Brian Spears says his children will enjoy a more sustainable planet, thanks in part to artificial intelligence and high-performance computing (HPC) simulations.
“I believe I will see fusion power in my lifetime, and I’m sure my daughters will see a world based on fusion,” said the 45-year-old principal investigator at Lawrence Livermore National Laboratory, who helped demonstrate to physics a clean and abundant source of energy. , which makes headlines around the world.
The results of the experiment arrived in Spears’ inbox at 5:30 a.m. on December 5 last year.
“I had to rub my eyes to make sure I wasn’t reading the numbers wrong,” he recalled.
After making sure, he ran downstairs to share the news with his wife, a chemical engineer at a lab that pioneered 3D glass printing and once worked on the fusion program.
“One of my friends described us as star trek household – I work on the warp core and she works on the replicator, he joked.
After the lab officially announced the news, Spears shared her excitement with the world in a wild tweet.
“Tired from an amazing day… Daughters send me screenshots of mommy and daddy’s latest work updates… Being a part of something amazing for humanity.”
in another tweet– he shared the technical details.
“Two million joules of laser energy were used to crush the capsule 100 times smoother than a mirror. It spread by half the thickness of a hair. In 100 trillionths of a second, we have produced ten petawatts of energy. It was the brightest thing in the solar system.”
AI helps to decide
A week before the experiment, Spears’ team analyzed his exact HPC design and then predicted the outcome with AI. Two atoms fuse into one, releasing energy in a process simply called ignition.
It was the most exciting of the thousands of AI predictions in what has become the two-step dance of modern science. Teams design experiments in HPC simulations and then use data from actual results to train AI models that refine the next simulation.
AI reveals details of experiments that are difficult for humans to see. For example, researchers tracked the effects of small imperfections in an exploding capsule and blasted it with 192 lasers to achieve fusion.
“You need artificial intelligence to understand the whole picture,” Spears said.
It is a large canvas filled with mathematics that describes the intricate details of atomic physics.
A single experiment may require hundreds of thousands of relatively small simulations. Each takes half a day on one supercomputer node.
The largest 3D simulations, called kitchen sinks, take up about half of Sierra, the sixth-fastest HPC system in the world, which contains 17,280 NVIDIA GPUs.
Edge AI Guides Experiments
AI is also helping scientists conduct experiments with autonomous driving. Neural networks can make split-second decisions about which way to run an experiment based on the results they process in real time.
Last year, for example, Spears, his colleagues, and NVIDIA collaborated on an AI-driven experiment that fired lasers up to three times per second. He created the type of proton beams that could one day treat cancer patients.
“Within a day, you can get a beam this bright that may have taken you months or years of human experiments to develop,” Spears said. “This approach of artificial intelligence at the edge will save orders of time for our experts.”
Controlling lasers that are fired many times per second will also be a key job in tomorrow’s fusion reactors.
Navigation in the flow of data
The impact of artificial intelligence will be widely felt in both scientific and industrial spheres, Spears believes.
“In the last decade, we have generated more simulation and experimental data than we are trained to deal with,” he said.
This deluge, once a burden for scientists, is now fuel for machine learning.
“AI is putting scientists back in the driver’s seat so we can move much faster,” he said.
Spears also directs AI initiative in a lab that depends on collaboration with companies including NVIDIA.
“NVIDIA is helping us look beyond the horizon so we can take the next step in using artificial intelligence for science,” he said.
It’s hard work with huge consequences, like leaving a more sustainable planet for the next generation.
When asked if his two daughters are planning careers in science, Spears is happy. They are both competitive swimmers, play the jazz trumpet, and are interested in everything from bioengineering to art.
“As we say in science, they are four pi, they cover the whole sky,” he said.