Prestigious Sloan Research Fellowships have been awarded to and , both assistant professors in the ��Ƶ of Computer Science.

The annual fellowships are given to early career researchers in Canada and the United States who the Alfred P. Sloan Foundation recognizes as individuals “whose creativity, innovation and research accomplishments make them stand out as the next generation of leaders.”

Ba received his PhD from the University of Toronto, where he was supervised by University Professor Emeritus . His research has had a major impact in the field of deep learning and focuses on the development of efficient learning algorithms for deep neural networks. These include the Adam Optimizer, a frequently used algorithm for training deep learning models; Lookahead, an algorithm that improves upon the generalization and stability of Adam and other ‘fast optimizers’; and Follow-the-Ridge, an optimization method for the situation of minimax optimization, in which multiple networks are simultaneously trained on different objectives.

Ba has also addressed the computational cost of training ensembles of neural networks, contributed to fundamental reinforcement learning algorithms, and advanced computer scientists’ theoretical understanding of deep neural networks.

“Through his innovative and wide-ranging research, Jimmy addresses some of the most challenging and important problems in training deep neural networks,” says , chair of the ��Ƶ of Computer Science. “His significant contributions have advanced our field by making deep learning more efficient and reliable. He is highly deserving of this honour.”

“I am honoured to receive the Sloan Fellowship in Computer Science this year,” says Ba, who is also a faculty member at the Vector Institute. “It would not be possible without the generous support of my colleagues and students, to whom I owe a debt of gratitude. I am very excited about the journey ahead!”

Sachdeva, an assistant professor in the ��Ƶ of Mathematical and Computational Sciences at the University of Toronto Mississauga and the tri-campus graduate ��Ƶ of Computer Science, received his PhD from Princeton University. As a theoretical computer scientist, he has made significant contributions to addressing the longstanding problem of maximum flow, a measure of how much material can flow through a network from a source to a destination when accounting for the limited capacity of its various parts. Common applications include the planning and optimization of telecommunications and transportation networks.

Previous work yielded many algorithms that describe how to move goods across a network; however, they were never particularly efficient. If the size of the network doubled, for example, the time needed to run the algorithm more than doubled.

Last year, Sachdeva, working with five other colleagues in Canada, the U.S., and Europe, found a vastly improved algorithm — one that the researchers say runs in “almost linear” time, meaning that the running time to find the solution grows in proportion to the size of the network being studied.

In a recent on the research in Quanta magazine, one researcher describes Sachdeva’s algorithm as “absurdly fast;” another called his team’s work a “tour de force.” Sachdeva himself describes it as a “mathematical breakthrough” and a “milestone.” And yet, the payoff may not be immediate, because the various previously-known solutions are good enough for many purposes. Nonetheless, he expects his team’s work will eventually lead to new software that may see widespread use.

“Sushant’s breakthroughs on the problem of maximum flow have garnered much-deserved recognition in the field of theoretical computer science and opened the door to new applications and lines of inquiry,” says de Lara. “This fellowship is a fitting tribute to his accomplishments.”

With files from .

������Ƶ

��Ƶ