{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/104957"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/104957","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Robustness and generalization guarantees for statistical learning of generative models","abstract":"We apply tools from the classical statistical learning theory to analyze theoretical properties of modern machine learning problems that are typically phrased in the context of generative models. By combining standard methods based on the theory of empirical processes with ideas from optimal transport and signal recovery, we formally address the generalization and robustness guarantees for the existing and newly suggested algorithms. More specifically, we consider the following three problems: First, we tackle the problem of domain adaptation, where the training data and the test data are drawn from two distributions that are related but not identical. We devise an empirical risk minimization algorithm based on local worst-case risks, and provide generalization and excess risk guarantees of the learned hypothesis, that are robust to drifts in generative models. Second, we consider the learning of coding schemes, where the goal is to minimize the reconstruction risk of the original signal. It turns out that the task can be viewed as approximating the signal-generating distributions by pushforwards of arbitrary distributions via reconstruction maps. We provide learning guarantees based on the notions of optimal transport and classic statistical learning, using reconstruction errors as hypotheses. Third, we propose a framework of assessing representation learning algorithms by evaluating their estimation capabilities of the representation generating the signal. Using polyhedral estimates from the signal recovery literature, we investigate the provably near-optimal guarantees of the topic model.","abstract_html":"We apply tools from the classical statistical learning theory to analyze theoretical properties of modern machine learning problems that are typically phrased in the context of generative models. By combining standard methods based on the theory of empirical processes with ideas from optimal transport and signal recovery, we formally address the generalization and robustness guarantees for the existing and newly suggested algorithms. More specifically, we consider the following three problems: First, we tackle the problem of domain adaptation, where the training data and the test data are drawn from two distributions that are related but not identical. We devise an empirical risk minimization algorithm based on local worst-case risks, and provide generalization and excess risk guarantees of the learned hypothesis, that are robust to drifts in generative models. Second, we consider the learning of coding schemes, where the goal is to minimize the reconstruction risk of the original signal. It turns out that the task can be viewed as approximating the signal-generating distributions by pushforwards of arbitrary distributions via reconstruction maps. We provide learning guarantees based on the notions of optimal transport and classic statistical learning, using reconstruction errors as hypotheses. Third, we propose a framework of assessing representation learning algorithms by evaluating their estimation capabilities of the representation generating the signal. Using polyhedral estimates from the signal recovery literature, we investigate the provably near-optimal guarantees of the topic model.","abstract_has_math":false,"creators":["Lee, Jaeho"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Electrical & Computer Engr","degree_department":null,"school":null,"contributors":["Raginsky, Maxim","Srikant, Rayadurgam","Veeravalli, Venugopal","Dokmanić, Ivan"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2019,"date_issued":"2019-08-23T20:28:03Z","date_published":"2019-08-23T20:28:03Z","updated_at":"2026-07-22T22:24:44Z","subjects":["statistical learning","minimax learning","learning a coding scheme","representation learning"],"languages":["en"],"rights":["Copyright 2019 Jaeho Lee"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/104957","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Raginsky, Maxim","Srikant, Rayadurgam","Veeravalli, Venugopal","Dokmanić, Ivan"]},{"key":"dc:creator","label":"Author","values":["Lee, Jaeho"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2019-08-23T20:28:03Z","2021-08-24T09:15:10Z","2019-01-15","2019-05"]},{"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":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"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":["statistical learning","minimax learning","learning a coding scheme","representation 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 2019 Jaeho Lee"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/104957"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["We apply tools from the classical statistical learning theory to analyze theoretical properties of modern machine learning problems that are typically phrased in the context of generative models. By combining standard methods based on the theory of empirical processes with ideas from optimal transport and signal recovery, we formally address the generalization and robustness guarantees for the existing and newly suggested algorithms. More specifically, we consider the following three problems: First, we tackle the problem of domain adaptation, where the training data and the test data are drawn from two distributions that are related but not identical. We devise an empirical risk minimization algorithm based on local worst-case risks, and provide generalization and excess risk guarantees of the learned hypothesis, that are robust to drifts in generative models. Second, we consider the learning of coding schemes, where the goal is to minimize the reconstruction risk of the original signal. It turns out that the task can be viewed as approximating the signal-generating distributions by pushforwards of arbitrary distributions via reconstruction maps. We provide learning guarantees based on the notions of optimal transport and classic statistical learning, using reconstruction errors as hypotheses. Third, we propose a framework of assessing representation learning algorithms by evaluating their estimation capabilities of the representation generating the signal. Using polyhedral estimates from the signal recovery literature, we investigate the provably near-optimal guarantees of the topic model.","Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2021-05-01","The student, Jaeho Lee, accepted the attached license on 2019-01-15 at 12:21.","The student, Jaeho Lee, submitted this Dissertation for approval on 2019-01-15 at 12:28.","This Dissertation was approved for publication on 2019-01-15 at 14:39.","DSpace SAF Submission Ingestion Package generated from Vireo submission #13358 on 2019-08-22 at 15:03:55","Made available in DSpace on 2019-08-23T20:28:03Z (GMT). No. of bitstreams: 3 LEE-DISSERTATION-2019.pdf: 483199 bytes, checksum: 9539329342b03f2b23b6b8c1d77e6101 (MD5) LICENSE.txt: 4206 bytes, checksum: b2f6ea597681f93ebcd70bc7fec874f5 (MD5) PROQUEST_LICENSE.txt: 4552 bytes, checksum: f3ef11d7d719c01e9ddc6220d8ac9094 (MD5) Previous issue date: 2019-01-15","Embargo set by: Seth Robbins for item 112074 Lift date: 2021-08-23T20:28:11Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","Embargo set by: Seth Robbins for item 112074 Lift date: 2021-08-23T20:29:33Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","Embargo set by: Seth Robbins for item 112074 Lift date: 2021-08-23T20:36:18Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","U of I Only Restriction Lifted for Item 112074 on 2021-08-24T09:15:10Z."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Robustness and generalization guarantees for statistical learning of generative models"]}]}],"canonical_facts":{"dc:contributor":["Raginsky, Maxim","Srikant, Rayadurgam","Veeravalli, Venugopal","Dokmanić, Ivan"],"dc:creator":["Lee, Jaeho"],"dc:date":["2019-08-23T20:28:03Z","2021-08-24T09:15:10Z","2019-01-15","2019-05"],"dc:description":["We apply tools from the classical statistical learning theory to analyze theoretical properties of modern machine learning problems that are typically phrased in the context of generative models. By combining standard methods based on the theory of empirical processes with ideas from optimal transport and signal recovery, we formally address the generalization and robustness guarantees for the existing and newly suggested algorithms. More specifically, we consider the following three problems: First, we tackle the problem of domain adaptation, where the training data and the test data are drawn from two distributions that are related but not identical. We devise an empirical risk minimization algorithm based on local worst-case risks, and provide generalization and excess risk guarantees of the learned hypothesis, that are robust to drifts in generative models. Second, we consider the learning of coding schemes, where the goal is to minimize the reconstruction risk of the original signal. It turns out that the task can be viewed as approximating the signal-generating distributions by pushforwards of arbitrary distributions via reconstruction maps. We provide learning guarantees based on the notions of optimal transport and classic statistical learning, using reconstruction errors as hypotheses. Third, we propose a framework of assessing representation learning algorithms by evaluating their estimation capabilities of the representation generating the signal. Using polyhedral estimates from the signal recovery literature, we investigate the provably near-optimal guarantees of the topic model.","Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2021-05-01","The student, Jaeho Lee, accepted the attached license on 2019-01-15 at 12:21.","The student, Jaeho Lee, submitted this Dissertation for approval on 2019-01-15 at 12:28.","This Dissertation was approved for publication on 2019-01-15 at 14:39.","DSpace SAF Submission Ingestion Package generated from Vireo submission #13358 on 2019-08-22 at 15:03:55","Made available in DSpace on 2019-08-23T20:28:03Z (GMT). No. of bitstreams: 3 LEE-DISSERTATION-2019.pdf: 483199 bytes, checksum: 9539329342b03f2b23b6b8c1d77e6101 (MD5) LICENSE.txt: 4206 bytes, checksum: b2f6ea597681f93ebcd70bc7fec874f5 (MD5) PROQUEST_LICENSE.txt: 4552 bytes, checksum: f3ef11d7d719c01e9ddc6220d8ac9094 (MD5) Previous issue date: 2019-01-15","Embargo set by: Seth Robbins for item 112074 Lift date: 2021-08-23T20:28:11Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","Embargo set by: Seth Robbins for item 112074 Lift date: 2021-08-23T20:29:33Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","Embargo set by: Seth Robbins for item 112074 Lift date: 2021-08-23T20:36:18Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","U of I Only Restriction Lifted for Item 112074 on 2021-08-24T09:15:10Z."],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/2142/104957"],"dc:language":["en"],"dc:rights":["Copyright 2019 Jaeho Lee"],"dc:subject":["statistical learning","minimax learning","learning a coding scheme","representation learning"],"dc:title":["Robustness and generalization guarantees for statistical learning of generative models"],"dc:type":["text"],"thesis:degree_discipline":["Electrical & Computer Engr"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:44Z"}