by Anna Fiorentino
It may not surprise you to learn that David Kaeli, Northeastern University distinguished professor and faculty member at the Institute for Experiential AI, recently earned a life-long achievement award from the Association of Computing Machinery (ACM). Not when you consider that we experience the impact of his early work each time we start up our car and use our phone or computer — when a digital circuit performs what is known as “branch prediction”.
“When a computer makes a decision about something, that decision turns into a conditional branch,” says Kaeli, who is also the director of the master’s in data science program. “My contribution was to discover how to predict them much more accurately.”
In January, Kaeli joined the esteemed ranks at the ACM, the world’s largest scientific and educational computing society. The distinguished professor in the College of Engineering and director of the Computer Architecture Research Laboratory is now among just 1 percent of ACM members to hold the honor of fellow in computing and information technology or outstanding service to ACM and the larger computing community.
But, despite his achievements, it blindsided Kaeli to learn about the award last month from his nominator and former collaborator, a faculty member at the Universitat Politècnica de Catalunya named Antonio Gonzalez.
“I was very pleasantly surprised to have received it,” he says. “I recognize the importance of the award by others who have received it before me, and everyone who’s helped me get to this point in my career along the way.”
Kaeli’s research has leveraged state-of-the-art high-performance computing and accelerators with help from his 44 doctoral students over his 28 years at Northeastern. But it was while at IBM that he experienced the genesis of his career.
“IBM sent me to a school called Systems Research Institute in New York City with some of their top engineers to learn about future technologies,” says Kaeli. “It convinced me to go get my Doctor of Philosophy and that I should be working on research.”
He moved over to IBM’s prestigious T.J. Watson Research Center while working toward his doctoral degree at Rutgers University, having already earned his bachelor’s from Rutgers and master’s from Syracuse University.
Before developing that fundamental accurate prediction of return instructions now implemented in nearly every microprocessor in production today, Kaeli also designed novel memory self-testing algorithms, performance evaluation methods, and a shared-memory microprocessor system for IBM. His contributions to branch prediction, compiler optimization, security, and graphics processing unit computing have accelerated critical applications in machine learning and fields including biomedical imaging and homeland security.
He offers unique historical insight into progress within the field of computing. Over the years, Kaeli watched the rate of how many transistors can fit on an integrated circuit — known as the Moore’s Law — enables key innovations. While this projection in semiconductor manufacturing, which has put more devices onto a chip to increase the levels of integration and computing capabilities, has allowed Kaeli to use those circuits to design the next generation of architectures, he notes that the discipline has reached an inflection point.
“We can no longer increase the density of the devices, so there’s important work that needs to be done across the field, from programming to silicon devices,” says Kaeli. “I’ve seen these advances, but we’re now at a turning point, and I think we’re going to see some major changes coming with new device technologies.”
As this area continues to grow exponentially, he looks forward to another chapter of collaborations with fellow researchers at the Institute for Experiential AI, Northeastern, and beyond.
“I’m excited about the class of applications that we’re trying to tackle,” says Kaeli. “A lot of them are focused on trying to make the world better and bringing the technology to bear on that is really important.”
Those are the types of collaborations, he says, that allowed him to accelerate his career. Kaeli’s work has appeared in more than 375 critically reviewed publications, and he is the author of seven books and 11 patents. He now serves as editor-in-chief of the ACM Transactions on Architecture and Code Optimization journal and leads the data management and analysis core for the National Institute of Environmental Health Sciences’ Puerto Rico Testsite for Exploring Contamination Threats Center. He also holds a courtesy appointment in the College of Computer and Information Science and an honorary professorship from City, University of London.
This is the second fellowship Kaeli has received from a prestigious professional association — the first being from the Institute of Electrical and Electronics Engineers. Kaeli was one of 71 ACM fellows recently honored for wide-ranging and fundamental contributions in areas including algorithms, computer science education, mobile and networked systems, data security and privacy, cryptography, and medical informatics that have continued to change the shape of technology today.