11 May 2023 — In a significant achievement for the National University of Singapore (NUS) School of Computing, Professor Reza Shokri, holding the distinguished position of NUS Presidential Young Professor in Computer Science, has been recognized as one of the inaugural recipients of the esteemed Asian Young Scientist Fellowship (AYSF) for 2023.
Established in 2022, the AYSF Fellowship honors twelve outstanding early-career scientists from across Asia every year. The Fellowship aims to fuel the innovative pursuits of these young researchers, encouraging them to explore pioneering and creative approaches in their fields of study. Emphasizing transformative and interdisciplinary research, the Fellowship provides each awardee with $100,000 USD over a two-year term to support their research endeavors.
Prof. Shokri's pioneering research at NUS delves into the intricate relationship between data privacy and reliable machine learning. His work has been pivotal in uncovering potential privacy risks that could emanate from machine learning algorithms. Despite its many advantages, Prof. Shokri's research has highlighted that machine learning could inadvertently expose sensitive information about their training datasets.
A highlight of Prof. Shokri's contributions is his establishment of quantitative metrics for assessing privacy in machine learning. These metrics present a systematic method to analyze and differentiate various privacy-preserving machine learning methods, assisting organizations in formulating their privacy strategies. Prof. Shokri has made fundamental contributions to the development of membership inference attacks, which are designed to ascertain if a particular data point was part of a machine learning model's training set. His analytical framework for examining these attacks offers vital insights to counter potential privacy risks associated with a broad range of identification and data reconstruction attacks. His research has additionally resulted in the creation of the Privacy Meter, an open-source tool designed for the quantitative evaluation of privacy risks in machine learning. Prof. Shokri has also been integral in laying the groundwork for safeguarding privacy in machine learning through the application of differential privacy. This methodology involves the deliberate addition of specified noise into computations, with the objective of reducing the potential threats linked to inference attacks in machine learning. He has further developed mathematical frameworks to meticulously track and regulate privacy loss in algorithms implementing differential privacy. Moreover, his work has explored the trade-off between privacy and various facets of trustworthy machine learning, such as fairness, robustness, and explainability.
Upon receiving this honor, Prof. Shokri expressed, "Being awarded the Asian Young Scientist Fellowship is truly a privilege. This recognition emphasizes the significance and potential influence of our work in data privacy and machine learning. Our fundamental goal is to capitalize on the advantages of machine learning without breaching privacy and data protection norms. I am excited about the opportunity to extend this research and wish to express my heartfelt gratitude to AYSF for their support."
This recognition accentuates not just Prof. Shokri's individual achievements but also the increasingly acknowledged importance of data privacy in the era of machine learning. As the integration of AI into our daily lives continues to grow, the contribution of committed scientists like Prof. Shokri becomes crucial in shaping a future where technological progress and privacy can coexist harmoniously. We congratulate Prof. Shokri on this well-deserved recognition and look forward to his continued contributions to the field.