{"id":{"repo_id":"buffalo","oai_identifier":"oai:ubir.buffalo.edu:10477/80874"},"canonical_url":"https://search.dev.ndltd.org/etd/buffalo/oai:ubir.buffalo.edu:10477/80874","repository":{"repo_id":"buffalo","name":"Buffalo","base_url":"https://ubir.buffalo.edu/oai/request"},"display":{"title":"Learning from Disparate Data: Applications in Biometrics and Sustainability","abstract":"Ph.D.","abstract_html":"Ph.D.","abstract_has_math":false,"creators":["Pokhriyal, Neeti"],"institution":"State University of New York at Buffalo","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Govindaraju, Venu","Computer Science and Engineering"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2019,"date_issued":"2019-10-29T16:47:46Z","date_published":"2019-10-29T16:47:46Z","updated_at":"2026-07-27T19:05:25Z","subjects":["computer science"],"languages":["eng"],"rights":["Users of works found in University at Buffalo Institutional Repository (UBIR) are responsible for identifying and contacting the copyright owner for permission to reuse. University at Buffalo Libraries do not manage rights for copyright-protected works and cannot assist with permissions.","Copyright retained by author."],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/10477/80874","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Govindaraju, Venu","Computer Science and Engineering"]},{"key":"dc:creator","label":"Author","values":["Pokhriyal, Neeti"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2019-10-29T16:47:46Z","2019","2019-07-31 14:58:17"]},{"key":"dc:publisher","label":"Institution","values":["State University of New York at Buffalo"]},{"key":"dc:type","label":"Dc Type","values":["Text","Dissertation"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["computer science"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Users of works found in University at Buffalo Institutional Repository (UBIR) are responsible for identifying and contacting the copyright owner for permission to reuse. University at Buffalo Libraries do not manage rights for copyright-protected works and cannot assist with permissions.","Copyright retained by author."]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/10477/80874"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Ph.D.","Many learning problems involve data coming from multiple sources, sensors, modalities or feature spaces, that describe the object of interest in a unique way, and typically exhibit heterogeneous properties. The varied data sources are termed as views, and the task of learning from such multi-view data is known as multi-view learning. The term, “view”, is an encompassing term referring to different sets of observations having distinct statistical properties. Views can correspond to data from different sensors, sources, modalities or feature spaces extracted using different algorithms.The primary objective of this thesis is to develop methods for multi-view learning in a supervised setting. We have explored i). Multi-kernel Learning using Gaussian Processes which employs Bayesian uncertainty to combine multiple views, and, ii). Dis- criminative Factorized Subspace Learning, which produces a low-dimensional factorized subspace, consisting of shared and per-view components. We have shown the efficacy of our algorithms on two distinct application areas - behavioral biometrics, and poverty prediction and mapping.Gaussian Process (GP) based learning algorithm, exploit the Bayesian uncertainty associated with GP Regression, to combine data from multiple views. An advantage of this methodology is that the different data ecosystems (to be combined) need not share any data amongst them. The individual datasets remain private within their specific ecosystems, and only the output predictions and the associated uncertainties are shared. This is very beneficial in the area of poverty mapping with different data entities, which usually belong to different private and public companies.The concept of uncertainty is attractive in biometrics for the algorithm/model de- signers as well as the practitioners. It is important to know the uncertainty information, especially in cases where biometric information is coming from multiple sensors (say different fingerprint sensors); or different modalities (say fusion of fingerprint, iris etc with soft biometrics like gait, swipe, etc.).Discriminative Factorized Subspace algorithm is proposed to mitigate the strict assumptions of existing subspace learning algorithms, i.e. the various views are either completely independent or fully dependent. Methods operating under the former assumption typically involve multi-kernel learning, while those following the latter assumption aim at learning a shared latent subspace or manifold. However, in real scenarios, these assumptions are almost never truly satisfied; and thus motivating the idea of factorized subspace learning.In each of the target domains, the proposed methods have shown a significant improvement in terms of predictive performance over state-of-art methods.","**To request an accessible version of the file(s) associated with this item, contact library@buffalo.edu. Please include the item's persistent URL [http://hdl.handle.net/. . .] in your request.**"]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Learning from Disparate Data: Applications in Biometrics and Sustainability"]}]}],"canonical_facts":{"dc:contributor":["Govindaraju, Venu","Computer Science and Engineering"],"dc:creator":["Pokhriyal, Neeti"],"dc:date":["2019-10-29T16:47:46Z","2019","2019-07-31 14:58:17"],"dc:description":["Ph.D.","Many learning problems involve data coming from multiple sources, sensors, modalities or feature spaces, that describe the object of interest in a unique way, and typically exhibit heterogeneous properties. The varied data sources are termed as views, and the task of learning from such multi-view data is known as multi-view learning. The term, “view”, is an encompassing term referring to different sets of observations having distinct statistical properties. Views can correspond to data from different sensors, sources, modalities or feature spaces extracted using different algorithms.The primary objective of this thesis is to develop methods for multi-view learning in a supervised setting. We have explored i). Multi-kernel Learning using Gaussian Processes which employs Bayesian uncertainty to combine multiple views, and, ii). Dis- criminative Factorized Subspace Learning, which produces a low-dimensional factorized subspace, consisting of shared and per-view components. We have shown the efficacy of our algorithms on two distinct application areas - behavioral biometrics, and poverty prediction and mapping.Gaussian Process (GP) based learning algorithm, exploit the Bayesian uncertainty associated with GP Regression, to combine data from multiple views. An advantage of this methodology is that the different data ecosystems (to be combined) need not share any data amongst them. The individual datasets remain private within their specific ecosystems, and only the output predictions and the associated uncertainties are shared. This is very beneficial in the area of poverty mapping with different data entities, which usually belong to different private and public companies.The concept of uncertainty is attractive in biometrics for the algorithm/model de- signers as well as the practitioners. It is important to know the uncertainty information, especially in cases where biometric information is coming from multiple sensors (say different fingerprint sensors); or different modalities (say fusion of fingerprint, iris etc with soft biometrics like gait, swipe, etc.).Discriminative Factorized Subspace algorithm is proposed to mitigate the strict assumptions of existing subspace learning algorithms, i.e. the various views are either completely independent or fully dependent. Methods operating under the former assumption typically involve multi-kernel learning, while those following the latter assumption aim at learning a shared latent subspace or manifold. However, in real scenarios, these assumptions are almost never truly satisfied; and thus motivating the idea of factorized subspace learning.In each of the target domains, the proposed methods have shown a significant improvement in terms of predictive performance over state-of-art methods.","**To request an accessible version of the file(s) associated with this item, contact library@buffalo.edu. Please include the item's persistent URL [http://hdl.handle.net/. . .] in your request.**"],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/10477/80874"],"dc:language":["eng"],"dc:publisher":["State University of New York at Buffalo"],"dc:rights":["Users of works found in University at Buffalo Institutional Repository (UBIR) are responsible for identifying and contacting the copyright owner for permission to reuse. University at Buffalo Libraries do not manage rights for copyright-protected works and cannot assist with permissions.","Copyright retained by author."],"dc:subject":["computer science"],"dc:title":["Learning from Disparate Data: Applications in Biometrics and Sustainability"],"dc:type":["Text","Dissertation"]},"updated_at":"2026-07-27T19:05:25Z"}