Identifying Planetary Transit Candidates in TESS Full-Frame Image Light Curves via Convolutional Neural Networks

TL;DR

This paper introduces a convolutional neural network that efficiently analyzes TESS light curves to identify planetary transit candidates, significantly reducing human effort and enabling large-scale exoplanet searches.

Contribution

The authors develop and publicly release a CNN model that detects transiting exoplanets in TESS data without prior transit parameters, achieving rapid inference and discovering new candidates.

Findings

Identified 181 new planet candidates.

Inference time of ~5ms per light curve on a GPU.

Model effectively dismisses false positives.

Abstract

The Transiting Exoplanet Survey Satellite (TESS) mission measured light from stars in ~75% of the sky throughout its two year primary mission, resulting in millions of TESS 30-minute cadence light curves to analyze in the search for transiting exoplanets. To search this vast data trove for transit signals, we aim to provide an approach that is both computationally efficient and produces highly performant predictions. This approach minimizes the required human search effort. We present a convolutional neural network, which we train to identify planetary transit signals and dismiss false positives. To make a prediction for a given light curve, our network requires no prior transit parameters identified using other methods. Our network performs inference on a TESS 30-minute cadence light curve in ~5ms on a single GPU, enabling large scale archival searches. We present 181 new planet candidates identified by our network, which pass subsequent human vetting designed to rule out false positives. Our neural network model is additionally provided as open-source code for public use and extension.