Shivam Jadoun|Nov 29, 2022
- Indian-American professor receives Amazon Research Award to study evolving machine learning systems
Indian-American professor receives Amazon Research Award to study evolving machine learning systems
Pavithra Prabhakar was one of 74 recipients of the awards from Amazon Research Award for designing a tool that minimises negative user experiences.
NEW DELHI: An Indian-American professor of computer science at a university in the US has received an Amazon Research Award to design a tool that minimises negative user experiences.
Pavithra Prabhakar, who is the Peggy and Gary Edwards chair in engineering, was one of 74 recipients of the awards from Amazon, which also includes an unrestricted gift, access to more than 300 Amazon public datasets, and Amazon Web Services’ artificial intelligence and machine learning services and tools, the Kansas State University said in a statement.
Prabhakar currently serves as a programme director at the National Science Foundation while on sabbatical from K-State. The tool itself would be utilised to minimise disruptive changes to the user experience of machine learning-based software systems as the product is refined and retrained over time. She explained the issue with an example of a search engine tool being retrained to provide superior automated results but in the process, losing some of the results the end user expects to see.
"The broad objective of the project is to automatically characterise how much two versions of machine learning-based systems are similar or different," Prabhakar said. "These systems are regularly retrained to achieve superior performance; however, this does not often translate to a better user experience. This can be mitigated by equipping the design team with an automated tool that could highlight where and by how much the systems changed between different versions, thereby aiding the team in making decisions regarding the acceptability of the changes from a user experience perspective."
The proposed research will build on foundational concepts from process algebra and control theory to define mathematical notions of distance between different versions of machine learning systems and develop algorithms for outputting the similarity and dissimilarity between them. This automated tool will benefit design teams in making critical decisions about improving the user experience of machine learning-based intelligent software systems.
Prabhakar obtained her doctorate in computer science and a master's degree in applied mathematics from the University of Illinois at Urbana-Champaign, followed by a Center for the Mathematics of Information postdoctoral fellowship at the California Institute of Technology. She was named a Michelle Munson-Serban Simu Keystone research scholar and received the dean's award for excellence in research from the Carl R. Ice College of Engineering.
Prabhakar's research has been recognized with several prestigious awards, including the NSF CAREER award, the Office of Naval Research Young Investigator Award and the Marie Curie Career Integration Grant from the European Union.
To get in touch, write to us at firstname.lastname@example.org.