We’d all love to soar like an eagle on occasion. Gautam Reddy figured out the math that makes it possible.
Soaring birds ‘cheat’ with some frequency. They find thermals – rising columns of air – which they can use to their advantage. By hopping from thermal to thermal, soaring birds can make the air work for them, pushing them high into the air without any flapping required.
Reddy spent his graduate career at UC San Diego working chiefly on this thermal soaring behavior of birds. He determined which variables can be used to efficiently find and harness thermals. Using these insights, he was able to engineer machine learning strategies that allowed gliders to train themselves to soar without supervision.
"In general, I use tools from reinforcement learning to understand how animals perform tasks critical for their survival,” says Reddy. “How do birds soar, how do insects navigate, how do fish forage? What information do they extract from an unpredictable environment and what algorithms do they use to make decisions based on this information?”
Now with support from the National Science Foundation and the Simons Foundation, Reddy is uncovering strategies that animals use across species. As an independently funded postdoctoral fellow at Harvard’s NSF-Simons Center, Reddy is able to work with any faculty member affiliated with the Center. This gives him access to faculty from across all of science, engineering, and math at Harvard.
“Harvard offers a fantastic environment for precisely this kind of research,” says Reddy. “There are so many labs working on diverse sets of organisms, and the Center provides the freedom and resources to start a broad set of exciting collaborations across its departments and other institutions."
“Gautam is already a fan favorite here,” says BN Queenan, Executive Director of Research at Harvard’s NSF-Simons Center for Quantitative Biology. “He has already kicked off all these amazing projects, but he goes about it in such a thoughtful, responsible way you really trust something great is going to happen. He’s quietly determined to solve all of biology.”
“I'm excited about coming up with mathematical models of biological phenomena we observe in our day-to-day lives,” says Reddy. So are we, Gautam!