{"id":{"repo_id":"radboud","oai_identifier":"oai:repository.ubn.ru.nl:2066/319701"},"canonical_url":"https://search.dev.ndltd.org/etd/radboud/oai:repository.ubn.ru.nl:2066/319701","repository":{"repo_id":"radboud","name":"Radboud University Nijmegen","base_url":"https://repository.ubn.ru.nl/oai/request"},"display":{"title":"Beyond the Security of Deep Learning: An Exploration of Stealthy Backdoor Attacks in Computer Vision","abstract":"Contains fulltext : 319701.pdf (Publisher’s version ) (Open Access)","abstract_html":"Contains fulltext : 319701.pdf (Publisher’s version ) (Open Access)","abstract_has_math":false,"creators":["Abad, G."],"institution":"S.l. : s.n.","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Batina, L.","Picek, S.","Urbieta, A."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-07-24T04:03:01Z","subjects":["Digital Security"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["9789464963922"],"render_values":[{"text":"9789464963922","href":null,"code":true}]}]},"links":{"outbound_url":"https://hdl.handle.net/2066/319701","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Batina, L.","Picek, S.","Urbieta, A."]},{"key":"dc:creator","label":"Author","values":["Abad, G."]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025"]},{"key":"dc:publisher","label":"Institution","values":["S.l. : s.n."]},{"key":"dc:type","label":"Dc Type","values":["Doctoral thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Digital Security"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://repository.ubn.ru.nl//bitstream/handle/2066/319701/319701.pdf","https://hdl.handle.net/2066/319701","9789464963922"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Contains fulltext : 319701.pdf (Publisher’s version ) (Open Access)","This thesis investigates the security of deep learning systems, with a particular focus on backdoor attacks—a form of data poisoning where models behave normally under typical inputs but produce attacker-controlled outputs when a specific trigger is present. The research systematically analyzes the effectiveness and stealthiness of such attacks across a range of modern machine learning settings, including convolutional neural networks, spiking neural networks with neuromorphic data, vision transformers, and federated learning systems. The findings show that backdoor success depends heavily on trigger design, model architecture, and training conditions. Larger models and those trained from scratch tend to be more vulnerable. In decentralized and neuromorphic contexts, novel attack strategies are introduced that exploit the structure of data and training workflows, achieving high attack success rates while remaining undetected. The evaluation of common defenses reveals that many are ineffective against more sophisticated or context-specific attacks. Overall, the work highlights the growing complexity of securing machine learning systems and the need for defense mechanisms that are robust across architectures, data modalities, and deployment scenarios.","Radboud University, 23 juni 2025","Promotor : Batina, L. Co-promotores : Picek, S., Urbieta, A.","xix, 274 p."]},{"key":"dc:title","label":"Title","values":["Beyond the Security of Deep Learning: An Exploration of Stealthy Backdoor Attacks in Computer Vision"]}]}],"canonical_facts":{"dc:contributor":["Batina, L.","Picek, S.","Urbieta, A."],"dc:creator":["Abad, G."],"dc:date":["2025"],"dc:description":["Contains fulltext : 319701.pdf (Publisher’s version ) (Open Access)","This thesis investigates the security of deep learning systems, with a particular focus on backdoor attacks—a form of data poisoning where models behave normally under typical inputs but produce attacker-controlled outputs when a specific trigger is present. The research systematically analyzes the effectiveness and stealthiness of such attacks across a range of modern machine learning settings, including convolutional neural networks, spiking neural networks with neuromorphic data, vision transformers, and federated learning systems. The findings show that backdoor success depends heavily on trigger design, model architecture, and training conditions. Larger models and those trained from scratch tend to be more vulnerable. In decentralized and neuromorphic contexts, novel attack strategies are introduced that exploit the structure of data and training workflows, achieving high attack success rates while remaining undetected. The evaluation of common defenses reveals that many are ineffective against more sophisticated or context-specific attacks. Overall, the work highlights the growing complexity of securing machine learning systems and the need for defense mechanisms that are robust across architectures, data modalities, and deployment scenarios.","Radboud University, 23 juni 2025","Promotor : Batina, L. Co-promotores : Picek, S., Urbieta, A.","xix, 274 p."],"dc:identifier":["https://repository.ubn.ru.nl//bitstream/handle/2066/319701/319701.pdf","https://hdl.handle.net/2066/319701","9789464963922"],"dc:publisher":["S.l. : s.n."],"dc:subject":["Digital Security"],"dc:title":["Beyond the Security of Deep Learning: An Exploration of Stealthy Backdoor Attacks in Computer Vision"],"dc:type":["Doctoral thesis"]},"updated_at":"2026-07-24T04:03:01Z"}