“There are higher uses for a PhD student than waiting around in a lab until 3am to be sure an experiment is run to the tip,” says Ant Rowstron, ARIA’s chief technology officer.
ARIA picked 12 projects to fund from the 245 proposals, doubling the quantity of funding it had intended to allocate due to the large number and top quality of submissions. Half the teams are from the UK; the remainder are from the US and Europe. A few of the teams are from universities, some from industry. Each will get around £500,000 (around $675,000) to cover 9 months’ work. At the tip of that point, they need to have the ability to display that their AI scientist was in a position to provide you with novel findings.
Winning teams include Lila Sciences, a US company that’s constructing what it calls an AI NanoScientist, a system that may design and run experiments to find the very best ways to compose and process quantum dots, that are nanometer-scale semiconductor particles utilized in medical imaging, solar panels and QLED TVs.
“We’re using the funds and time to prove some extent,” says Rafa Gómez-Bombarelli at Lila Sciences: “The grant lets us design an actual AI robotics loop around a focused scientific problem, generate evidence that it really works, and document the playbook so others can reproduce and extend it.”
One other team, from the University of Liverpool, UK, is constructing a robot chemist, which runs multiple experiments without delay and uses a vision language model to assist troubleshoot when the robot makes an error.
And Humanis AI, a startup based in London, is developing an AI scientist called ThetaWorld, which is using LLMs to design experiments to check the physical and chemical interactions which might be vital for the performance of batteries. The experiments will then be run in an automatic lab by Sandia National Laboratories within the US.
Taking the temperature
In comparison with the £5 million projects spanning 2-3 years that ARIA often funds, £500,000 is small change. But that was the thought, says Rowstron: It’s an experiment on ARIA’s part too. By funding a spread of projects for a brief period of time, the agency is taking the temperature on the leading edge to find out how the best way science is finished is changing, and how briskly. What it learns will grow to be the baseline for funding future large-scale projects.
Rowstron acknowledges there’s plenty of hype, especially now that the majority of the highest AI firms have teams focused on science. When results are shared by press release and never peer review, it may possibly be hard to know what the technology can and might’t do. “That’s all the time a challenge for a research agency attempting to fund the frontier,” he says. “To do things on the frontier we have to know what the frontier is.”
