Vibha Raghvendran

How can a machine learning model be trained to predict the characteristics of a planetary system from a host star's light curve, and how can its predictions be validated using the physics-based Mandel-Agol model?

Started May 12, 2025

Abstract or project description

I will develop a machine learning model to predict key characteristics of an exoplanetary system from a star's light curve: the transit midpoint, the orbital period, the scaled semi-major axis, the impact parameter, and the planet-to-star radius ratio. The model will be trained on the flux time series of the host star to infer these parameters. To validate and refine the model's predictions, I will use the Mandel–Agol physics-based transit model to generate a synthetic flux curve from the predicted parameters. This synthetic curve will then be compared to the original observed flux time series, and the resulting discrepancy will be used as a loss function to iteratively improve the machine learning model's performance.