Please use this identifier to cite or link to this item:
https://hdl.handle.net/2440/120659
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Carneiro, Gustavo | - |
dc.contributor.advisor | Suter, David | - |
dc.contributor.advisor | Sasdelli, Michele | - |
dc.contributor.author | Williams, Jerome Oskar | - |
dc.date.issued | 2019 | - |
dc.identifier.uri | http://hdl.handle.net/2440/120659 | - |
dc.description.abstract | Improving efficiency in deep learning models implies achieving a more accurate model for a given computational budget, or conversely a faster, leaner model without losing accuracy. In order to improve efficiency, we can use regularization to to improve generalization to the real world, and compression to improve speed. Due to the information-restricting nature of regularization, these two methods are related. Firstly we present a novel autoencoder architecture as a method of regularization for Pedestrian Detection. Secondly, we present a hyperparameter-free, iterative compression method based on measuring the information content of the model with the Information Bottleneck principle. | en |
dc.language.iso | en | en |
dc.subject | Machine learning | en |
dc.subject | neural network | en |
dc.subject | deep learning | en |
dc.subject | computer vision | en |
dc.subject | regularization | en |
dc.subject | compression | en |
dc.subject | information bottleneck | en |
dc.subject | autoencoder | en |
dc.subject | pedestrian detection | en |
dc.subject | region of interest | en |
dc.subject | convolutional | en |
dc.subject | statistics | en |
dc.subject | efficiency | en |
dc.title | Efficient Deep Learning Models with Autoencoder Regularization and Information Bottleneck Compression | en |
dc.type | Thesis | en |
dc.contributor.school | School of Computer Science | en |
dc.provenance | This electronic version is made publicly available by the University of Adelaide in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. This thesis may incorporate third party material which has been used by the author pursuant to Fair Dealing exceptions. If you are the owner of any included third party copyright material you wish to be removed from this electronic version, please complete the take down form located at: http://www.adelaide.edu.au/legals | en |
dc.description.dissertation | Thesis (MPhil) -- University of Adelaide, School of Computer Science, 2019 | en |
Appears in Collections: | Research Theses |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Williams2019_MPhil.pdf | Thesis | 1.98 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.