New Rotation Period Measurements for Main-Sequence Stars Using Deep Learning

We propose a new framework to predict stellar properties from light curves using deep learning model. We analyze the light-curve data from the Kepler space mission and develop a novel tool for deriving the stellar rotation periods for main-sequence stars. Using this tool, we provide rotation periods for ~85000 Kepler stars with an average error of 1.6 Days. Our model, LightPred, is a novel deep-learning model designed to extract stellar rotation periods from light curves. The model utilizes a dual-branch architecture combining Long Short-Term Memory (LSTM) and Transformer components to capture both temporal and global features within the data. We train LightPred on a both self-supervised contrastive pre-training,  and a dataset of simulated light curves generated using a realistic spot model. Our evaluation demonstrates that LightPred outperforms classical methods like the Autocorrelation Function (ACF) in terms of accuracy and robustness.  We apply LightPred to the Kepler dataset, generating the largest catalog to date of stellar rotation periods for main-sequence stars.  Using error analysis, we were able to remove potential contaminations,  false positive predictions and confirm tidal synchronization in eclipsing binaries with orbital periods shorter than 10 days.  Our findings highlight the potential of deep learning in extracting fundamental stellar properties from light curves, opening new avenues for understanding stellar evolution and population demographics.