University of Utah electrical and computer engineering assistant professors Weilu Gao and Cunxi Yu were recently published in the Laser and Photonics Review Journal for a collaborative paper they wrote on end-to-end computing system design. The paper investigates the use of machine learning and AI in the automation of system design from scratch.
“This work deals with end-to-end design from scratch,” says Yu, “so we are talking about from materials all the way to an actual computing system. There is a lot of processing involved, which is a really complicated task, but we are using machine learning or AI to do everything automatically.”
The team is using reinforcement learning, one of the most popular techniques in AI, to help the AI agent learn by itself how to build a system from scratch. They hope that by demonstrating the potential of carefully designed AI able to achieve this type of design and build automation, that the technique can be utilized to simplify engineering efforts across the industry.
“When you design the AI carefully, it should be able to carry out the design tasks of what would be comparable to thousands of engineers in significantly less time,” explains Yu. “Additionally, there is currently a large barrier of entry to engineering in terms of the background and technical knowledge needed for any role; this technique could help lower that barrier to allow for more people to be involved.”
“We are still in the framework stages and anticipate carrying out experimental verification in the near future to prove this concept and its potential,” says Gao.
To read Gao and Yu’s published paper on this work, click here.
Learn more about Electrical and Computer Engineering faculty research topics and discover ways to get involved in undergraduate research.