More data than you want, fewer data than you need: machine learning approaches to starlight subtraction with MagAO-X

Joseph D. Long; Jared R. Males; Laird M. Close; Olivier Guyon; Sebastiaan Y. Haffert; Alycia J. Weinberger; Jay Kueny; Kyle Van Gorkom; Eden McEwen; Logan Pearce; Maggie Kautz; Jialin Li; Jennifer Lumbres; Alexander Hedglen; Lauren Schatz; Avalon McLeod; Isabella Doty; Warren B. Foster; Roswell Roberts; Katie Twitchell

doi:10.1117/12.3018424

More data than you want, fewer data than you need: machine learning approaches to starlight subtraction with MagAO-X

Joseph D. Long, Jared R. Males, Laird M. Close, Olivier Guyon, Sebastiaan Y. Haffert, Alycia J. Weinberger, Jay Kueny, Kyle Van Gorkom, Eden McEwen, Logan Pearce, Maggie Kautz, Jialin Li, Jennifer Lumbres, Alexander Hedglen, Lauren Schatz, Avalon McLeod, Isabella Doty, Warren B. Foster, Roswell Roberts, Katie Twitchell

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Scopus citations

Abstract

High-contrast imaging data analysis depends on removing residual starlight from the host star to reveal planets and disks. Most observers do this with principal components analysis (i.e. KLIP) using modes computed from the science images themselves. These modes may not be orthogonal to planet and disk signals, leading to over-subtraction. The wavefront sensor data recorded during the observation provide an independent signal with which to predict the instrument point-spread function (PSF). MagAO-X is an extreme adaptive optics (ExAO) system for the 6.5-meter Magellan Clay telescope and a technology pathfinder for ExAO with GMagAO-X on the upcoming Giant Magellan Telescope. MagAO-X is designed to save all sensor information, including kHz-speed wavefront measurements. Our software and compressed data formats were designed to record the millions of training samples required for machine learning with high throughput. The large volume of image and sensor data lets us learn a PSF model incorporating all the information available. This allows us to probe smaller star-planet separations at greater sensitivities, which will be needed for rocky planet imaging.

Original language	English (US)
Title of host publication	Adaptive Optics Systems IX
Editors	Kathryn J. Jackson, Dirk Schmidt, Elise Vernet
Publisher	SPIE
ISBN (Electronic)	9781510675179
DOIs	https://doi.org/10.1117/12.3018424
State	Published - 2024
Event	Adaptive Optics Systems IX 2024 - Yokohama, Japan Duration: Jun 16 2024 → Jun 22 2024

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	13097
ISSN (Print)	0277-786X
ISSN (Electronic)	1996-756X

Conference

Conference	Adaptive Optics Systems IX 2024
Country/Territory	Japan
City	Yokohama
Period	6/16/24 → 6/22/24

Keywords

Adaptive optics
automatic differentiation
exoplanets
high-contrast imaging
machine learning

ASJC Scopus subject areas

Electronic, Optical and Magnetic Materials
Condensed Matter Physics
Computer Science Applications
Applied Mathematics
Electrical and Electronic Engineering

Access to Document

10.1117/12.3018424

Cite this

Long, J. D., Males, J. R., Close, L. M., Guyon, O., Haffert, S. Y., Weinberger, A. J., Kueny, J., Van Gorkom, K., McEwen, E., Pearce, L., Kautz, M., Li, J., Lumbres, J., Hedglen, A., Schatz, L., McLeod, A., Doty, I., Foster, W. B., Roberts, R., & Twitchell, K. (2024). More data than you want, fewer data than you need: machine learning approaches to starlight subtraction with MagAO-X. In K. J. Jackson, D. Schmidt, & E. Vernet (Eds.), Adaptive Optics Systems IX Article 130972X (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 13097). SPIE. https://doi.org/10.1117/12.3018424

More data than you want, fewer data than you need: machine learning approaches to starlight subtraction with MagAO-X. / Long, Joseph D.; Males, Jared R.; Close, Laird M. et al.
Adaptive Optics Systems IX. ed. / Kathryn J. Jackson; Dirk Schmidt; Elise Vernet. SPIE, 2024. 130972X (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 13097).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Long, JD, Males, JR, Close, LM , Guyon, O, Haffert, SY, Weinberger, AJ, Kueny, J, Van Gorkom, K, McEwen, E, Pearce, L, Kautz, M, Li, J, Lumbres, J, Hedglen, A, Schatz, L, McLeod, A, Doty, I, Foster, WB, Roberts, R & Twitchell, K 2024, More data than you want, fewer data than you need: machine learning approaches to starlight subtraction with MagAO-X. in KJ Jackson, D Schmidt & E Vernet (eds), Adaptive Optics Systems IX., 130972X, Proceedings of SPIE - The International Society for Optical Engineering, vol. 13097, SPIE, Adaptive Optics Systems IX 2024, Yokohama, Japan, 6/16/24. https://doi.org/10.1117/12.3018424

Long JD, Males JR, Close LM , Guyon O, Haffert SY, Weinberger AJ et al. More data than you want, fewer data than you need: machine learning approaches to starlight subtraction with MagAO-X. In Jackson KJ, Schmidt D, Vernet E, editors, Adaptive Optics Systems IX. SPIE. 2024. 130972X. (Proceedings of SPIE - The International Society for Optical Engineering). doi: 10.1117/12.3018424

@inproceedings{66ea3aa9f39e48ad9e9e8a80086d0b98,

title = "More data than you want, fewer data than you need: machine learning approaches to starlight subtraction with MagAO-X",

abstract = "High-contrast imaging data analysis depends on removing residual starlight from the host star to reveal planets and disks. Most observers do this with principal components analysis (i.e. KLIP) using modes computed from the science images themselves. These modes may not be orthogonal to planet and disk signals, leading to over-subtraction. The wavefront sensor data recorded during the observation provide an independent signal with which to predict the instrument point-spread function (PSF). MagAO-X is an extreme adaptive optics (ExAO) system for the 6.5-meter Magellan Clay telescope and a technology pathfinder for ExAO with GMagAO-X on the upcoming Giant Magellan Telescope. MagAO-X is designed to save all sensor information, including kHz-speed wavefront measurements. Our software and compressed data formats were designed to record the millions of training samples required for machine learning with high throughput. The large volume of image and sensor data lets us learn a PSF model incorporating all the information available. This allows us to probe smaller star-planet separations at greater sensitivities, which will be needed for rocky planet imaging.",

keywords = "Adaptive optics, automatic differentiation, exoplanets, high-contrast imaging, machine learning",

author = "Long, {Joseph D.} and Males, {Jared R.} and Close, {Laird M.} and Olivier Guyon and Haffert, {Sebastiaan Y.} and Weinberger, {Alycia J.} and Jay Kueny and {Van Gorkom}, Kyle and Eden McEwen and Logan Pearce and Maggie Kautz and Jialin Li and Jennifer Lumbres and Alexander Hedglen and Lauren Schatz and Avalon McLeod and Isabella Doty and Foster, {Warren B.} and Roswell Roberts and Katie Twitchell",

note = "Publisher Copyright: {\textcopyright} 2024 SPIE.; Adaptive Optics Systems IX 2024 ; Conference date: 16-06-2024 Through 22-06-2024",

year = "2024",

doi = "10.1117/12.3018424",

language = "English (US)",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

publisher = "SPIE",

editor = "Jackson, {Kathryn J.} and Dirk Schmidt and Elise Vernet",

booktitle = "Adaptive Optics Systems IX",

}

TY - GEN

T1 - More data than you want, fewer data than you need

T2 - Adaptive Optics Systems IX 2024

AU - Long, Joseph D.

AU - Males, Jared R.

AU - Close, Laird M.

AU - Guyon, Olivier

AU - Haffert, Sebastiaan Y.

AU - Weinberger, Alycia J.

AU - Kueny, Jay

AU - Van Gorkom, Kyle

AU - McEwen, Eden

AU - Pearce, Logan

AU - Kautz, Maggie

AU - Li, Jialin

AU - Lumbres, Jennifer

AU - Hedglen, Alexander

AU - Schatz, Lauren

AU - McLeod, Avalon

AU - Doty, Isabella

AU - Foster, Warren B.

AU - Roberts, Roswell

AU - Twitchell, Katie

PY - 2024

Y1 - 2024

N2 - High-contrast imaging data analysis depends on removing residual starlight from the host star to reveal planets and disks. Most observers do this with principal components analysis (i.e. KLIP) using modes computed from the science images themselves. These modes may not be orthogonal to planet and disk signals, leading to over-subtraction. The wavefront sensor data recorded during the observation provide an independent signal with which to predict the instrument point-spread function (PSF). MagAO-X is an extreme adaptive optics (ExAO) system for the 6.5-meter Magellan Clay telescope and a technology pathfinder for ExAO with GMagAO-X on the upcoming Giant Magellan Telescope. MagAO-X is designed to save all sensor information, including kHz-speed wavefront measurements. Our software and compressed data formats were designed to record the millions of training samples required for machine learning with high throughput. The large volume of image and sensor data lets us learn a PSF model incorporating all the information available. This allows us to probe smaller star-planet separations at greater sensitivities, which will be needed for rocky planet imaging.

AB - High-contrast imaging data analysis depends on removing residual starlight from the host star to reveal planets and disks. Most observers do this with principal components analysis (i.e. KLIP) using modes computed from the science images themselves. These modes may not be orthogonal to planet and disk signals, leading to over-subtraction. The wavefront sensor data recorded during the observation provide an independent signal with which to predict the instrument point-spread function (PSF). MagAO-X is an extreme adaptive optics (ExAO) system for the 6.5-meter Magellan Clay telescope and a technology pathfinder for ExAO with GMagAO-X on the upcoming Giant Magellan Telescope. MagAO-X is designed to save all sensor information, including kHz-speed wavefront measurements. Our software and compressed data formats were designed to record the millions of training samples required for machine learning with high throughput. The large volume of image and sensor data lets us learn a PSF model incorporating all the information available. This allows us to probe smaller star-planet separations at greater sensitivities, which will be needed for rocky planet imaging.

KW - Adaptive optics

KW - automatic differentiation

KW - exoplanets

KW - high-contrast imaging

KW - machine learning

UR - http://www.scopus.com/inward/record.url?scp=85200382827&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85200382827&partnerID=8YFLogxK

U2 - 10.1117/12.3018424

DO - 10.1117/12.3018424

M3 - Conference contribution

AN - SCOPUS:85200382827

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Adaptive Optics Systems IX

A2 - Jackson, Kathryn J.

A2 - Schmidt, Dirk

A2 - Vernet, Elise

PB - SPIE

Y2 - 16 June 2024 through 22 June 2024

ER -

More data than you want, fewer data than you need: machine learning approaches to starlight subtraction with MagAO-X

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this