Revolutionizing Data Science with Geometric Techniques
Justin Solomon, an associate professor at MIT’s Department of Electrical Engineering and Computer Science (EECS) and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL), is pioneering the use of geometric methods to solve complex problems in data science, machine learning, and beyond.
Geometry’s Role in Modern Data Science
Solomon’s work illustrates how ancient geometric principles, established by Euclid over 2,000 years ago, are now integral to contemporary data science. He leverages geometric tools to compare datasets, revealing underlying structures that affect machine-learning model performance, a critical aspect in fields like statistics and AI.
Geometric Data Processing Group: A Diverse Research Spectrum
Leading the Geometric Data Processing Group, Solomon and his team tackle a wide array of problems, ranging from processing 2D and 3D geometric data for medical imaging and autonomous vehicles to high-dimensional statistical research for developing generative AI models.
Graphics and Optimal Transport: Foundations of Solomon’s Research
Solomon’s journey into this field began with a passion for computer graphics. His exploration of optimal transport problems in computer graphics gradually expanded to other applications, setting the foundation for his diverse research group at MIT.
MIT: A Hub for Interdisciplinary Research
Attracted to MIT for its potential for impactful, interdisciplinary research, Solomon collaborates with brilliant students and colleagues, solving practical problems across various disciplines.
Fostering Diversity in Geometric Research
Committed to making geometric research accessible to underserved students, Solomon initiated the Summer Geometry Initiative, a program to introduce undergraduates, predominantly from underrepresented backgrounds, to geometry research. This initiative aims to diversify the field and bring in new perspectives.
Applying Geometry to Unsupervised Machine Learning
Solomon’s future endeavors include applying geometric tools to enhance unsupervised machine learning models, especially in the context of 3D data processing and inference.
Music: Solomon’s Artistic Escape
Apart from his academic pursuits, Solomon is an avid musician, playing classical music on the piano and cello. He is currently a member of the New Philharmonia Orchestra in Newton, Massachusetts, finding a harmonious balance between his analytical and artistic interests.
Impact on the Field of Machine Learning
Solomon’s work exemplifies the crucial role of geometric principles in modern data science and machine learning, underscoring the need for a diverse array of researchers to tackle the ever-growing challenges in these fields.
References:
https://news.mit.edu/2023/justin-solomon-pushing-geometric-boundaries-1212
