Searching for exoplanets using artificial intelligence

Kyle A. Pearson, Leon Palafox, Caitlin Griffith

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

In the last decade, over a million stars were monitored to detect transiting planets. Manual interpretation of potential exoplanet candidates is labour intensive and subject to human error, the results of which are difficult to quantify. Here we present a new method of detecting exoplanet candidates in large planetary search projects that, unlike current methods, uses a neural network.Neural networks, also called 'deep learning' or 'deep nets', are designed to give a computer perception into a specific problem by training it to recognize patterns. Unlike past transit detection algorithms, deep nets learn to recognize planet features instead of relying on hand-coded metrics that humans perceive as the most representative. Our convolutional neural network is capable of detecting Earth-like exoplanets in noisy time series data with a greater accuracy than a least-squares method. Deep nets are highly generalizable allowing data to be evaluated from different time series after interpolation without compromising performance. As validated by our deep net analysis of Kepler light curves, we detect periodic transits consistent with the true period without any model fitting. Our study indicates that machine learning will facilitate the characterization of exoplanets in future analysis of large astronomy data sets.

Original languageEnglish (US)
Pages (from-to)478-491
Number of pages14
JournalMonthly Notices of the Royal Astronomical Society
Volume474
Issue number1
DOIs
StatePublished - Feb 1 2018

Fingerprint

artificial intelligence
extrasolar planets
planet
time series
least squares method
transit
astronomy
interpolation
planets
labor
learning
machine learning
light curve
education
stars
analysis
method
detection
project

Keywords

  • Methods: data analysis
  • Planets and satellites: detection
  • Techniques: photometric

ASJC Scopus subject areas

  • Astronomy and Astrophysics
  • Space and Planetary Science

Cite this

Searching for exoplanets using artificial intelligence. / Pearson, Kyle A.; Palafox, Leon; Griffith, Caitlin.

In: Monthly Notices of the Royal Astronomical Society, Vol. 474, No. 1, 01.02.2018, p. 478-491.

Research output: Contribution to journalArticle

@article{a5656dc2d5c648eeae8b1996c79862ad,
title = "Searching for exoplanets using artificial intelligence",
abstract = "In the last decade, over a million stars were monitored to detect transiting planets. Manual interpretation of potential exoplanet candidates is labour intensive and subject to human error, the results of which are difficult to quantify. Here we present a new method of detecting exoplanet candidates in large planetary search projects that, unlike current methods, uses a neural network.Neural networks, also called 'deep learning' or 'deep nets', are designed to give a computer perception into a specific problem by training it to recognize patterns. Unlike past transit detection algorithms, deep nets learn to recognize planet features instead of relying on hand-coded metrics that humans perceive as the most representative. Our convolutional neural network is capable of detecting Earth-like exoplanets in noisy time series data with a greater accuracy than a least-squares method. Deep nets are highly generalizable allowing data to be evaluated from different time series after interpolation without compromising performance. As validated by our deep net analysis of Kepler light curves, we detect periodic transits consistent with the true period without any model fitting. Our study indicates that machine learning will facilitate the characterization of exoplanets in future analysis of large astronomy data sets.",
keywords = "Methods: data analysis, Planets and satellites: detection, Techniques: photometric",
author = "Pearson, {Kyle A.} and Leon Palafox and Caitlin Griffith",
year = "2018",
month = "2",
day = "1",
doi = "10.1093/mnras/stx2761",
language = "English (US)",
volume = "474",
pages = "478--491",
journal = "Monthly Notices of the Royal Astronomical Society",
issn = "0035-8711",
publisher = "Oxford University Press",
number = "1",

}

TY - JOUR

T1 - Searching for exoplanets using artificial intelligence

AU - Pearson, Kyle A.

AU - Palafox, Leon

AU - Griffith, Caitlin

PY - 2018/2/1

Y1 - 2018/2/1

N2 - In the last decade, over a million stars were monitored to detect transiting planets. Manual interpretation of potential exoplanet candidates is labour intensive and subject to human error, the results of which are difficult to quantify. Here we present a new method of detecting exoplanet candidates in large planetary search projects that, unlike current methods, uses a neural network.Neural networks, also called 'deep learning' or 'deep nets', are designed to give a computer perception into a specific problem by training it to recognize patterns. Unlike past transit detection algorithms, deep nets learn to recognize planet features instead of relying on hand-coded metrics that humans perceive as the most representative. Our convolutional neural network is capable of detecting Earth-like exoplanets in noisy time series data with a greater accuracy than a least-squares method. Deep nets are highly generalizable allowing data to be evaluated from different time series after interpolation without compromising performance. As validated by our deep net analysis of Kepler light curves, we detect periodic transits consistent with the true period without any model fitting. Our study indicates that machine learning will facilitate the characterization of exoplanets in future analysis of large astronomy data sets.

AB - In the last decade, over a million stars were monitored to detect transiting planets. Manual interpretation of potential exoplanet candidates is labour intensive and subject to human error, the results of which are difficult to quantify. Here we present a new method of detecting exoplanet candidates in large planetary search projects that, unlike current methods, uses a neural network.Neural networks, also called 'deep learning' or 'deep nets', are designed to give a computer perception into a specific problem by training it to recognize patterns. Unlike past transit detection algorithms, deep nets learn to recognize planet features instead of relying on hand-coded metrics that humans perceive as the most representative. Our convolutional neural network is capable of detecting Earth-like exoplanets in noisy time series data with a greater accuracy than a least-squares method. Deep nets are highly generalizable allowing data to be evaluated from different time series after interpolation without compromising performance. As validated by our deep net analysis of Kepler light curves, we detect periodic transits consistent with the true period without any model fitting. Our study indicates that machine learning will facilitate the characterization of exoplanets in future analysis of large astronomy data sets.

KW - Methods: data analysis

KW - Planets and satellites: detection

KW - Techniques: photometric

UR - http://www.scopus.com/inward/record.url?scp=85045564206&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85045564206&partnerID=8YFLogxK

U2 - 10.1093/mnras/stx2761

DO - 10.1093/mnras/stx2761

M3 - Article

AN - SCOPUS:85045564206

VL - 474

SP - 478

EP - 491

JO - Monthly Notices of the Royal Astronomical Society

JF - Monthly Notices of the Royal Astronomical Society

SN - 0035-8711

IS - 1

ER -