{"id":{"repo_id":"msu","oai_identifier":"oai:d.lib.msu.edu:etd_53154"},"canonical_url":"https://search.ndltd.org/etd/msu/oai:d.lib.msu.edu:etd_53154","repository":{"repo_id":"msu","name":"Michigan State University","base_url":"https://d.lib.msu.edu/oai"},"display":{"title":"RANDOMIZED METHODS FOR HIGH-DIMENSIONAL COMPUTATION : CONCENTRATION INEQUALITIES, SPECTRAL APPROXIMATION, AND STREAMING APPLICATIONS","abstract":"Thesis (Ph.D.)--Michigan State University. Mathematics - Doctor of Philosophy, 2026","abstract_html":"Thesis (Ph.D.)--Michigan State University. Mathematics - Doctor of Philosophy, 2026","abstract_has_math":false,"creators":["Boahen, Edem Kofi Konadu"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Iwen, Mark MI","Wong, Willie WW","Nelson, Brent BN","Huang, Longxiu LH"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-11-18T18:59:41Z","subjects":["Mathematics"],"languages":["eng"],"rights":["Attribution 4.0 International"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:53154","local:Boahen_grad.msu_0128D_22178"],"render_values":[{"text":"etd:53154","href":null,"code":true},{"text":"local:Boahen_grad.msu_0128D_22178","href":null,"code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://doi.org/doi:10.25335/01sv-f175","outbound_label":"DOI","outbound_source":"dc:identifier","record_url":"https://doi.org/doi:10.25335/01sv-f175","record_label":"Repository record","record_kind":"doi","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://d.lib.msu.edu/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ad.lib.msu.edu%3Aetd_53154","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Iwen, Mark MI","Wong, Willie WW","Nelson, Brent BN","Huang, Longxiu LH"]},{"key":"dc:creator","label":"Author","values":["Boahen, Edem Kofi Konadu"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026"]},{"key":"dc:relation","label":"Dc Relation","values":["Electronic Theses & Dissertations"]},{"key":"dc:type","label":"Dc Type","values":["text","Electronic dissertations"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Mathematics"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Attribution 4.0 International"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:53154","local:Boahen_grad.msu_0128D_22178","https://doi.org/doi:10.25335/01sv-f175"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Thesis (Ph.D.)--Michigan State University. Mathematics - Doctor of Philosophy, 2026","Modern data analysis frequently confronts the challenge of extracting meaningful structure fromextremely large and high-dimensional datasets. Whether the task is learning low-dimensional representations, identifying large clusters in networks, or approximating dominant directions of variation in a dataset, the underlying computational problems often involve operating on objects far too large to store or process directly. Classical algorithms with polynomial-time guarantees are increasingly impractical in such settings. This has driven significant interest in randomized methods and sketching techniques, which reduce the size or complexity of data while approximately preserving the structural properties of interest. At a high level, randomized embeddings and sketches serve as compressed “views” of data: they transform massive objects into smaller representations that retain sufficient information for downstream tasks. Over the past two decades, these methods have reshaped algorithm design in numerical linear algebra, signal processing, and machine learning. However, despite great progress, key challenges remain. Many sketching methods either rely on restrictive assumptions, or their guarantees degrade in infinite or streaming regimes. Furthermore, while randomized algorithms are known for practical speed, ensuring strong theoretical guarantees especially in large-scale or streaming contexts still remains a very active research area. This thesis contributes to this area by studying how structured randomness and compressed representations can be leveraged to design algorithms that are both scalable and provably reliable. Specifically, the work spans three interrelated directions: extending the Johnson-Lindenstrauss lemma to infinite sets and strengthening associated concentration inequalities; developing one-pass algorithms for sparse eigenvector approximation in massive matrices; and designing sketch-based power methods for cluster recovery in the stochastic block model with streaming guarantees. Together, these projects advance both the theoretical underpinnings and the algorithmic toolkit of randomized linear algebra and sketching, offering new perspectives on how to reconcile efficiency with rigorous guarantees in large-scale data analysis.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf","1 online resource (98 p.) : ill."]},{"key":"dc:title","label":"Title","values":["RANDOMIZED METHODS FOR HIGH-DIMENSIONAL COMPUTATION : CONCENTRATION INEQUALITIES, SPECTRAL APPROXIMATION, AND STREAMING APPLICATIONS"]}]}],"canonical_facts":{"dc:contributor":["Iwen, Mark MI","Wong, Willie WW","Nelson, Brent BN","Huang, Longxiu LH"],"dc:creator":["Boahen, Edem Kofi Konadu"],"dc:date":["2026"],"dc:description":["Thesis (Ph.D.)--Michigan State University. Mathematics - Doctor of Philosophy, 2026","Modern data analysis frequently confronts the challenge of extracting meaningful structure fromextremely large and high-dimensional datasets. Whether the task is learning low-dimensional representations, identifying large clusters in networks, or approximating dominant directions of variation in a dataset, the underlying computational problems often involve operating on objects far too large to store or process directly. Classical algorithms with polynomial-time guarantees are increasingly impractical in such settings. This has driven significant interest in randomized methods and sketching techniques, which reduce the size or complexity of data while approximately preserving the structural properties of interest. At a high level, randomized embeddings and sketches serve as compressed “views” of data: they transform massive objects into smaller representations that retain sufficient information for downstream tasks. Over the past two decades, these methods have reshaped algorithm design in numerical linear algebra, signal processing, and machine learning. However, despite great progress, key challenges remain. Many sketching methods either rely on restrictive assumptions, or their guarantees degrade in infinite or streaming regimes. Furthermore, while randomized algorithms are known for practical speed, ensuring strong theoretical guarantees especially in large-scale or streaming contexts still remains a very active research area. This thesis contributes to this area by studying how structured randomness and compressed representations can be leveraged to design algorithms that are both scalable and provably reliable. Specifically, the work spans three interrelated directions: extending the Johnson-Lindenstrauss lemma to infinite sets and strengthening associated concentration inequalities; developing one-pass algorithms for sparse eigenvector approximation in massive matrices; and designing sketch-based power methods for cluster recovery in the stochastic block model with streaming guarantees. Together, these projects advance both the theoretical underpinnings and the algorithmic toolkit of randomized linear algebra and sketching, offering new perspectives on how to reconcile efficiency with rigorous guarantees in large-scale data analysis.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."],"dc:format":["application/pdf","1 online resource (98 p.) : ill."],"dc:identifier":["etd:53154","local:Boahen_grad.msu_0128D_22178","https://doi.org/doi:10.25335/01sv-f175"],"dc:language":["eng"],"dc:relation":["Electronic Theses & Dissertations"],"dc:rights":["Attribution 4.0 International"],"dc:subject":["Mathematics"],"dc:title":["RANDOMIZED METHODS FOR HIGH-DIMENSIONAL COMPUTATION : CONCENTRATION INEQUALITIES, SPECTRAL APPROXIMATION, AND STREAMING APPLICATIONS"],"dc:type":["text","Electronic dissertations"]},"updated_at":"2026-11-18T18:59:41Z"}