A Self-supervised GAN for Unsupervised Few-shot Object Recognition
It addresses the problem of recognizing objects with limited labeled data for machine learning applications, representing an incremental advancement in few-shot learning.
This paper tackles unsupervised few-shot object recognition by extending a vanilla GAN with self-supervised loss functions, achieving significant performance improvements over state-of-the-art methods on Mini-Imagenet and Tiered-Imagenet datasets.
This paper addresses unsupervised few-shot object recognition, where all training images are unlabeled, and test images are divided into queries and a few labeled support images per object class of interest. The training and test images do not share object classes. We extend the vanilla GAN with two loss functions, both aimed at self-supervised learning. The first is a reconstruction loss that enforces the discriminator to reconstruct the probabilistically sampled latent code which has been used for generating the "fake" image. The second is a triplet loss that enforces the discriminator to output image encodings that are closer for more similar images. Evaluation, comparisons, and detailed ablation studies are done in the context of few-shot classification. Our approach significantly outperforms the state of the art on the Mini-Imagenet and Tiered-Imagenet datasets.