Detrending Exoplanetary Transit Light Curves with Long Short-Term Memory Networks

Abstract

The precise derivation of transit depths from transit light curves is a key component for measuring exoplanet transit spectra, and henceforth for the study of exoplanet atmospheres. However, it is still deeply affected by various kinds of systematic errors and noise. In this paper we propose a new detrending method by reconstructing the stellar flux baseline during transit time. We train a probabilistic Long Short-Term Memory (LSTM) network to predict the next data point of the light curve during the out-of-transit, and use this model to reconstruct a transit-free light curve - i.e. including only the systematics - during the in-transit. By making no assumption about the instrument, and using only the transit ephemeris, this provides a general way to correct the systematics and perform a subsequent transit fit. The name of the proposed model is TLCD-LSTM, standing for Transit Light Curve Detrending LSTM. Here we present the first results on data from six transit observations of HD 189733b with the IRAC camera on board the Spitzer Space Telescope, and discuss some of its possible further applications.

12 pages, 10 figures, 4 tables, accepted for publication in The Astronomical Journal

References in corpus (23)

• Adam: A Method for Stochastic Optimization
• Detection of Planetary Transits Across a Sun-like Star
• Fast and scalable Gaussian process modeling with applications to astronomical time series
• A map of the day-night contrast of the extrasolar planet HD 189733b
• Identifying Exoplanets with Deep Learning: A Five Planet Resonant Chain around Kepler-80 and an Eighth Planet around Kepler-90
• The climate of HD 189733b from fourteen transits and eclipses measured by Spitzer
• Parameter Estimation from Time-Series Data with Correlated Errors: A Wavelet-Based Method and its Application to Transit Light Curves
• A Gaussian process framework for modelling instrumental systematics: application to transmission spectroscopy
• Spitzer Secondary Eclipses of the Dense, Modestly-irradiated, Giant Exoplanet HAT-P-20b Using Pixel-Level Decorrelation
• Analytic Approximations for Transit Light Curve Observables, Uncertainties, and Covariances
• A new approach to analysing HST spatial scans: the transmission spectrum of HD 209458 b
• Deep Recurrent Neural Networks for Supernovae Classification
• ExoGAN: Retrieving Exoplanetary Atmospheres Using Deep Convolutional Generative Adversarial Networks
• Scientific Domain Knowledge Improves Exoplanet Transit Classification with Deep Learning
• Of `Cocktail Parties' and Exoplanets
• Machine Learning Techniques for Stellar Light Curve Classification
• Rapid Classification of TESS Planet Candidates with Convolutional Neural Networks
• Supervised detection of exoplanets in high-contrast imaging sequences
• A new look at Spitzer primary transit observations of the exoplanet HD189733b
• Repeatability of Spitzer/IRAC exoplanetary eclipses with Independent Component Analysis
• Selective Principal Component Extraction and Reconstruction: A Novel Method for Ground Based Exoplanet Spectroscopy
• Aldebaran b's temperate past uncovered in planet search data
• Speech Recognition with Deep Recurrent Neural Networks

Cited by in corpus (7)

• TESS Transit Timing of Hundreds of Hot Jupiters
• On The Compatibility of Ground-based and Space-based Data: WASP-96 b, An Example
• KELT-11 b: Abundances of water and constraints on carbon-bearing molecules from the Hubble transmission spectrum
• The transmission spectrum of WASP-17 b from the optical to the near-infrared wavelengths: combining STIS, WFC3 and IRAC datasets
• PyLightcurve-torch: a transit modelling package for deep learning applications in PyTorch
• Atmospheric characterization of hot Jupiters using hierarchical models of Spitzer observations
• Characterization of Gravitational Waves Signals Using Neural Networks