Science 101: Autonomous Discovery

Have you ever pulled an “all-nighter” trying to finish a big project or homework assignment? By dawn, your brain was probably fried. But robots don’t get tired. Artificial intelligence doesn’t need coffee. They can keep working on a problem 24/7.

That’s why researchers at the U.S. Department of Energy’s Argonne National Laboratory are pioneering the next generation of scientific experimentation. It’s called autonomous discovery, and it harnesses the power of robotics, machine learning and artificial intelligence (AI) to solve big problems in energy, human health and materials faster than ever before. No robot coffee required.

Let’s say you need a new type of plastic that is easier to reuse. There are thousands of possible polymers that might work, and there are potentially harmful chemicals involved in the testing process. In the past, we might have had several researchers working at a lab bench to test one polymer at a time. And that process would take years.

With autonomous discovery, human researchers turn the problem over to robotics and AI. The researchers use a process called machine learning to train the AI with huge data sets. This gives the AI information on everything we already know about chemical processes and potential polymer structures. The AI uses that information to make the best decision about which experiments to run next.

Robotic arms and sampling machines run 24 hours a day, keeping researchers safely away from chemicals or other potential risks. With machine learning, AI gets smarter as it goes, looking for new patterns and new experiments to run.

Human scientists are free to use their big, creative brains to interpret and act on the most interesting results. Autonomous discovery will help scientists find solutions to complex problems in a matter of days or weeks instead of years or lifetimes.

Human ingenuity and imagination remain at the heart of this innovation. Argonne’s scientists and technicians are finding creative ways to apply the tools of autonomous discovery:

We’re creating “digital twins” of labs. These virtual-reality models help researchers test machines and figure out the best way to install equipment in real-life biology and chemistry labs.
We’re exploring everything from new treatments for drug-resistant bacteria to discovery of new materials for energy storage.
We’re using some of the world’s fastest supercomputers to run the AI and process data coming from large-scale experiments.

Autonomous labs are helping bring scientific solutions into our lives with never-before-seen speed, efficiency and accuracy. This will not just revolutionize the way we do science; it will propel us into a new era of discovery.