f-VAEGAN-D2: A Feature Generating Framework for Any-Shot Learning
This work addresses the problem of scarce labeled data in computer vision by enabling feature generation for any-shot learning, which is incremental as it builds on existing VAE and GAN methods.
The paper tackles any-shot learning (zero-shot and few-shot) by developing a feature generating framework that uses a conditional generative model combining VAE and GANs, with an unconditional discriminator to leverage unlabeled data, achieving state-of-the-art results on five datasets including CUB, SUN, AWA, and ImageNet.
When labeled training data is scarce, a promising data augmentation approach is to generate visual features of unknown classes using their attributes. To learn the class conditional distribution of CNN features, these models rely on pairs of image features and class attributes. Hence, they can not make use of the abundance of unlabeled data samples. In this paper, we tackle any-shot learning problems i.e. zero-shot and few-shot, in a unified feature generating framework that operates in both inductive and transductive learning settings. We develop a conditional generative model that combines the strength of VAE and GANs and in addition, via an unconditional discriminator, learns the marginal feature distribution of unlabeled images. We empirically show that our model learns highly discriminative CNN features for five datasets, i.e. CUB, SUN, AWA and ImageNet, and establish a new state-of-the-art in any-shot learning, i.e. inductive and transductive (generalized) zero- and few-shot learning settings. We also demonstrate that our learned features are interpretable: we visualize them by inverting them back to the pixel space and we explain them by generating textual arguments of why they are associated with a certain label.