We are upgrading the repository! A content freeze is in effect until November 22nd, 2024 - no new submissions will be accepted; however, all content already published will remain publicly available. Please reach out to repository@u.library.arizona.edu with your questions, or if you are a UA affiliate who needs to make content available soon. Note that any new user accounts created after September 22, 2024 will need to be recreated by the user in November after our migration is completed.

Show simple item record

dc.contributor.advisorAkoglu, Ali
dc.contributor.authorMixter, John Edward
dc.creatorMixter, John Edward
dc.date.accessioned2024-01-27T18:59:39Z
dc.date.available2024-01-27T18:59:39Z
dc.date.issued2023
dc.identifier.citationMixter, John Edward. (2023). Neural Network Reduction for Efficient Execution on Edge Devices (Doctoral dissertation, University of Arizona, Tucson, USA).
dc.identifier.urihttp://hdl.handle.net/10150/670868
dc.description.abstractAs the size of neural networks increase, the resources needed to support their execution also increase. This presents a barrier for creating neural networks that can be trained and executed within resource limited embedded systems. To reduce the resources needed to execute neural networks, weight reduction is often the first target. A network that has been significantly pruned can be executed on-chip, that is, in low SWaP hardware. But, this does not enable either training or pruning in embedded hardware which first requires a full-sized network to fit within the restricted resources. We introduce two methods of network reduction that allows neural networks to be grown and trained within edge devices, Artificial Neurogenesis and Synaptic Input Consolidation.
dc.language.isoen
dc.publisherThe University of Arizona.
dc.rightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
dc.rights.urihttp://rightsstatements.org/vocab/InC/1.0/
dc.subjectMachine Learning
dc.subjectNeural Network Growth
dc.subjectNeural Network Reduction
dc.subjectNeural Networks
dc.titleNeural Network Reduction for Efficient Execution on Edge Devices
dc.typeElectronic Dissertation
dc.typetext
thesis.degree.grantorUniversity of Arizona
thesis.degree.leveldoctoral
dc.contributor.committeememberHariri, Salim
dc.contributor.committeememberTandon, Ravi
thesis.degree.disciplineGraduate College
thesis.degree.disciplineElectrical & Computer Engineering
thesis.degree.namePh.D.
refterms.dateFOA2024-01-27T18:59:39Z


Files in this item

Thumbnail
Name:
azu_etd_21112_sip1_m.pdf
Size:
1.318Mb
Format:
PDF

This item appears in the following Collection(s)

Show simple item record