with limited labeled data in semi-supervised learning, though it appears incremental as it builds upon existing GAN frameworks.
The paper tackled the problem of semi-supervised learning and loss function learning by proposing discriminative adversarial networks (DAN), which use two discriminators instead of a generator and discriminator. The results show that DAN significantly boosts predictor performance for small labeled sets and outperforms standard loss functions like pairwise and negative log-likelihood in tasks such as classification and ranking.
We propose discriminative adversarial networks (DAN) for semi-supervised learning and loss function learning. Our DAN approach builds upon generative adversarial networks (GANs) and conditional GANs but includes the key differentiator of using two discriminators instead of a generator and a discriminator. DAN can be seen as a framework to learn loss functions for predictors that also implements semi-supervised learning in a straightforward manner. We propose instantiations of DAN for two different prediction tasks: classification and ranking. Our experimental results on three datasets of different tasks demonstrate that DAN is a promising framework for both semi-supervised learning and learning loss functions for predictors. For all tasks, the semi-supervised capability of DAN can significantly boost the predictor performance for small labeled sets with minor architecture changes across tasks. Moreover, the loss functions automatically learned by DANs are very competitive and usually outperform the standard pairwise and negative log-likelihood loss functions for both semi-supervised and supervised learning.