Justin Solomon, an associate professor in the MIT Department of Electrical Engineering and Computer Science and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL), employs modern geometric techniques to solve intricate problems often unrelated to shapes. Using these geometric methods, data sets can be compared and the high-dimensional space in which the data is arranged can be understood. This approach helps determine whether a machine-learning model will operate effectively on different data sets.
Solomon and his Geometric Data Processing Group utilize geometric techniques to solve a broad range of issues. About half the team works on 2D and 3D geometric data problems, such as aligning 3D organ scans and enabling autonomous vehicles to recognize pedestrians using data from LiDAR sensors. The other half uses geometric tools to conduct high-dimensional statistical research, which aids in building improved generative AI models.
Solomon’s interest in computer graphics from an early age inspired him to study both mathematics and computer science at Stanford University, helping him develop algorithms for 3D face recognition. A summer internship at Pixar Animation Studios allowed him to improve the realism of animated films and change the look of animated content.
While pursuing a computer science PhD at Stanford, Solomon studied optimal transport, a problem involving the movement of a distribution of an item to another distribution as efficiently as possible. He was drawn to MIT for the chance to collaborate on complex, practical problems that could impact many fields.
Outside his research endeavors, Solomon strives to make geometric research accessible to underserved students who typically lack research opportunities in high school or college. The Summer Geometry Initiative, a program he launched, serves this very purpose. This six-week research program mainly involves undergraduates from underrepresented backgrounds. The program completed its third year in 2023.
Solomon believes in the importance of diversity in the field of geometric research as he continues to apply geometric techniques to a growing list of problems in machine learning and statistics. He looks forward to using geometry to improve unsupervised machine learning models by helping computers decipher complex, unlabeled 3D scenes.
Apart from academic pursuits, Solomon is a classical music enthusiast, with a keen interest in composer Dmitri Shostakovich. His love for music connects with his research field, computer graphics, both of which he finds analytical and closely related to artistic practice.
