{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/99324"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/99324","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Analytical guarantees for reduced precision fixed-point margin hyperplane classifiers","abstract":"Margin hyperplane classifiers such as support vector machines are strong predictive models having gained considerable success in various classification tasks. Their conceptual simplicity makes them suitable candidates for the design of embedded machine learning systems. Their accuracy and resource utilization can effectively be traded off each other through precision. We analytically capture this trade-off by means of bounds on the precision requirements of general margin hyperplane classifiers. In addition, we propose a principled precision reduction scheme based on the trade-off between input and weight precisions. Our analysis is supported by simulation results illustrating the gains of our approach in terms of reducing resource utilization. For instance, we show that a linear margin classifier with precision assignment dictated by our approach and applied to the `two vs. four' task of the MNIST dataset is ~2x more accurate than a standard 8 bit low-precision implementation in spite of using ~2x10^4 fewer 1 bit full adders and ~2x10^3 fewer bits for data and weight representation.","abstract_html":"Margin hyperplane classifiers such as support vector machines are strong predictive models having gained considerable success in various classification tasks. Their conceptual simplicity makes them suitable candidates for the design of embedded machine learning systems. Their accuracy and resource utilization can effectively be traded off each other through precision. We analytically capture this trade-off by means of bounds on the precision requirements of general margin hyperplane classifiers. In addition, we propose a principled precision reduction scheme based on the trade-off between input and weight precisions. Our analysis is supported by simulation results illustrating the gains of our approach in terms of reducing resource utilization. For instance, we show that a linear margin classifier with precision assignment dictated by our approach and applied to the `two vs. four' task of the MNIST dataset is ~2x more accurate than a standard 8 bit low-precision implementation in spite of using ~2x10^4 fewer 1 bit full adders and ~2x10^3 fewer bits for data and weight representation.","abstract_has_math":false,"creators":["Sakr, Charbel"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Electrical & Computer Engr","degree_department":null,"school":null,"contributors":["Shanbhag, Naresh R."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2018,"date_issued":"2018-03-13T15:45:14Z","date_published":"2018-03-13T15:45:14Z","updated_at":"2026-07-22T22:24:37Z","subjects":["Fixed-point machine learning"],"languages":["en"],"rights":["Copyright 2017 Charbel Sakr"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/99324","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Shanbhag, Naresh R."]},{"key":"dc:creator","label":"Author","values":["Sakr, Charbel"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2018-03-13T15:45:14Z","2017-11-10","2017-12"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical & Computer Engr"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis"]},{"key":"thesis:degree_name","label":"Degree Name","values":["M.S."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Fixed-point machine learning"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2017 Charbel Sakr"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/99324"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Margin hyperplane classifiers such as support vector machines are strong predictive models having gained considerable success in various classification tasks. Their conceptual simplicity makes them suitable candidates for the design of embedded machine learning systems. Their accuracy and resource utilization can effectively be traded off each other through precision. We analytically capture this trade-off by means of bounds on the precision requirements of general margin hyperplane classifiers. In addition, we propose a principled precision reduction scheme based on the trade-off between input and weight precisions. Our analysis is supported by simulation results illustrating the gains of our approach in terms of reducing resource utilization. For instance, we show that a linear margin classifier with precision assignment dictated by our approach and applied to the `two vs. four' task of the MNIST dataset is ~2x more accurate than a standard 8 bit low-precision implementation in spite of using ~2x10^4 fewer 1 bit full adders and ~2x10^3 fewer bits for data and weight representation.","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2018-03-13 without embargo terms","The student, Charbel Sakr, accepted the attached license on 2017-11-10 at 12:11.","The student, Charbel Sakr, submitted this Thesis for approval on 2017-11-10 at 12:16.","This Thesis was approved for publication on 2017-11-10 at 14:27.","DSpace SAF Submission Ingestion Package generated from Vireo submission #11730 on 2018-03-13 at 10:08:31","Made available in DSpace on 2018-03-13T15:45:14Z (GMT). No. of bitstreams: 2 SAKR-THESIS-2017.pdf: 3911673 bytes, checksum: 581462cb8585a034d80d04aad8538bd1 (MD5) LICENSE.txt: 4209 bytes, checksum: baa08cc49c0df5b657837b5e3310574f (MD5) Previous issue date: 2017-11-10"]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Analytical guarantees for reduced precision fixed-point margin hyperplane classifiers"]}]}],"canonical_facts":{"dc:contributor":["Shanbhag, Naresh R."],"dc:creator":["Sakr, Charbel"],"dc:date":["2018-03-13T15:45:14Z","2017-11-10","2017-12"],"dc:description":["Margin hyperplane classifiers such as support vector machines are strong predictive models having gained considerable success in various classification tasks. Their conceptual simplicity makes them suitable candidates for the design of embedded machine learning systems. Their accuracy and resource utilization can effectively be traded off each other through precision. We analytically capture this trade-off by means of bounds on the precision requirements of general margin hyperplane classifiers. In addition, we propose a principled precision reduction scheme based on the trade-off between input and weight precisions. Our analysis is supported by simulation results illustrating the gains of our approach in terms of reducing resource utilization. For instance, we show that a linear margin classifier with precision assignment dictated by our approach and applied to the `two vs. four' task of the MNIST dataset is ~2x more accurate than a standard 8 bit low-precision implementation in spite of using ~2x10^4 fewer 1 bit full adders and ~2x10^3 fewer bits for data and weight representation.","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2018-03-13 without embargo terms","The student, Charbel Sakr, accepted the attached license on 2017-11-10 at 12:11.","The student, Charbel Sakr, submitted this Thesis for approval on 2017-11-10 at 12:16.","This Thesis was approved for publication on 2017-11-10 at 14:27.","DSpace SAF Submission Ingestion Package generated from Vireo submission #11730 on 2018-03-13 at 10:08:31","Made available in DSpace on 2018-03-13T15:45:14Z (GMT). No. of bitstreams: 2 SAKR-THESIS-2017.pdf: 3911673 bytes, checksum: 581462cb8585a034d80d04aad8538bd1 (MD5) LICENSE.txt: 4209 bytes, checksum: baa08cc49c0df5b657837b5e3310574f (MD5) Previous issue date: 2017-11-10"],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/2142/99324"],"dc:language":["en"],"dc:rights":["Copyright 2017 Charbel Sakr"],"dc:subject":["Fixed-point machine learning"],"dc:title":["Analytical guarantees for reduced precision fixed-point margin hyperplane classifiers"],"dc:type":["text"],"thesis:degree_discipline":["Electrical & Computer Engr"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:37Z"}