One-shot learning
Date:
In this talk I have presented about one-shot learning - metric learning, siamese networks - and we had a hands-on implementation of siamese network training.
Human can learn even from a single instance, but computers can not. The existing model architectures suggest that computer are really very good in learning from high dimensional and huge amount of data. However, for every problem, we can neither have a high dimensional data nor the huge amount of data. Hence, how can we teach computers to learn from the less data, is still an open research problem. Some work in this direction is published in past few years. Researchers named this “One-shot learning”.