{"id":{"repo_id":"vt","oai_identifier":"oai:vtechworks.lib.vt.edu:10919/137535"},"canonical_url":"https://search.dev.ndltd.org/etd/vt/oai:vtechworks.lib.vt.edu:10919/137535","repository":{"repo_id":"vt","name":"Virginia Tech","base_url":"https://vtechworks.lib.vt.edu/oai/request"},"display":{"title":"Advancing Catalysis Theory with Theory-infused Deep Learning","abstract":"Despite recent advances of data acquisition and algorithms development, machine learning (ML) faces tremendous challenges to being adopted in practical catalyst design, largely due to its limited generalizability and poor explainability. We developed a theory-infused neural network (TinNet) approach that integrates deep learning algorithms with catalysis theory. Incorporation of scientific knowledge of physical interactions into learning from data opens up new avenues for interpretable discovery of novel motifs with desired catalytic properties. The TinNet framework offers a robust platform for transforming ab initio data into physicochemical insights, enabling the design of novel catalytic materials. Its architecture, deeply rooted in the physics of electronic interactions, transcends algorithmic boundaries. TinNet not only sheds light on the fundamental characteristics of active sites but also enhances prediction accuracy in harmony with the physical principles governing catalytic surfaces. In this dissertation, We will highlight the role of TinNet in the rapid development of new catalytic materials, emphasizing its crucial contribution to sustainable chemical processes. The framework is particularly flexible at predicting surface reactivity, electronic structures, and cohesive energies, thus guiding the design and synthesis of an array of metallic systems, from single-atom alloys to complex high-entropy alloys. TinNet's unique fusion of theory and data-driven algorithms signifies a transformative step in the field of heterogeneous catalysis, one that could redefine the future of material design and sustainability.","abstract_html":"Despite recent advances of data acquisition and algorithms development, machine learning (ML) faces tremendous challenges to being adopted in practical catalyst design, largely due to its limited generalizability and poor explainability. We developed a theory-infused neural network (TinNet) approach that integrates deep learning algorithms with catalysis theory. Incorporation of scientific knowledge of physical interactions into learning from data opens up new avenues for interpretable discovery of novel motifs with desired catalytic properties. The TinNet framework offers a robust platform for transforming ab initio data into physicochemical insights, enabling the design of novel catalytic materials. Its architecture, deeply rooted in the physics of electronic interactions, transcends algorithmic boundaries. TinNet not only sheds light on the fundamental characteristics of active sites but also enhances prediction accuracy in harmony with the physical principles governing catalytic surfaces. In this dissertation, We will highlight the role of TinNet in the rapid development of new catalytic materials, emphasizing its crucial contribution to sustainable chemical processes. The framework is particularly flexible at predicting surface reactivity, electronic structures, and cohesive energies, thus guiding the design and synthesis of an array of metallic systems, from single-atom alloys to complex high-entropy alloys. TinNet's unique fusion of theory and data-driven algorithms signifies a transformative step in the field of heterogeneous catalysis, one that could redefine the future of material design and sustainability.","abstract_has_math":false,"creators":["Wang, Shih-Han"],"institution":"Virginia Tech","degree_name":"Doctor of Philosophy","degree_level":"doctoral","degree_discipline":"Chemical Engineering","degree_department":"Chemical Engineering","school":null,"contributors":[],"advisors":[],"committee_chairs":["Achenie, Luke E. K.","Xin, Hongliang"],"committee_members":["Deshmukh, Sanket A.","Bai, Xianming"],"year":2025,"date_issued":"2025-08-19","date_published":"2025-08-19","updated_at":"2026-07-22T22:20:40Z","subjects":["interpretable deep learning","reactivity descriptor","chemisorption model","tight binding theory","catalysis"],"languages":["en"],"rights":["In Copyright"],"rights_urls":["http://rightsstatements.org/vocab/InC/1.0/"],"identifier_entries":[{"key":"dc:identifier.other","label":"Dc Identifier Other","values":["vt_gsexam:44514"],"render_values":[{"text":"vt_gsexam:44514","href":null,"code":true}]}]},"links":{"outbound_url":"https://hdl.handle.net/10919/137535","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.committeechair","label":"Committee Chair","values":["Achenie, Luke E. K.","Xin, Hongliang"]},{"key":"dc:contributor.committeemember","label":"Committee Member","values":["Deshmukh, Sanket A.","Bai, Xianming"]},{"key":"dc:contributor.department","label":"Department","values":["Chemical Engineering"]},{"key":"dc:creator","label":"Author","values":["Wang, Shih-Han"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2025-08-20T08:00:26Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2025-08-20T08:00:26Z"]},{"key":"dc:date.issued","label":"Date","values":["2025-08-19"]},{"key":"dc:publisher","label":"Institution","values":["Virginia Tech"]},{"key":"dc:type","label":"Dc Type","values":["Dissertation"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Chemical Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Doctor of Philosophy"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["Virginia Polytechnic Institute and State University"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["interpretable deep learning","reactivity descriptor","chemisorption model","tight binding theory","catalysis"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["In Copyright"]},{"key":"dc:rights.uri","label":"Rights URI","values":["http://rightsstatements.org/vocab/InC/1.0/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.other","label":"Dc Identifier Other","values":["vt_gsexam:44514"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/10919/137535"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Despite recent advances of data acquisition and algorithms development, machine learning (ML) faces tremendous challenges to being adopted in practical catalyst design, largely due to its limited generalizability and poor explainability. We developed a theory-infused neural network (TinNet) approach that integrates deep learning algorithms with catalysis theory. Incorporation of scientific knowledge of physical interactions into learning from data opens up new avenues for interpretable discovery of novel motifs with desired catalytic properties. The TinNet framework offers a robust platform for transforming ab initio data into physicochemical insights, enabling the design of novel catalytic materials. Its architecture, deeply rooted in the physics of electronic interactions, transcends algorithmic boundaries. TinNet not only sheds light on the fundamental characteristics of active sites but also enhances prediction accuracy in harmony with the physical principles governing catalytic surfaces. In this dissertation, We will highlight the role of TinNet in the rapid development of new catalytic materials, emphasizing its crucial contribution to sustainable chemical processes. The framework is particularly flexible at predicting surface reactivity, electronic structures, and cohesive energies, thus guiding the design and synthesis of an array of metallic systems, from single-atom alloys to complex high-entropy alloys. TinNet's unique fusion of theory and data-driven algorithms signifies a transformative step in the field of heterogeneous catalysis, one that could redefine the future of material design and sustainability."]},{"key":"dc:description.abstractgeneral","label":"General Abstract","values":["Designing new materials for chemical reactions is essential for creating sustainable technologies such as clean energy and environmentally friendly manufacturing. However, finding the right materials—called catalysts—is often a long, complex, and expensive process. In recent years, scientists have increasingly turned to machine learning, a type of computer-based prediction technique, to accelerate this discovery. Unfortunately, most machine learning tools work like a \"black box\", making predictions without explaining why or how they work. This lack of transparency makes it difficult for scientists to trust and use them in real-world applications. In this dissertation, We present a new approach that combines machine learning with well-established scientific principles. This method, called TinNet, allows the computer not only to learn from data but also to follow known rules from chemistry and physics. By doing so, TinNet produces more accurate predictions and helps scientists understand why certain materials perform better than others. This approach is especially useful for understanding how metals behave at the atomic level, which is crucial for designing better catalysts. TinNet has been applied to predict how well different metal surfaces react with other substances and how stable they are. By combining scientific knowledge with computer algorithms, this work opens the door to faster, smarter, and more reliable development of new materials that support a cleaner and more sustainable future."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["Doctor of Philosophy"]},{"key":"dc:format.medium","label":"Dc Format Medium","values":["ETD"]},{"key":"dc:title","label":"Title","values":["Advancing Catalysis Theory with Theory-infused Deep Learning"]}]}],"canonical_facts":{"dc:contributor.committeechair":["Achenie, Luke E. K.","Xin, Hongliang"],"dc:contributor.committeemember":["Deshmukh, Sanket A.","Bai, Xianming"],"dc:contributor.department":["Chemical Engineering"],"dc:creator":["Wang, Shih-Han"],"dc:date.accessioned":["2025-08-20T08:00:26Z"],"dc:date.available":["2025-08-20T08:00:26Z"],"dc:date.issued":["2025-08-19"],"dc:description.abstract":["Despite recent advances of data acquisition and algorithms development, machine learning (ML) faces tremendous challenges to being adopted in practical catalyst design, largely due to its limited generalizability and poor explainability. We developed a theory-infused neural network (TinNet) approach that integrates deep learning algorithms with catalysis theory. Incorporation of scientific knowledge of physical interactions into learning from data opens up new avenues for interpretable discovery of novel motifs with desired catalytic properties. The TinNet framework offers a robust platform for transforming ab initio data into physicochemical insights, enabling the design of novel catalytic materials. Its architecture, deeply rooted in the physics of electronic interactions, transcends algorithmic boundaries. TinNet not only sheds light on the fundamental characteristics of active sites but also enhances prediction accuracy in harmony with the physical principles governing catalytic surfaces. In this dissertation, We will highlight the role of TinNet in the rapid development of new catalytic materials, emphasizing its crucial contribution to sustainable chemical processes. The framework is particularly flexible at predicting surface reactivity, electronic structures, and cohesive energies, thus guiding the design and synthesis of an array of metallic systems, from single-atom alloys to complex high-entropy alloys. TinNet's unique fusion of theory and data-driven algorithms signifies a transformative step in the field of heterogeneous catalysis, one that could redefine the future of material design and sustainability."],"dc:description.abstractgeneral":["Designing new materials for chemical reactions is essential for creating sustainable technologies such as clean energy and environmentally friendly manufacturing. However, finding the right materials—called catalysts—is often a long, complex, and expensive process. In recent years, scientists have increasingly turned to machine learning, a type of computer-based prediction technique, to accelerate this discovery. Unfortunately, most machine learning tools work like a \"black box\", making predictions without explaining why or how they work. This lack of transparency makes it difficult for scientists to trust and use them in real-world applications. In this dissertation, We present a new approach that combines machine learning with well-established scientific principles. This method, called TinNet, allows the computer not only to learn from data but also to follow known rules from chemistry and physics. By doing so, TinNet produces more accurate predictions and helps scientists understand why certain materials perform better than others. This approach is especially useful for understanding how metals behave at the atomic level, which is crucial for designing better catalysts. TinNet has been applied to predict how well different metal surfaces react with other substances and how stable they are. By combining scientific knowledge with computer algorithms, this work opens the door to faster, smarter, and more reliable development of new materials that support a cleaner and more sustainable future."],"dc:description.degree":["Doctor of Philosophy"],"dc:format.medium":["ETD"],"dc:identifier.other":["vt_gsexam:44514"],"dc:identifier.uri":["https://hdl.handle.net/10919/137535"],"dc:language.iso":["en"],"dc:publisher":["Virginia Tech"],"dc:rights":["In Copyright"],"dc:rights.uri":["http://rightsstatements.org/vocab/InC/1.0/"],"dc:subject":["interpretable deep learning","reactivity descriptor","chemisorption model","tight binding theory","catalysis"],"dc:title":["Advancing Catalysis Theory with Theory-infused Deep Learning"],"dc:type":["Dissertation"],"thesis:degree_discipline":["Chemical Engineering"],"thesis:degree_level":["doctoral"],"thesis:degree_name":["Doctor of Philosophy"],"thesis:institution_name":["Virginia Polytechnic Institute and State University"]},"updated_at":"2026-07-22T22:20:40Z"}