Texas ECE PhD student Jiaqi Gu won the First Place at the Association for Computing Machinery (ACM) Student Research Competition Grand Finals (Graduate Category). His research competition title is "Light in Artificial Intelligence: Efficient Neurocomputing with Optical Neural Networks." The award will be presented at the annual ACM Awards Banquet.
In this project, researchers propose a holistic solution to enable efficient neuromorphic computing with optical neural networks (ONN). They explore various photonic circuit designs and software-hardware co-optimization methods to enable high-performance photonic accelerators with lower area cost, better energy efficiency, and higher variation-robustness. Their ONN project includes a compact and power-efficient fast-Fourier-transform-based ONN architecture FFT-ONN (ASP-DAC’20 Best Paper Award, TCAD’20), a noise-aware robustness training scheme ROQ (DATE’20), and a scalable and efficient ONN on-chip learning framework FLOPS (DAC’20 Best Paper Candidate). Their project also has a photonic neural chip tape-out under measurement.
The ACM-wide Student Research Competition Grand Finals brings together first place winners from all ACM Special Interest Groups.
This is the second time a Texas ECE PhD student has won First Place in the ACM Student Research Competition Grand Finals (Graduate Category). Meng Li took first place in 2018.