Siamese Networks for One Shot Learning using Kernel Based Activation functions

EasyChair Preprint 893, version 1

Abstract

The lack of a large amount of training data has always been the constraining factor in solving a lot of problems in machine learning, making One Shot Learning one of the most intriguing ideas in machine learning. It aims to learn information about object categories from one, or only a few, training examples, and for certain image classification tasks, has successfully been able to get results comparable to human beings. In this paper, we have experimented with an architecture of One Shot Learning using Siamese neural networks and improve on their performance using Kafnets (kernel-based non-parametric activation functions for neural networks). We have experimented with Kafnets and showed how decision boundaries improve, and how well it converges with less number of layers and epochs.

Keyphrases: Kernels, computer vision, decision boundary, machine learning, one-shot learning

@booklet{EasyChair:893,
  author    = {Shruti Jadon and Aditya Acrot Srinivasan},
  title     = {Siamese Networks for One Shot Learning using Kernel Based Activation functions},
  howpublished = {EasyChair Preprint 893},
  year      = {EasyChair, 2019}}