Dr. Olivera Kotevska is a research scientist in the Computer Science and Mathematics Division (CSMD) at Oak Ridge National Laboratory (ORNL) in Tennessee, USA. She earned her Ph.D. in Computer Science from Université Grenoble Alpes, France, and conducted research at the National Institute of Standards and Technology (NIST) in Washington, DC, during her doctoral studies. Her research focuses on machine learning, secure AI, and their intersection with scientific applications. She has held various organizational roles, including chairing IEEE subcommittees, editorial board of ACM Transactions on Internet Technology and Sensors IoT Data Analytics journals, and organizing numerous workshops and programs at premier AI conferences. Dr. Kotevska was awarded IEEE Senior Membership and received the ORNL’s CSMD Outstanding Mentorship and Community Outreach Awards.
Presentation Title:
Enabling Trustworthy Federated Learning Thought Privacy and Explainability
Presentation Abstract:
Federated Learning (FL) enables collaborative model training without sharing raw data, making it well-suited for privacy-sensitive applications. However, trust remains a challenge due to risks of information leakage and limited transparency—especially when clients have varying privacy needs. This presentation introduces an approach that combines differential privacy with client-level explainability to address heterogeneous privacy requirements. By supporting heterogeneous privacy budgets and interpretable model behavior, the framework improves trust without sacrificing performance. We demonstrate its effectiveness in high-stakes scenarios, highlighting a practical path toward deploying secure, transparent, and trustworthy FL systems across diverse and distributed environments.