Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/77810
Title: Improving the robustness of machine learning system through convolutional neural network
Authors: Chia, Daryl Jing
Keywords: DRNTU::Engineering::Electrical and electronic engineering
Issue Date: 2019
Abstract: This report is to study what is a Convolutional Neural Network and carry out a multi-layer network surgery on the AlexNet Convolutional Neural Network. The AlexNet network consists of five convolutional layers and three fully connected layers. Noise error layers are then created to fit into the various layers. These error layers with number of noise element being roughly 10% of total elements per layer are then injected into the network to see how the output will be affected. This is to find out how robust each layer towards error. This report will cover the methods and steps that I took to carry out the surgery, create the error layers of the correct size and injecting of the errors into the network. I will break down the codes that I used to achieve the above chunk by chunk in the Methodology Chapter. Thereafter, I will be discussing the results and findings of each testing in each individual layer of the network. The report will end with a conclusion and recommended future works.
URI: http://hdl.handle.net/10356/77810
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
FYP Final Report (Chia Jing Daryl).pdf
  Restricted Access
5.76 MBAdobe PDFView/Open

Page view(s)

145
Updated on Jun 27, 2022

Download(s)

6
Updated on Jun 27, 2022

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.