Skip to content Skip to footer

Justin Solomon, an associate professor in the MIT Department of Electrical Engineering and Computer Science (EECS) and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL), is using advanced geometric techniques to deal with complex issues that don’t seemingly have any connection with geometry. Solomon explains that geometric terms like distance, similarity, and curvature are often used to describe data, making geometry a useful tool in data science.

Solomon’s Geometric Data Processing Group divides their work into two main sections; processing two and three-dimensional geometric data and high-dimensional statistical research. Examples of their work include aligning 3D organ scans in medical imaging and utilizing generative AI models. These models learn to create new images by sampling from parts of a dataset with example images, a process essentially geometric in nature.

From an early age, Solomon was interested in computer graphics, and high school and university internships further cemented his interest. Solomon’s ongoing work at Pixar throughout college and graduate school expanded his skills in the animation of cloth and fluids, as well as techniques to change the ‘look’ of animated content.

During his PhD at Stanford, Solomon focused on ‘optimal transport’, a concept that seeks to move a distribution of an item to another distribution as efficiently as possible. This research ultimately attracted Solomon to MIT, where he works on complex problems with far-reaching effects on many disciplines.

Solomon aims to open up the field of geometric research to those from underrepresented backgrounds. To establish this, he started the Summer Geometry Initiative, a six-week paid research program for undergraduates, focusing on candidates from underrepresented communities. His efforts have seen incoming PhD classes’ composition change, emphasizing the value of ensuring diversity within geometry-related research fields.

Solomon believes his methods can be particularly beneficial for tackling unsupervised machine learning, where models learn to recognize patterns without labeled training data. Because labeling 3D data can be costly, models that incorporate geometric insight can help computers decipher complex, unlabeled 3D scenes more efficiently.

Outside of academics, Solomon is also a talented musician and currently plays the cello with the New Philharmonia Orchestra. Solomon notes the similarity between music and his research field, computer graphics, both being analytical and closely linked to artistic practice.

Leave a comment

0.0/5