After winning $100,000 in funding per team in June, six UCLA students are creating new technology for self-driving vehicles, artificial intelligence and microchip design.
Qualcomm, a communications and computing company, gives the award through a fellowship to 16 two-person teams annually based on project proposals they submit. All the UCLA recipients this year are doctoral students at the UCLA Samueli School of Engineering.
Qualcomm received 392 applications from scholars in a variety of research areas, including AI, autonomous driving and hardware and theory.
Prospective fellows submitted a one-page abstract and a three-page proposal and gave a final PowerPoint presentation to apply for the fellowship, said Seth Zhao, one of the UCLA students to receive the award.
“It’s a really long process,” Zhao said. “It’s also really challenging, and I think many other great Ph.D. students have great proposals, but they finally selected us for the project.”
Sujoy Ghosh, a recipient of the award, said his team’s research aims to reduce the size and increase the efficiency of chip components, which power almost all electronic devices and are used in everything from smartphones to data centers.
Equipment offered at UCLA helped his team succeed, Ghosh said. The research produced by fellows will be accessible to the public, he added.
“We definitely had a team which was able to advise us on how to proceed with this project, and on top of that, we had the industry partnerships, their software, their technical support, their advising and our professor,” Ghosh said.
Research funding at higher education institutions primarily comes from federal grants, but industry funding still contributes more than $6 billion a year – representing about 6% of total academic research funding, according to the National Center for Science and Engineering Statistics.
Zhao and Zewei Zhou have also used the fellowship funding to research new technologies to improve autonomous vehicle safety – another area of interest in the Qualcomm fellowship.
Current consumer autonomous driving systems are not advanced enough to tackle the diverse challenges that arrive on the road, and advanced AI can help them to respond to such complex situations, Zhou said. He added that these advanced models are currently too slow to respond quickly enough for the constantly changing conditions and that his group hopes to speed up these models for use in cars.
Samyak Chakrabarty, an award recipient, said in an emailed statement that his team aims to increase efficiency in AI by redesigning the chips used for computations on hardware in cars and robots.
Currently, many AI models are impractical without an internet connection to faraway data centers containing powerful processors, but Chakrabarty said his work could allow them to operate independently in technology such as cars, robots and satellites.
“One of the biggest bottlenecks in AI hardware is the energy and time spent moving data between memory and processors,” Chakrabarty said in the statement. “Current solutions throw huge amounts of resources at the problem, but that only gets you so far.”
Comments are closed.