Date of Award

Spring 3-3-2019

Document Type

Capstone

Degree Name

Bachelor of Science

Department

Computer Science & Mathematics; College of Arts & Sciences

First Advisor

Carlos Ortiz

Second Advisor

Vitaly Ford

Abstract

This paper has two objectives. 1. It simplifies the Mathematics behind a simple Neural Network. Furthermore, it explores how Neural Networks can be modeled using Ordinary Differential Equations(ODE). 2. It implements a simple example of an ODE Neural network using diffeqflux.jl library. My paper is based on the paper "Neural Ordinary Differential equations"[1] paper and contains multiple extracts from this paper and hence the work in chapter 4 should not be considered original work as it aims to explain the mathematics in the original paper and all credit is due to the authors of the paper [1]. This paper[1] was among on the 5 papers to be recognized at the 2018 annual conference NeurIPS(Neural Information Processing Systems)

Download

Share

COinS