Your Colleagues

Ratha named fellow of National Academy of Inventors

By TESSIE MAR

Published January 11, 2022

Print
headshot of Nalini Ratha.

Nalini Ratha, Empire Innovation Professor in the Department of Computer Science and Engineering, has been selected as a National Academy of Inventors Fellow for the class of 2021.

The NAI fellows program includes 1,403 fellows across the globe from more than 250 universities, and governmental and nonprofit research institutions. Fellows further research and advancements in innovation through their 42,700-and-counting U.S patents, creating licensed technologies, companies and jobs.

In a joint statement, Venu Govindaraju, vice president for research and economic development, and Kemper Lewis, dean of the School of Engineering and Applied Sciences, praised Ratha’s scholarship.

“Nalini Ratha is a leader in biometrics and has achieved worldwide recognition for his cutting-edge work in biometrics recognition, security and privacy, and performance evaluation, both in terms of highly citation publications and a long list of innovations captured through his patents,” Govindaraju and Lewis said. “His research and innovation leadership, as well as his numerous professional contributions, make him uniquely and exceptionally qualified for the elevation to NAI fellow.”

Ratha, who joined the UB faculty in July 2020 after working for IBM for nearly 25 years, holds more than 85 patents as either an inventor or co-inventor of a wide range of biometric, cognitive and privacy enhancements, and artificial intelligence technology-related inventions.

He has dedicated his career to studying biometric security and privacy, improving fundamental biometrics recognition and developing performance evaluation for biometric systems. His cancelable biometrics techniques to improve privacy in biometrics data continues to attract significant academic interest after two decades.

Notable accomplishments include designing an 11-point attack through his analysis of security threats to biometric systems. This strategy is used internationally and has grown to encompass adversarial attacks in deep learning systems. Ratha also devised a method to compute confidence intervals in performance evaluation by using “subsets bootstrap” — a variant of the bootstrap technique that is cited in biometrics standards documents.  

Scholars view Ratha’s research as an example of practical trustworthy AI systems, such as reducing bias in face analytics, detection and mitigation of adversarial attacks and learning on encrypted data.

During his time as a research staff member at IBM Thomas J. Watson Research Center, he worked on the company’s DiF face dataset creation. To improve adversarial attacks in face-recognition systems, he developed novel techniques to detect and mitigate the attack to bring back the original recognition performance in midst of adversarial noise. More recently, he has been working on using fully homomorphic encryption methods applied to machine learning to build models that can handle private encrypted data.

He is a fellow of the Institute of Electrical and Electronics Engineers (IEEE) and International Association of Pattern Recognition (IAPR), and a distinguished scientist of the Association for Computing Machinery (ACM).

Ratha has been recognized for his excellence in innovation through such awards as the Research Division award and an Outstanding Technical Innovation Award. He was named an IBM Research Master Inventor in 2018 and was a recipient of the IEEE Biometrics Council Leadership Award.

As editor-in-chief of the biometric journal IEEE’s Transactions on Biometrics, Behavior, and Identity Science, Ratha is heavily involved in the publication of research. He is also co-author of the textbook “A Guide to Biometrics Selection and System Design” and co-edited four books: “Deep Learning-Based Face Analytics,” “Domain Adaptation in Visual Understanding,” “Advances in Biometrics: Sensors, Algorithms and Systems,” and “Automatic Fingerprint Recognition Systems.”