Danfeng (Daphne) Yao presentation

Dr. Danfeng (Daphne) Yao is a Professor of Computer Science at Virginia Tech. She is a Turner Faculty Fellow and CACI Faculty Fellow. Her research interests include building cyber defenses and data science solutions for healthcare, with a shared focus on accuracy and deployability. Her tool CryptoGuard helps large software companies and Apache projects harden their cryptographic code. Her patents on anomaly detection are influential in the industry, cited by patents from major cybersecurity firms and technology companies, including FireEye, Symantec, Qualcomm, Cisco, IBM, SAP, Boeing, and Palo Alto Networks. Daphne received her Ph.D. from Brown University (Computer Science), Master’s from Princeton University (Chemistry) and Indiana University (Computer Science), and B.S. from Peking University in China (Chemistry). Daphne is an IEEE Fellow for her contributions to enterprise data security and high-precision vulnerability screening. She is also an ACM Distinguished Scientist and a senior member of the National Academy of Inventors (NAI). She received the prestigious ACM CODASPY Lasting Research Award in 2021. Daphne is passionate about mentoring. As an ACM SIGSAC officer, she has led many community-level research and mentoring initiatives since 2017.

Abstract:
As the medical industry rapidly rolls out AI machine learning (ML) products that directly impact patients and their families, comprehensive and objective evaluation is a must. However, assessing and improving the correctness and fairness of AI/ML models are challenging due to many factors, e.g., data imbalance in the training data, lack of diverse test cases, and needing medical knowledge for ground truth. I will share our recent work towards trustworthy ML in digital health. We developed a data enrichment method that produces customized prediction models for underrepresented patient groups with improved accuracy. In another study, we used synthetic records to quantify how models respond to critical health conditions for mortality risk and cancer survivability predictions. To illustrate the importance of working with practitioners, I will share my experience developing secure coding tools and AI models for software developers. Finally, I will discuss my future research directions, including human-AI teaming, benchmark development, and community building to ensure AI/ML trustworthiness in critical applications.

Question? Contact Christine Hopper.