{"id":{"repo_id":"calpoly","oai_identifier":"oai:digitalcommons.calpoly.edu:theses-4678"},"canonical_url":"https://search.ndltd.org/etd/calpoly/oai:digitalcommons.calpoly.edu:theses-4678","repository":{"repo_id":"calpoly","name":"Cal Poly","base_url":"https://digitalcommons.calpoly.edu/do/oai/"},"display":{"title":"Filling Gaps in Scientific Data Sets Using Physics Informed Neural Networks: A Case Study in Velocity Fields","abstract":"Gaps in scientific data sets are a persistent issue for researchers in a variety of fields, and while nothing makes up for missing out on real data, well-simulated synthetic data can be a useful tool. In the world of image processing, machine learning techniques have become quite sophisticated at taking an image with a missing component and filling in that space with something believable. The aim of this thesis is to take machine learning techniques similar to what gets used in image processing and repurpose them to infill gaps in scientific data sets in a realistic manner. This thesis compares and contrasts some machine learning techniques, such as basic neural networks, convolutional neural networks, and autoencoders, before focusing on neural networks. Two approaches were considered for this problem: an approach inspired by the idea of taking the weighted average of a point's neighbors to impute that point and an approach grounded in spatial coordinates. Then, a physics informed component is added to both of these networks to enforce certain physical realism standards. The physics informed component makes the neural networks marginally more accurate but noticeably more realistic and physically possible. Each of the networks, both before and after the physics informed component is added, is tested on a variety of velocity fields. Finally, a statistical analysis is done on the networks as compared to less sophisticated infill methods to demonstrate that they do perform well.","abstract_html":"Gaps in scientific data sets are a persistent issue for researchers in a variety of fields, and while nothing makes up for missing out on real data, well-simulated synthetic data can be a useful tool. In the world of image processing, machine learning techniques have become quite sophisticated at taking an image with a missing component and filling in that space with something believable. The aim of this thesis is to take machine learning techniques similar to what gets used in image processing and repurpose them to infill gaps in scientific data sets in a realistic manner. This thesis compares and contrasts some machine learning techniques, such as basic neural networks, convolutional neural networks, and autoencoders, before focusing on neural networks. Two approaches were considered for this problem: an approach inspired by the idea of taking the weighted average of a point&#x27;s neighbors to impute that point and an approach grounded in spatial coordinates. Then, a physics informed component is added to both of these networks to enforce certain physical realism standards. The physics informed component makes the neural networks marginally more accurate but noticeably more realistic and physically possible. Each of the networks, both before and after the physics informed component is added, is tested on a variety of velocity fields. Finally, a statistical analysis is done on the networks as compared to less sophisticated infill methods to demonstrate that they do perform well.","abstract_has_math":false,"creators":["Saunders, Ellen"],"institution":null,"degree_name":"MS in Mathematics","degree_level":null,"degree_discipline":"Mathematics","degree_department":null,"school":null,"contributors":["Paul Choboter","Mathematics","College of Science and Mathematics"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":null,"date_issued":"","date_published":null,"updated_at":"2026-11-19T07:12:29Z","subjects":["Neural Networks","Infill","Velocity Fields","Numerical Analysis and Computation","Numerical Analysis and Scientific Computing","Oceanography","OS and Networks","Other Applied Mathematics","Special Functions","Theory and Algorithms"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://digitalcommons.calpoly.edu/theses/3015","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://digitalcommons.calpoly.edu/theses/3015","record_label":"Repository record","record_kind":"record","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://digitalcommons.calpoly.edu/do/oai/?verb=GetRecord&metadataPrefix=oai_etdms&identifier=oai%3Adigitalcommons.calpoly.edu%3Atheses-4678","prefix":"oai_etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Paul Choboter","Mathematics","College of Science and Mathematics"]},{"key":"dc:creator","label":"Author","values":["Saunders, Ellen"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.available","label":"Dc Date Available","values":["2025-06-10T07:00:00Z"]},{"key":"dc:date.created","label":"Dc Date Created","values":["2025-06-01T07:00:00Z"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Mathematics"]},{"key":"thesis:degree_name","label":"Degree Name","values":["MS in Mathematics"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Neural Networks","Infill","Velocity Fields","Numerical Analysis and Computation","Numerical Analysis and Scientific Computing","Oceanography","OS and Networks","Other Applied Mathematics","Special Functions","Theory and Algorithms"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://digitalcommons.calpoly.edu/theses/3015"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Gaps in scientific data sets are a persistent issue for researchers in a variety of fields, and while nothing makes up for missing out on real data, well-simulated synthetic data can be a useful tool. In the world of image processing, machine learning techniques have become quite sophisticated at taking an image with a missing component and filling in that space with something believable. The aim of this thesis is to take machine learning techniques similar to what gets used in image processing and repurpose them to infill gaps in scientific data sets in a realistic manner. This thesis compares and contrasts some machine learning techniques, such as basic neural networks, convolutional neural networks, and autoencoders, before focusing on neural networks. Two approaches were considered for this problem: an approach inspired by the idea of taking the weighted average of a point's neighbors to impute that point and an approach grounded in spatial coordinates. Then, a physics informed component is added to both of these networks to enforce certain physical realism standards. The physics informed component makes the neural networks marginally more accurate but noticeably more realistic and physically possible. Each of the networks, both before and after the physics informed component is added, is tested on a variety of velocity fields. Finally, a statistical analysis is done on the networks as compared to less sophisticated infill methods to demonstrate that they do perform well."]},{"key":"dc:title","label":"Title","values":["Filling Gaps in Scientific Data Sets Using Physics Informed Neural Networks: A Case Study in Velocity Fields"]}]}],"canonical_facts":{"dc:contributor":["Paul Choboter","Mathematics","College of Science and Mathematics"],"dc:creator":["Saunders, Ellen"],"dc:date.available":["2025-06-10T07:00:00Z"],"dc:date.created":["2025-06-01T07:00:00Z"],"dc:description.abstract":["Gaps in scientific data sets are a persistent issue for researchers in a variety of fields, and while nothing makes up for missing out on real data, well-simulated synthetic data can be a useful tool. In the world of image processing, machine learning techniques have become quite sophisticated at taking an image with a missing component and filling in that space with something believable. The aim of this thesis is to take machine learning techniques similar to what gets used in image processing and repurpose them to infill gaps in scientific data sets in a realistic manner. This thesis compares and contrasts some machine learning techniques, such as basic neural networks, convolutional neural networks, and autoencoders, before focusing on neural networks. Two approaches were considered for this problem: an approach inspired by the idea of taking the weighted average of a point's neighbors to impute that point and an approach grounded in spatial coordinates. Then, a physics informed component is added to both of these networks to enforce certain physical realism standards. The physics informed component makes the neural networks marginally more accurate but noticeably more realistic and physically possible. Each of the networks, both before and after the physics informed component is added, is tested on a variety of velocity fields. Finally, a statistical analysis is done on the networks as compared to less sophisticated infill methods to demonstrate that they do perform well."],"dc:identifier":["https://digitalcommons.calpoly.edu/theses/3015"],"dc:subject":["Neural Networks","Infill","Velocity Fields","Numerical Analysis and Computation","Numerical Analysis and Scientific Computing","Oceanography","OS and Networks","Other Applied Mathematics","Special Functions","Theory and Algorithms"],"dc:title":["Filling Gaps in Scientific Data Sets Using Physics Informed Neural Networks: A Case Study in Velocity Fields"],"thesis:degree_discipline":["Mathematics"],"thesis:degree_name":["MS in Mathematics"]},"updated_at":"2026-11-19T07:12:29Z"}