For Connor Coley, figuring out why things work the way they do is only half of science's allure. The other half, he says, is figuring out how he can make them work better. His drive and curiosity led him to work on problems in structural biology, renewable processes for converting carbon dioxide to methanol, and fracking—all before he could legally buy beer. "I couldn't figure out what I wasn't interested in," he jokes.

These days Coley's interests still vary—so much so that the Massachusetts Institute of Technology graduate student divides his time between William H. Green's molecular simulation lab and Klavs F. Jensen's drug discovery lab. In just four years, he's spearheaded a project in the white-hot field of machine learning, which Coley hopes to apply to chemical synthesis for drug discovery and development.

Machine learning aims to create artificial intelligence systems that make decisions with little intervention from people. Coley's efforts in this arena have blossomed into a collaboration between MIT and eight drug industry partners, known as the Machine Learning for Pharmaceutical Discovery & Synthesis Consortium. While most other chemists working in the field of machine learning and chemical synthesis use rules devised by experts to guide their systems, Coley relies on reactions in databases, such as those in U.S. patent filings, to teach the computer what transformations will and won't take place without being influenced by human bias.

Before Coley had even finished his second year of graduate school, he was using machine learning to predict the major products of many different types of chemical reactions, with the computer accurately predicting the correct reaction product 72% of the time. "Many smart chemists had worked on and thought about this problem for decades," Green says. "Several other groups had been trying to solve it using machine learning for years before Connor even entered graduate school. But Connor almost immediately saw the way forward and implemented it."

"Connor is the main force behind MIT's efforts in the area of machine learning and chemistry," adds MIT computer science professor Regina Barzilay, who is involved with the consortium. "He deeply understands both fields and comes up with creative ideas in their intersection."