Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Computing Transiting Exoplanet Parameters with 1D Convolutional Neural Networks

Santiago Iglesias Álvarez, Enrique Díez Alonso, María Luisa Sánchez Rodríguez, Javier Rodríguez Rodríguez, Saúl Pérez Fernández, Francisco Javier de Cos Juez

TL;DR

This work addresses the challenge of rapidly characterizing transiting exoplanets from stellar light curves by leveraging two 1D CNNs: one processes complete light curves to estimate the orbital period $P$, and the other processes phase-folded light curves to estimate $(R_p/R____/)^2$ and $a/R_____$. The models are trained on simulated, TESS-like data generated with the Batman transit model (including limb darkening via Mandel–Agol), and are validated on real TESS light curves, demonstrating high predictive accuracy (e.g., $R^2$ up to 0.991 for $P$) and strong agreement with traditional TLS results while drastically reducing computation time. Key findings include comparable accuracy to TLS on large test sets and real data, with a dramatic speedup (approximately 30 s for 150k light curves) and no reliance on prior parameter settings. The work thus offers a scalable path toward real-time exoplanet characterization in large survey datasets.

Abstract

The transit method allows the detection and characterization of planetary systems by analyzing stellar light curves. Convolutional neural networks appear to offer a viable solution for automating these analyses. In this research, two 1D convolutional neural network models, which work with simulated light curves in which transit-like signals were injected, are presented. One model operates on complete light curves and estimates the orbital period, and the other one operates on phase-folded light curves and estimates the semimajor axis of the orbit and the square of the planet-to-star radius ratio. Both models were tested on real data from TESS light curves with confirmed planets to ensure that they are able to work with real data. The results obtained show that 1D CNNs are able to characterize transiting exoplanets from their host star's detrended light curve and, furthermore, reducing both the required time and computational costs compared with the current detection and characterization algorithms.

Computing Transiting Exoplanet Parameters with 1D Convolutional Neural Networks

TL;DR

This work addresses the challenge of rapidly characterizing transiting exoplanets from stellar light curves by leveraging two 1D CNNs: one processes complete light curves to estimate the orbital period , and the other processes phase-folded light curves to estimate and . The models are trained on simulated, TESS-like data generated with the Batman transit model (including limb darkening via Mandel–Agol), and are validated on real TESS light curves, demonstrating high predictive accuracy (e.g., up to 0.991 for ) and strong agreement with traditional TLS results while drastically reducing computation time. Key findings include comparable accuracy to TLS on large test sets and real data, with a dramatic speedup (approximately 30 s for 150k light curves) and no reliance on prior parameter settings. The work thus offers a scalable path toward real-time exoplanet characterization in large survey datasets.

Abstract

The transit method allows the detection and characterization of planetary systems by analyzing stellar light curves. Convolutional neural networks appear to offer a viable solution for automating these analyses. In this research, two 1D convolutional neural network models, which work with simulated light curves in which transit-like signals were injected, are presented. One model operates on complete light curves and estimates the orbital period, and the other one operates on phase-folded light curves and estimates the semimajor axis of the orbit and the square of the planet-to-star radius ratio. Both models were tested on real data from TESS light curves with confirmed planets to ensure that they are able to work with real data. The results obtained show that 1D CNNs are able to characterize transiting exoplanets from their host star's detrended light curve and, furthermore, reducing both the required time and computational costs compared with the current detection and characterization algorithms.
Paper Structure (7 sections, 10 equations, 14 figures, 4 tables)

This paper contains 7 sections, 10 equations, 14 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Materials and Methods
  3. Transit Shape
  4. Light Curve Simulation
  5. Convolutional Neural Networks (CNNs): Our 1D CNN Models
  6. Model Training and Test Results and Discussion
  7. Conclusions

Figures (14)

  • Figure S1: Theoretical transit model without limb darkening.
  • Figure S2: Example of a transit shape with the Mandel and Agol model computed with TLS from the simulated light curve shown in the upper panel of Figure \ref{['fig: lcs']}. Black: phase-folded light curve. Red: TLS model.
  • Figure S3: Example of the 2 views of a simulated light curve. Top panel: complete light curve. Bottom panel: phase-folded light curve. Phase is computed as $(t-t_0)/P$.
  • Figure S4: Zoom of the light curves of Figure \ref{['fig: lcs']}. Top panel: complete light curve. Bottom panel: phase-folded light curve.
  • Figure S5: TOI-192 light curve. Top panel: Full light curve. Bottom panel: phase-folded light curve computed with TLS.
  • ...and 9 more figures