In the deep learning community, there are various other approaches that have been proposed for one-shot learning, such as generative modeling using GANs, image deformation meta-networks, representative based metric learning, and so on. So far, we have seen models doing classification using one-shot learning, but there are advancements currently being made in object detection and semantic segmentation as well. In this section, we will touch upon some of the recent papers from major machine learning-based conferences (for example, CVPR, NeurIPS, ICLR, and so on).
Metric-based learning is one of the older methods to approach one-shot learning. Though this area is old, there are plenty of aspects of it that are still being explored. The research work on the topic Revisiting local descriptor based image-to-class measure for few-shot learning (https://arxiv.org...
