Have a UROP opening you would like to submit?
Please fill out the form.
Spectroscopic Line Detection and Classification Using Machine Learning
PSFC: Plasma Science & Fusion Center
Both in astrophysics and in fusion devices, electromagnetic radiation emitted by plasmas allows us to probe extreme physical environments. In particular, atomic spectra offer the means to gain insights into plasma dynamics and equilibria in cases where more direct measurements may be unfeasible. For this reason, increasing our confidence in atomic line recognition is clearly desirable. In this project, we propose to focus on the automatic detection of atomic lines from spectroscopic measurements in the ultraviolet range on the C-Mod tokamak at MIT. The UROP student will work with and expand on a set of Python routines to analyze previously collected data containing atomic signatures in real experimental settings. Comparison with theoretical atomic predictions will allow the development and validation of pattern recognition techniques. We expect that the project will have a computational emphasis, using computer vision and machine learning methods to discover highly-ionized states of atoms in the plasma. The project will make use of tokamak data, but may also be interesting to students interested in observational astrophysics, since similar data analysis challenges are faced in the two fields.
Some familiarity with Python is required. Some background in quantum mechanics and spectroscopy may be helpful. No prior knowledge of tokamak or plasma physics is expected.