Deep learning interpretability analysis for carbon star identification in Gaia DR3

Abstract

Context. A large fraction of Asymptotic Giant Branch (AGB) stars develop carbon-rich atmospheres during their evolution. Based on their color and luminosity, these carbon stars can be easily distinguished from many other kinds of stars. However, numerous G, K, and M giants also occupy the same region as carbon stars on the HR diagram. Despite this, their spectra exhibit differences, especially in the prominent CN molecular bands. Target. We aim to distinguish carbon stars from other kinds of stars using $Gaia$'s XP spectra, while providing attributional interpretations of key features necessary for identification, and even discovering additional new spectral key features. Method. We propose a classification model named `GaiaNet', an improved one-dimensional convolutional neural network specifically designed for handling $Gaia$'s XP spectra. We utilized the SHAP interpretability model to determine SHAP values for each feature in a spectrum, enabling us to explain the output of the `GaiaNet' model and provide further meaningful analysis. Compared to four traditional machine-learning methods, the `GaiaNet' model exhibits an average classification accuracy improvement of approximately 0.3% on the validation set, with the highest accuracy reaching 100%. Utilizing the SHAP model, we present a clear spectroscopic heatmap highlighting molecular band absorption features primarily distributed around CN773.3 and CN895.0, and summarize five key feature regions for carbon star identification. Upon applying the trained classification model to the CSTAR sample with Gaia `xp_sampled_mean' spectra, we obtained 451 new candidate carbon stars as a by-product.

23 pages, 22 figures

References in corpus (17)

• XGBoost: A Scalable Tree Boosting System
• A 3D Dust Map Based on Gaia, Pan-STARRS 1 and 2MASS
• Gaia Data Release 3: Astrophysical parameters inference system (Apsis) I -- methods and content overview
• Spectroscopy of New High Proper Motion Stars in the Northern Sky. I. New Nearby Stars, New High Velocity Stars, and an Enhanced Classification Scheme for M Dwarfs
• s-Process Nucleosynthesis in Carbon Stars
• Internal calibration of Gaia BP/RP low-resolution spectra
• Self-consistent Stellar Radial Velocities from LAMOST Medium-Resolution Survey (MRS) DR7
• Gaia Data Release 3: The second Gaia catalogue of Long-Period Variable candidates
• Carbon stars identified from LAMOST DR4 using Machine Learning
• Characterisation of Galactic carbon stars and related stars from Gaia-EDR3
• Carbon stars from LAMOST DR2 data
• Properties of carbon stars in the Solar neighbourhood based on Gaia DR2 astrometry
• Metallic-Line Stars Identified from Low Resolution Spectra of LAMOST DR5
• Search for carbon stars and DZ white dwarfs in SDSS spectra survey through machine learning
• Hunting for C-rich long-period variable stars in the Milky Way's bar-bulge using unsupervised classification of Gaia BP/RP spectra
• Identification, mass and age of primary red clump stars from spectral features derived with the LAMOST DR7
• Value-added catalog of M-giant stars in LAMOST DR9

Cited by in corpus (2)

• Interpreting anomaly detection of SDSS spectra
• Explainable machine learning classification of \textit{Chandra} X-ray sources: SHAP analysis of multi-wavelength features