{"id":{"repo_id":"gatech","oai_identifier":"oai:repository.gatech.edu:1853/67193"},"canonical_url":"https://search.dev.ndltd.org/etd/gatech/oai:repository.gatech.edu:1853/67193","repository":{"repo_id":"gatech","name":"Georgia Tech","base_url":"https://repository.gatech.edu/server/oai/request"},"display":{"title":"Energy-aware DNN Quantization for Processing-In-Memory Architecture","abstract":"With increasing computational cost of deep neural network (DNN), many efforts to develop energy-efficient intelligent system have been proposed from dedicated hardware platforms to model compression algorithms. Recently, hardware-aware quantization algorithms have shown further improvement in the energy efficiency of DNN by considering hardware architectures and algorithms together. In this work, a genetic algorithm-based energy-aware DNN quantization framework for Processing-In-Memory (PIM) architectures, named EGQ, is presented. The key contribution of the research is to design a fitness function that can reduce the number of analog-to-digital converter (ADC) access, which is one of the main energy overhead in PIM. EGQ automatically optimizes layer-wise weight and activation bitwidth with negligible accuracy loss while considering the dynamic energy in PIM. The research demonstrates the effectiveness of EGQ on several DNN models VGG-19, ResNet-18, ResNet-50, MobileNet-V2, and SqueezeNet. Also, the area, dynamic energy, and energy efficiency in the compressed models with various memory technologies are analyzed. EGQ shows 15%-103% higher energy efficiency with 2% accuracy loss than other PIM-aware quantization algorithms.","abstract_html":"With increasing computational cost of deep neural network (DNN), many efforts to develop energy-efficient intelligent system have been proposed from dedicated hardware platforms to model compression algorithms. Recently, hardware-aware quantization algorithms have shown further improvement in the energy efficiency of DNN by considering hardware architectures and algorithms together. In this work, a genetic algorithm-based energy-aware DNN quantization framework for Processing-In-Memory (PIM) architectures, named EGQ, is presented. The key contribution of the research is to design a fitness function that can reduce the number of analog-to-digital converter (ADC) access, which is one of the main energy overhead in PIM. EGQ automatically optimizes layer-wise weight and activation bitwidth with negligible accuracy loss while considering the dynamic energy in PIM. The research demonstrates the effectiveness of EGQ on several DNN models VGG-19, ResNet-18, ResNet-50, MobileNet-V2, and SqueezeNet. Also, the area, dynamic energy, and energy efficiency in the compressed models with various memory technologies are analyzed. EGQ shows 15%-103% higher energy efficiency with 2% accuracy loss than other PIM-aware quantization algorithms.","abstract_has_math":false,"creators":["Kang, Beomseok"],"institution":"Georgia Institute of Technology","degree_name":null,"degree_level":"Masters","degree_discipline":null,"degree_department":"Electrical and Computer Engineering","school":null,"contributors":[],"advisors":["Mukhopadhyay, Saibal"],"committee_chairs":[],"committee_members":["Yu, Shimeng","Krishna, Tushar"],"year":2022,"date_issued":"2022-05-13","date_published":"2022-05-13","updated_at":"2026-07-27T19:50:35Z","subjects":["Quantization","Deep Neural Network","Genetic Algorithm","Processing-In-Memory"],"languages":["en_US"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/1853/67193","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Mukhopadhyay, Saibal"]},{"key":"dc:contributor.committeemember","label":"Committee Member","values":["Yu, Shimeng","Krishna, Tushar"]},{"key":"dc:contributor.department","label":"Department","values":["Electrical and Computer Engineering"]},{"key":"dc:creator","label":"Author","values":["Kang, Beomseok"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2022-08-25T13:31:40Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2022-08-25T13:31:40Z"]},{"key":"dc:date.issued","label":"Date","values":["2022-05-13"]},{"key":"dc:publisher","label":"Institution","values":["Georgia Institute of Technology"]},{"key":"dc:type","label":"Dc Type","values":["Text"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Masters"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Quantization","Deep Neural Network","Genetic Algorithm","Processing-In-Memory"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en_US"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["http://hdl.handle.net/1853/67193"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["With increasing computational cost of deep neural network (DNN), many efforts to develop energy-efficient intelligent system have been proposed from dedicated hardware platforms to model compression algorithms. Recently, hardware-aware quantization algorithms have shown further improvement in the energy efficiency of DNN by considering hardware architectures and algorithms together. In this work, a genetic algorithm-based energy-aware DNN quantization framework for Processing-In-Memory (PIM) architectures, named EGQ, is presented. The key contribution of the research is to design a fitness function that can reduce the number of analog-to-digital converter (ADC) access, which is one of the main energy overhead in PIM. EGQ automatically optimizes layer-wise weight and activation bitwidth with negligible accuracy loss while considering the dynamic energy in PIM. The research demonstrates the effectiveness of EGQ on several DNN models VGG-19, ResNet-18, ResNet-50, MobileNet-V2, and SqueezeNet. Also, the area, dynamic energy, and energy efficiency in the compressed models with various memory technologies are analyzed. EGQ shows 15%-103% higher energy efficiency with 2% accuracy loss than other PIM-aware quantization algorithms."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["M.S."]},{"key":"dc:format.mimetype","label":"Dc Format Mimetype","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Energy-aware DNN Quantization for Processing-In-Memory Architecture"]}]}],"canonical_facts":{"dc:contributor.advisor":["Mukhopadhyay, Saibal"],"dc:contributor.committeemember":["Yu, Shimeng","Krishna, Tushar"],"dc:contributor.department":["Electrical and Computer Engineering"],"dc:creator":["Kang, Beomseok"],"dc:date.accessioned":["2022-08-25T13:31:40Z"],"dc:date.available":["2022-08-25T13:31:40Z"],"dc:date.issued":["2022-05-13"],"dc:description.abstract":["With increasing computational cost of deep neural network (DNN), many efforts to develop energy-efficient intelligent system have been proposed from dedicated hardware platforms to model compression algorithms. Recently, hardware-aware quantization algorithms have shown further improvement in the energy efficiency of DNN by considering hardware architectures and algorithms together. In this work, a genetic algorithm-based energy-aware DNN quantization framework for Processing-In-Memory (PIM) architectures, named EGQ, is presented. The key contribution of the research is to design a fitness function that can reduce the number of analog-to-digital converter (ADC) access, which is one of the main energy overhead in PIM. EGQ automatically optimizes layer-wise weight and activation bitwidth with negligible accuracy loss while considering the dynamic energy in PIM. The research demonstrates the effectiveness of EGQ on several DNN models VGG-19, ResNet-18, ResNet-50, MobileNet-V2, and SqueezeNet. Also, the area, dynamic energy, and energy efficiency in the compressed models with various memory technologies are analyzed. EGQ shows 15%-103% higher energy efficiency with 2% accuracy loss than other PIM-aware quantization algorithms."],"dc:description.degree":["M.S."],"dc:format.mimetype":["application/pdf"],"dc:identifier.uri":["http://hdl.handle.net/1853/67193"],"dc:language.iso":["en_US"],"dc:publisher":["Georgia Institute of Technology"],"dc:subject":["Quantization","Deep Neural Network","Genetic Algorithm","Processing-In-Memory"],"dc:title":["Energy-aware DNN Quantization for Processing-In-Memory Architecture"],"dc:type":["Text"],"thesis:degree_level":["Masters"]},"updated_at":"2026-07-27T19:50:35Z"}