An emeritus professor taps his engineering acumen to explore the materials used by Van Gogh, Vermeer, and more. Read more about ‘Computational Art History’ Finds Clues in the Canvas
Ziv Goldfeld receives IBM 2020 University Award
Ziv Goldfeld, assistant professor in electrical and computer engineering, received an IBM 2020 University Award for the theoretical machine learning research he is doing jointly with collaborators from the IBM-MIT Watson AI Lab. Goldfeld was nominated by his IBM collaborators, and the award affirms IBM's interest and commitment to their work.
“We are looking at several machine learning projects,” Goldfeld said. “All revolve around and work towards developing a comprehensive statistical learning theory that can explain and provide strong performance guarantees for modern machine learning methods.”
Goldfeld’s team is also interested in developing a toolkit for visualizing and interpreting the operation of deep learning systems. The goal is to “open the hood” of deep neural networks, which form the backbone of cutting-edge artificial intelligence technologies.
“By making the decisions of AI mechanisms more transparent to end users and other stakeholders, and certifying their performance against potential failure modes,” Goldfeld said, “we hope to facilitate the integration of AI in human-facing services and unlocking the valuable societal benefits this entails.”
The research collaboration with the IBM-MIT Watson AI Lab draws on the strengths of the various groups. Goldfeld is focused on theoretical aspects of the project while IBM brings experts with broad practical experience and access to state-of-the-art computing resources. Goldfeld describes the collaboration as extremely fruitful.
The award will allow Goldfeld to fund a student or postdoc to join the research team.
The IBM Faculty Awards are aimed at fostering collaboration between university researchers and IBM research, development and services organizations to promote curriculum innovation and to stimulate growth in strategic disciplines. IBM strongly encourages work resulting from the awards to be placed in the public domain.