The Georgia Institute of Technology, Sandia National Laboratories, and Pacific Northwest National Laboratory have jointly launched a new research center to address some of the most challenging problems in today's artificial intelligence (AI), thanks to the US Department of Defense. Provided $5.5 million in funding. Energy (DoE).

Artificial Intelligence Research and Cooperation at Georgia Institute of Technology.
Artificial Intelligence Research and Cooperation at Georgia Institute of Technology.


AI enables computer systems to learn from experience without explicit programming automatically. This technique can perform tasks that were only humans can perform: see, identify patterns, make decisions, and take action.

The new Common Design Center, funded by the US Department of Energy's Science Office, the Artificial Intelligence-Centered Architecture and Algorithm Center (ARIAA), will promote collaboration between scientists in the three organizations as they develop AI applications. Critical core technologies to DoE mission priorities such as network security, grid resiliency, graphical analysis, and scientific simulation.

PNNL senior research scientist Roberto Gioiosa will serve as the director of the center and will lead the overall vision, strategy, and research direction. Tushar Krishna, assistant professor of the School of Electrical and Computer Engineering (ECE) at Georgia Institute of Technology, and Siva Rajamanickam, a computer scientist at Sandia, will serve as deputy director.

Each organization brings unique advantages to collaboration: PNNL has expertise in power grid simulation, chemistry, and network security, and computer architecture and programming models, as well as computing resources such as systems for testing emerging architectures. Have research experience. Sandia has expertise in computer systems for software simulation, machine learning algorithms, graphical analysis, and sparse linear algebra.

Computer facilities and test platform systems to support early access to emerging computing architectures for code development and test.

Sandia's Rajamanickam said: "In the past few years, strategic cooperation between Sandia and Georgia Institute of Technology has done possible new projects like the Center."

Dedicated hardware allows AI tasks to run faster and consume less power than traditional computing devices such as CPUs and GPUs. ARIAA revolves around a concept called "collaborative design," which refers to the need for researchers to weigh and balance the power of hardware and software and to summarize the architectures and algorithms that might best solve the current problem.

Krishna's lab will lead the design and evaluation of reconfigurable hardware accelerators to accommodate the rapidly evolving needs of AI applications. In particular, the effective running of sparse computing will be the focus. Sparse calculations are critical to multiple computing areas of interest to DoE because they significantly reduce the number of predictions for massive data problems. One way to consider sparsity is that there may be millions or even billions of users on social media sites, but users only care about updates from hundreds of friends.

Post a Comment

Previous Post Next Post