Graph Theory and Optimal Transport for Data Science
Directed by James Murphy
This project will focus on data-driven applications of graph theory and optimal transport to problems in anomaly detection and machine learning in image and network data. We will focus on designing new algorithms that enjoy robust performance guarantees and efficient complexity. We will also work on specific data challenges sponsored by the National Geospatial Agency and collaborators at Tufts Medical Center. Students can expect to learn and develop new mathematics, create new coding tools based on these ideas, and apply them to cutting-edge real-world problems. This project is highly interdisciplinary and is ideal for students with interests in applied math, data science, and machine learning.
Candidates should have at least taken Calculus III, Linear Algebra and have some working knowledge of statistics and basic coding in Python. Ideal candidates will have taken courses in probability theory and real analysis and be strong programmers.
About the Professor
I am an assistant professor of mathematics at Tufts University with adjunct appointments in computer science and electrical & computer engineering. I work on problems in applied harmonic analysis, mathematics of data science, statistical & machine learning, high dimensional probability theory, and graph theory. In addition to mathematics, I work on data science problems in image & signal processing, molecular dynamics, and computational biology. I enjoy working with junior researchers at all levels and look forward to welcoming REU students to my team.