For more than a decade, MIT Associate Professor Rafael Gómez-Bombarelli has used artificial intelligence to create new things. As technology has increased, so have his desires.
Now, a newly appointed professor of materials science and engineering believes AI is poised to revolutionize science in unprecedented ways. His work at MIT and beyond is dedicated to accelerating that future.
“We are at the second inflection point,” Gómez-Bombarelli says. “The first one was in 2015 with the first wave of learning representation, generative AI, and big data in other areas of science. Those are some techniques that I started to bring to my lab at MIT. Now I think we are in the second place of change, mixing language and combining many methods into common scientific intelligence. We need rules, to balance all the reasons of language and to balance about language. structures, and to discuss about recipes of integration.”
Gómez Bombarelli’s research combines physics-based simulations with methods such as machine learning and generative AI to discover new materials with promising real-world applications. His work led to new materials for batteries, catalysts, plastics, and organic light-emitting diodes (OLED). He has also founded several companies and served on scientific advisory boards to pioneer the use of AI in drug discovery, robotics, and more. His latest company, Lila Sciences, works to create a scientific intelligence platform for the life sciences, chemical, and materials science industries.
All that work is designed to ensure that the future of scientific research is smoother and more productive than today’s research.
“Scientific AI is one of the most exciting and ambitious applications of AI,” said Gómez-Bombarelli. “Some applications of AI have a lot of negativity and ambiguity. Scientific AI is about bringing about a better future over time.”
From testing to simulation
Gómez-Bombarelli grew up in Spain and focused on natural sciences from an early age. In 2001, he won the Chemistry Olympics competition, placing him on the chemistry course list, which he studied as an undergraduate at his hometown college, the University of Salamanca. Gómez-Bombarelli stuck around for his PhD, where he investigated the activity of chemicals that damage DNA.
“My PhD started as an experiment, and I was bitten by the simulation and computer science bug about halfway through,” he says. “I started simulating the same chemical reactions I was measuring in the lab. I like the way programming programs your brain; it felt like a natural way to program human thinking. Programming is also very limited to what you can do with your hands or scientific instruments.”
Next, Gómez-Bombarelli went to Scotland for a postdoctoral position, where he studied quantum effects in biology. Through that work, he connected with Alán Aspuru-Guzik, a professor of chemistry at Harvard University, whom he joined for his next postdoc in 2014.
“I was one of the first people to use artificial AI in chemistry in 2016, and I was in the first group to use neural networks to understand molecules in 2015,” said Gómez-Bombarelli. “It was the early, early days of serious scientific study.”
Gómez-Bombarelli also began working to remove the manual parts of molecular simulations in order to perform high-throughput experiments. He and his collaborators ended up running hundreds of thousands of calculations across the board, finding hundreds of promising test items.
After two years in the lab, Gómez-Bombarelli and Aspuru-Guzik started a general-purpose computing company, which eventually focused on producing light-emitting biodies. Gómez-Bombarelli joined the company full-time and calls it the hardest thing she’s ever done in her career.
“It was amazing to do something tangible,” he says. “Also, after seeing Aspuru-Guzik run the lab, I didn’t want to be a professor. My father was a professor of languages, and I thought it was a menial job. Then I saw Aspuru-Guzik had a team of 40 people, and he was on the road 120 days a year. It was crazy.
In 2018, Aspuru-Guzik suggested that Gómez-Bombarelli apply for a new position at MIT’s Department of Materials Science and Engineering. But, in his trepidation about professional work, Gómez-Bombarelli let the deadline pass. Aspuru-Guzik met him in his office, slammed his hands on the table, and told him, “You need to apply for this.” It was enough to get Gómez-Bombarelli to put together a formal request.
Fortunately in his early days, Gómez-Bombarelli had spent a lot of time thinking about how to create value from the acquisition of building materials. During the interview, he says, he was drawn to the energy and spirit of collaboration at MIT. He also began to appreciate research opportunities.
“Everything I was doing as a postdoc and in the company would be part of what I could do at MIT,” he said. “I was making products, and I still do. All of a sudden, my whole work environment was part of this new area of things that I could explore and do.”
It has been nine years since Gómez Bombarelli joined MIT. Today his lab focuses on how the structure, composition, and reactivity of atoms affect the properties of materials. He also used advanced simulation to create new objects and helped develop tools to combine deep learning with physics-based modeling.
“Physics-based simulations make data and AI algorithms better when you give them data,” says Gómez Bombarelli’s. “There are all kinds of virtuous cycles between AI and simulation.”
The research team he built is computer-only – no physical testing.
“It is a blessing because we can have a large scope and do many things at the same time,” he said. “We love working with testers and try to be good partners with them. We also love creating computational tools that help testers test ideas from AI.”
Gómez-Bombarelli also remains focused on real-world applications of his inventions. His lab works closely with companies and organizations such as MIT’s Industrial Liaison Program to understand the real needs of the private sector and the practical barriers to commercial development.
Accelerating science
As excitement surrounding artificial intelligence has exploded, Gómez-Bombarelli has seen the field mature. Companies like Meta, Microsoft, and Google’s DeepMind are now regularly doing physics-based simulations reminiscent of what he was working on back in 2016. In November, the US Department of Energy launched the Genesis Mission to accelerate scientific discovery, national security, and energy governance using AI.
“Simulation AI has gone from something that might work to a scientific theory of consensus,” Gómez-Bombarelli said. “We’re at a turning point. People think in natural language, we write papers in natural language, and it turns out that these kinds of big languages that have learned natural language have opened up the ability to accelerate science. We’ve seen that scaling works for simulation. We’ve seen that scaling works for language. Now we’re going to see how scaling works for science.”
When he first came to MIT, Gómez-Bombarelli says he was blown away by how non-competitive things are between researchers. He tries to bring that same positive-sum thinking to his research group, which is made up of about 25 graduate students and postdocs.
“We’ve naturally grown into a really diverse group, with a diverse mindset,” Gomez-Bombarelli said. “Everyone has their own career aspirations and strengths and weaknesses. Figuring out how to help people become the best versions of themselves is exciting. Now I’m the one who insists that people apply for professional positions after the deadline. I think I passed that stick.”
