Imagine walking into a lab where robots are conducting careful experiments. They are assisting researchers in the discovery of metal alloys that can snap back into shape after crumbling, or that can snake through easily into human blood vessels.
This is not a scene from a Steven Spielberg movie or a chapter from an Isaac Asimov novel.
It is indeed going to be a reality at the US’s Texas A&M University. The Department of Materials Science and Engineering of the university has received a $2 million grant to build a first-of-its-kind self-driven research lab, named the Aggie AI Foundry.
Automated system test material
The facility will be fully automated, equipped with artificial intelligence (AI) brainpower and robotic workers. The computer-controlled arms of the robots are expected to test novel metal alloys by mixing, melting, and testing materials based on directions from an AI brain.
These are expected to take these laborious tasks away from humans, which still take a toll on them, and speed up the discovery and development of new metal alloys beyond current capabilities.
”Dr. Karaman and I, along with our collaborators and students, know what it takes to design materials faster. And we have shown it,” said Raymundo Arróyave, a professor in the department. “But the process can still take a toll on humans.”
The Foundry is a brainchild of Arróyave and Ibrahim Karaman, another materials scientist in the department. By building this lab, they hope to free the students for more creative work.
”You want students to be creative, but imagine they are like line cooks in a big, fancy restaurant. They want to be chefs, and we have them chopping onions and cutting tomatoes. We want to free human brains and hands to conduct more science and write papers that share these discoveries with the world.”
Trained AI brain directs robots
In the lab, AI training will first take place to enable it to propose ideas for transformative alloys. These training sessions will use physical models, algorithms, and human knowledge built through years of trial and error.
Based on the directions from the AI brain, the robotics will mix, melt, and test the materials for the desired qualities using computer-controlled robotic arms. These arms will function in the assembly line, taking samples from a wall cabinet of elements from the entire periodic table, a statement from the university read.
These automated processes are expected to increase the rate at which metal alloys are developed from months to a week or less.
Researchers noted that, with the current systems available, it wasn’t possible to test materials using only a computer. The work must be done on actual materials, they added.
”The computer can predict some behavior, but there’s no computational tool anywhere in the world that can predict how an alloy is going to behave,” Arróyave said.
But once this lab is up and running, AI can try a trillion possible ways of mixing several dozen elements from the periodic table, a feat yet not achievable for the human brain. While AI can discover hidden correlations between elements, make suggestions, and run the experiments quickly, that could fasten the pace of new alloy discoveries.
”While it’s natural to have questions about AI, it can help scientists explore possibilities and make discoveries at a scale we’ve never seen before,” Arróyave said.
The grant came from the WoodNext Foundation, which manages the philanthropy of tech innovator and Roku CEO and Founder Anthony Wood and his wife Susan.