Virtual Co-Design Research Center Works To Improve AI Technologies

Sandia National Laboratories, Pacific Northwest National Laboratory, and the Georgia Institute of Technology are partnering to launch a virtual research center that will combine hardware design and software development to improve artificial intelligence technologies. The program – Artificial Intelligence-focused Architectures and Algorithms – will receive $5.5 million in funding over the next three years from The Department of Energy’s Office of Advanced Scientific Computing Research.

“The center will focus on the most challenging basic problems facing the young field, with the intention of speeding advances in cybersecurity, electric grid resilience, physics and chemistry simulations and other DOE priorities,” said Sandia project lead Siva Rajamanickam, an expert in high-performance computing. “A co-design center is a wonderful opportunity because people of diverse backgrounds — hardware designers, theoretical computer scientists, mathematicians and domain scientists — come together to develop solutions to a very challenging problem, the co-design of machine learning accelerators.”

The collaborative environment is intended to encourage researchers at the three locations to simulate and evaluate artificial intelligence hardware when employed on current or future supercomputers. It is expected that the researchers will be able to improve AI and machine-learning methods as they replace or augment more traditional computation methods.

Researchers at the three locations will bring their specialties to the project:

  • Sandia – with a strong background in high-performance computing, the lab will develop methods to use emerging machine-learning devices effectively and provide access to computer facilities and test beds to AI researchers. A focus of the center will be on sparse computations – a type of computation that utilizes the principle that in real life there might be many interactions but only a few that may affect the outcome to a problem;
  • PNNL – has expertise in simulations related to power grids, chemistry, and cybersecurity; and,
  • Georgia Tech – has experience in developing custom hardware accelerators for machine learning. It will focus on using this hardware for sparse linear algebra.