Justin Solomon, an Associate Professor in the Massachusetts Institute of Technology (MIT)’s Department of Electrical Engineering and Computer Science (EECS) and member of the Computer Science and Artificial Intelligence Laboratory (CSAIL), is leveraging geometric techniques to tackle complex problems in data science. Quite often, these problems are seemingly unrelated to shapes. For example, when a statistician wants to compare two datasets for predicting the potential performance of a machine-learning model, Solomon explains that these datasets could share a geometric structure based on how the data are ordered in high-dimensional space. Using geometric tools to compare them could provide insights into whether the same model could be applicable to both datasets.
Solomon leads the Geometric Data Processing Group, which focuses on applying geometric techniques to a wide range of problems. Half of his team deals with problems that involve processing 2D and 3D geometric data such as aligning 3D organ scans in medical imaging and allowing autonomous vehicles to identify pedestrians using spatial data gathered by LiDAR sensors.
The other half of Solomon’s team carries out high-dimensional statistical research using geometric tools, like developing more efficient generative AI models. For instance, they learn to create new images by sampling specific parts of a dataset filled with example images – a fundamentally geometric problem.
Solomon’s interest in computer graphics started when he was a high-school student. He interned at a research lab where he developed algorithms for 3D face recognition, and then double-majored in math and computer science at Stanford University. At Stanford, Solomon also worked on physical simulation of cloth and fluids to enhance the realism of animated films.
Later, as a graduate student, he focused on a problem known as optimal transport, seeking the most efficient way to move a distribution of some item to another distribution. Initially, he applied optimal transport to computer graphics, but eventually, it found use in other applications which influenced the structure of his research group at MIT.
Solomon aims to make geometric research accessible to people who usually have limited exposure to it, specifically underserved students. He launched the Summer Geometry Initiative, a paid research program for undergraduates usually from underrepresented backgrounds that provides a hands-on introduction to geometry research. The program, wrapped up its third summer in 2023.
Solomon hopes this initiative will help diversify the field of geometric research, providing more perspectives in dealing with problems in machine learning and statistics using geometric techniques. He is particularly interested in improving unsupervised machine learning models with geometric tools.
Beyond his research, Solomon is also a classical music enthusiast. He often plays the piano or cello, occasionally joining a symphony in whatever city he lives in. He believes music shares an analytical nature with his research field, computer graphics, creating a mutually beneficial relationship between them.
