Inferring Vector Magnetic Fields from Stokes Profiles of GST/NIRIS Using a Convolutional Neural Network

We propose a new machine learning approach to Stokes inversion based on a convolutional neural network (CNN) and the Milne-Eddington (ME) method. The Stokes measurements used in this study were taken by the Near InfraRed Imaging Spectropolarimeter (NIRIS) on the 1.6 m Goode Solar Telescope (GST) at the Big Bear Solar Observatory. By learning the latent patterns in the training data prepared by the physics-based ME tool, the proposed CNN method is able to infer vector magnetic fields from the Stokes profiles of GST/NIRIS. Experimental results show that our CNN method produces smoother and cleaner magnetic maps than the widely used ME method. Furthermore, the CNN method is 4~6 times faster than the ME method, and is able to produce vector magnetic fields in near real-time, which is essential to space weather forecasting. Specifically, it takes ~50 seconds for the CNN method to process an image of 720 x 720 pixels comprising Stokes profiles of GST/NIRIS. Finally, the CNN-inferred results are highly correlated to the ME-calculated results and are closer to the ME's results with the Pearson product-moment correlation coefficient (PPMCC) being closer to 1 on average than those from other machine learning algorithms such as multiple support vector regression and multilayer perceptrons (MLP). In particular, the CNN method outperforms the current best machine learning method (MLP) by 2.6% on average in PPMCC according to our experimental study. Thus, the proposed physics-assisted deep learning-based CNN tool can be considered as an alternative, efficient method for Stokes inversion for high resolution polarimetric observations obtained by GST/NIRIS.

[1]  S. Solanki,et al.  Retrieval of the full magnetic vector with the He I multiplet at 1083 nm. Maps of an emerging flux region , 2004 .

[2]  A. Ramos,et al.  Returning magnetic flux in sunspot penumbrae , 2012, 1211.6335.

[3]  R. Davies,et al.  Astronomical Society of the Pacific Conference Series , 2010 .

[4]  Thomas G. Dietterich Adaptive computation and machine learning , 1998 .

[5]  Philip R. Goode,et al.  Diffraction‐limited Polarimetry from the Infrared Imaging Magnetograph at Big Bear Solar Observatory , 2006 .

[6]  J. C. del Toro Iniesta,et al.  Inversion of Stokes profiles , 1992 .

[7]  K. Pearson VII. Note on regression and inheritance in the case of two parents , 1895, Proceedings of the Royal Society of London.

[8]  F. Teng Application of Kernel Based Machine Learning to the Inversion Problem of Photospheric Magnetic Fields , 2015 .

[9]  Chang Liu,et al.  Evolution of Photospheric Vector Magnetic Field Associated with Moving Flare Ribbons as Seen by GST , 2018, The Astrophysical Journal.

[10]  A. Ramos,et al.  Sparse inversion of Stokes profiles. I. Two-dimensional Milne-Eddington inversions , 2015, 1503.07666.

[11]  Friedrich Wöger,et al.  The Advanced Technology Solar Telescope: design and early construction , 2012, Other Conferences.

[12]  Manuel Collados European Solar Telescope (EST): project status , 2008, Astronomical Telescopes + Instrumentation.

[13]  F. Galton Regression Towards Mediocrity in Hereditary Stature. , 1886 .

[14]  Ying Guo,et al.  Real Time Stokes Inversion Using Multiple Support Vector Regression , 2004, KES.

[15]  A. Skumanich,et al.  Stokes profile analysis and vector magnetic fields. I. Inversion of photospheric lines , 1987 .

[16]  J. Heasley,et al.  The determination of vector magnetic fields from stokes profiles , 1977 .

[17]  Chang Liu,et al.  Witnessing magnetic twist with high-resolution observation from the 1.6-m New Solar Telescope , 2015, Nature Communications.

[18]  B. R. Cobo,et al.  Spatial deconvolution of spectropolarimetric data: an application to quiet Sun magnetic elements , 2015, 1505.03219.

[19]  M. McHugh Interrater reliability: the kappa statistic , 2012, Biochemia medica.

[20]  J. Schou,et al.  VFISV: Very Fast Inversion of the Stokes Vector for the Helioseismic and Magnetic Imager , 2009, 0901.2702.

[21]  A. Ariste,et al.  Fast Inversion of Spectral Lines Using Principal Components Analysis. II. Inversion of Real Stokes Data , 2001 .

[22]  A. Asensio Ramos,et al.  Stokes Inversion based on Convolutional Neural Networks , 2019, Astronomy & Astrophysics.

[23]  H. Socas-Navarro,et al.  Strategies for Spectral Profile Inversion Using Artificial Neural Networks , 2004, astro-ph/0410567.

[24]  Roberto Gilmozzi,et al.  Ground-based and Airborne Telescopes VII , 2008 .

[25]  Chang Liu,et al.  Transient rotation of photospheric vector magnetic fields associated with a solar flare , 2018, Nature Communications.

[26]  P. R. Goode,et al.  NIRIS: The Second Generation Near-Infrared Imaging Spectro-polarimeter for the 1.6 Meter New Solar Telescope , 2012 .

[27]  Chang Liu,et al.  High-resolution observations of flare precursors in the low solar atmosphere , 2017, Nature Astronomy.

[28]  Hector Socas-Navarro Measuring solar magnetic fields with artificial neural networks , 2003, Neural Networks.

[29]  Philip R. Goode,et al.  The 1.6 m off-axis New Solar Telescope (NST) in Big Bear , 2012, Other Conferences.

[30]  E. L. Degl'innocenti Polarization in spectral lines: III: Resonance polarization in the non-magnetic, collisionless regime , 1984 .

[31]  M. Landolfi,et al.  On the diagnostic of magnetic fields in sunspots through the interpretation of stokes parameters profiles , 1984 .

[32]  Horst Balthasar,et al.  The inversion of Stokes profiles with artificial neural networks , 2001 .

[33]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[34]  Yan Xu,et al.  ULTRA-NARROW NEGATIVE FLARE FRONT OBSERVED IN HELIUM-10830 Å USING THE 1.6 m NEW SOLAR TELESCOPE , 2016, 1601.04729.

[35]  M. Kopf,et al.  Zeeman-Tomography of the Solar Photosphere -- 3-Dimensional Surface Structures Retrieved from Hinode Observations , 2008, 0803.1048.

[36]  N. Draper,et al.  Applied Regression Analysis , 1966 .