https://towardsdatascience.com/advances-in-few-shot-learning-a-guided-tour-36bc10a68b77

great brief summary over matching networks, prototypical networks, model-agnostic meta-learning(MAML). well, the first two topics are well explained but MAML section needs a lot more thinking to understand. Also, I think MAML is closer to the topic of meta-learning rather than few shot learning.


https://arxiv.org/pdf/2008.06365.pdf

“An Overview of Deep Learning Architectures inFew-Shot Learning Domain” paper, released on 2020.

It lists up few shot learning methods in a more broader sense, meaning it mentions data augmentation, transfer learning, and other techniques as a way of few shot learning. It does dig in deeper on the “narrow definition” of few shot learning methods as well. Nice to check on various methods of few shot learning but the descriptions of each method are quite hard to understand.

surprisingly, this paper doesn’t mention prototypical networks.


https://towardsdatascience.com/few-shot-learning-with-prototypical-networks-87949de03ccd

tutorial on building a prototypical network with code examples.


0 Comments

Leave a Reply

Your email address will not be published.