Autoencoding Galaxy Spectra II: Redshift Invariance and Outlier Detection

Abstract

We present an unsupervised outlier detection method for galaxy spectra based on the spectrum autoencoder architecture spender, which reliably captures spectral features and provides highly realistic reconstructions for SDSS galaxy spectra. We interpret the sample density in the autoencoder latent space as a probability distribution, and identify outliers as low-probability objects with a normalizing flow. However, we found that the latent-space position is not, as expected from the architecture, redshift invariant, which introduces stochasticity into the latent space and the outlier detection method. We solve this problem by adding two novel loss terms during training, which explicitly link latent-space distances to data-space distances, preserving locality in the autoencoding process. Minimizing the additional losses leads to a redshift-invariant, non-degenerate latent space distribution with clear separations between common and anomalous data. We inspect the spectra with the lowest probability and find them to include blends with foreground stars, extremely reddened galaxies, galaxy pairs and triples, and stars that are misclassified as galaxies. We release the newly trained spender model and the latent-space probability for the entire SDSS-I galaxy sample to aid further investigations.

arXiv comment12 pages, 7 figures, submitted to AJ. Code available at https://github.com/pmelchior/spender

References in corpus (19)

• Array Programming with NumPy
• The Astropy Project: Sustaining and Growing a Community-oriented Open-source Project and the Latest Major Release (v5.0) of the Core Package
• The Sixteenth Data Release of the Sloan Digital Sky Surveys: First Release from the APOGEE-2 Southern Survey and Full Release of eBOSS Spectra
• Spectral Classification and Redshift Measurement for the SDSS-III Baryon Oscillation Spectroscopic Survey
• Dust Attenuation Curves in the Local Universe: Demographics and New Laws for Star-forming Galaxies and High-redshift Analogs
• GALEX-SDSS-WISE Legacy Catalog (GSWLC): Star Formation Rates, Stellar Masses and Dust Attenuations of 700,000 Low-redshift Galaxies
• Recovering galaxy star formation and metallicity histories from spectra using VESPA
• Strongly star forming galaxies in the local Universe with nebular He II 4686 emission
• An Atlas of Galaxy Spectral Energy Distributions from the Ultraviolet to the Mid-Infrared
• The Completed SDSS-IV extended Baryon Oscillation Spectroscopic Survey: Large-scale Structure Catalogs for Cosmological Analysis
• The weirdest SDSS galaxies: results from an outlier detection algorithm
• Discovery of 90 Type Ia supernovae among 700,000 Sloan spectra: the Type-Ia supernova rate versus galaxy mass and star-formation rate at redshift ~0.1
• Dimensionality Reduction of SDSS Spectra with Variational Autoencoders
• The Galah Survey: Classification and diagnostics with t-SNE reduction of spectral information
• Anomaly detection for machine learning redshifts applied to SDSS galaxies
• An optical overview of blazars with LAMOST I: Hunting changing-look blazars and new redshift estimates
• Elevated Hot Gas and High-Mass X-ray Binary Emission in Low Metallicity Galaxies: Implications for Nebular Ionization and Intergalactic Medium Heating in the Early Universe
• Autoencoding Galaxy Spectra I: Architecture
• Resolved and Integrated Stellar Masses in the SDSS-IV/MaNGA Survey, Paper II: Applications of PCA-based stellar mass estimates

Cited by in corpus (10)

• AstroCLIP: A Cross-Modal Foundation Model for Galaxies
• A review of unsupervised learning in astronomy
• Autoencoding Galaxy Spectra I: Architecture
• PopSED: Population-Level Inference for Galaxy Properties from Broadband Photometry with Neural Density Estimation
• Galaxy Spectra Networks (GaSNet). III. Generative pre-trained network for spectrum reconstruction, redshift estimate and anomaly detection
• Inferring redshift and galaxy properties via a multi-task neural net with probabilistic outputs: An application to simulated MOONS spectra
• Ly$\alpha$NNA: A Deep Learning Field-level Inference Machine for the Lyman-$\alpha$ Forest
• SpecPT (Spectroscopy Pre-trained Transformer) Model for Extragalactic Spectroscopy: I. Architecture and Automated Redshift Measurement
• Automated quasar continuum estimation using neural networks: a comparative study of deep-learning architectures
• Extracting latent representations from X-ray spectra. Classification, regression, and accretion signatures of Chandra sources