Enhancing Siamese Networks Training with Importance Sampling

Ajay Shrestha, Ausif Mahmood

2019

Abstract

The accuracy of machine learning (ML) model is determined to a great extent by its training dataset. Yet the dataset optimization is often not the center of the focus to improve ML models. Datasets used in the training process can have a huge impact on the convergence of the training process and accuracy of the models. In this paper, we propose and implement importance sampling, a Monte Carlo method for variance reduction on training siamese networks to improve the accuracy of the image recognition. We demonstrate empirically that our approach can achieve improvement in training and testing errors on MNIST dataset compared to training when importance sampling is not used. Unlike standard convolution neural networks (CNN), siamese networks scale efficiently when the number of classes for image recognition increases. This paper is the first known attempt to combine importance sampling with siamese network and shows its effectiveness towards getting better accuracy.

Paper Citation

in Harvard Style

Shrestha A. and Mahmood A. (2019). Enhancing Siamese Networks Training with Importance Sampling.In Proceedings of the 11th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART, ISBN 978-989-758-350-6, pages 610-615. DOI: 10.5220/0007371706100615

in Bibtex Style

@conference{icaart19,
author={Ajay Shrestha and Ausif Mahmood},
title={Enhancing Siamese Networks Training with Importance Sampling},
booktitle={Proceedings of the 11th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART,},
year={2019},
pages={610-615},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0007371706100615},
isbn={978-989-758-350-6},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 11th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART,
TI - Enhancing Siamese Networks Training with Importance Sampling
SN - 978-989-758-350-6
AU - Shrestha A.
AU - Mahmood A.
PY - 2019
SP - 610
EP - 615
DO - 10.5220/0007371706100615