Analyzing Decision Polygons of DNN-based Classification Methods

Jongyoung Kim, Seongyoun Woo, Wonjun Lee, Donghwan Kim, Chulhee Lee

2020

Abstract

Deep neural networks have shown impressive performance in various applications, including many pattern recognition problems. However, their working mechanisms have not been fully understood and adversarial examples indicate some fundamental problems with DNN-based classification methods. In this paper, we investigate the decision modeling mechanism of deep neural networks, which use the ReLU function. We derive some equations that show how each layer of deep neural networks expands the input dimension into higher dimensional spaces and generates numerous decision polygons. In this paper, we investigate the decision polygon formulations and present some examples that show interesting properties of DNN based classification methods.

Paper Citation

in Harvard Style

Kim J., Woo S., Lee W., Kim D. and Lee C. (2020). Analyzing Decision Polygons of DNN-based Classification Methods.In Proceedings of the 17th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO, ISBN 978-989-758-442-8, pages 346-351. DOI: 10.5220/0009888203460351

in Bibtex Style

@conference{icinco20,
author={Jongyoung Kim and Seongyoun Woo and Wonjun Lee and Donghwan Kim and Chulhee Lee},
title={Analyzing Decision Polygons of DNN-based Classification Methods},
booktitle={Proceedings of the 17th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,},
year={2020},
pages={346-351},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0009888203460351},
isbn={978-989-758-442-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,
TI - Analyzing Decision Polygons of DNN-based Classification Methods
SN - 978-989-758-442-8
AU - Kim J.
AU - Woo S.
AU - Lee W.
AU - Kim D.
AU - Lee C.
PY - 2020
SP - 346
EP - 351
DO - 10.5220/0009888203460351