Improving Siamese Networks for One Shot Learning using Kernel Based Activation functions

TL;DR

This paper enhances Siamese networks for one-shot learning by integrating kernel-based activation functions, leading to improved accuracy and faster convergence with fewer epochs.

Contribution

It introduces the use of kernel activation functions (Kafnets) in Siamese networks to improve embedding quality and learning efficiency in one-shot learning tasks.

Findings

Kernel activation functions outperform ReLU in embedding quality.

Faster convergence with fewer training epochs.

Higher accuracy in one-shot learning benchmarks.

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. This process of learning in deep learning is usually accomplished by proper objective function, i.e; loss function and embeddings extraction i.e; architecture. In this paper, we discussed about metrics based deep learning architectures for one shot learning such as Siamese neural networks and present a method to improve on their accuracy using Kafnets (kernel-based non-parametric activation functions for neural networks) by learning proper embeddings with relatively less number of epochs. Using kernel activation functions, we are able to achieve strong results which exceed those of ReLU based deep learning models in terms of embeddings structure, loss convergence, and accuracy.