arXiv preprint ·
Stellar parameters and abundances for 3.2 million DESI DR1 spectra using machine learning
- Published
- Source
- arXiv preprint, posted before peer review
- Original
- Read it at arXiv
- Topics
Abstract
We present a machine-learning pipeline for determining stellar parameters and chemical abundances from spectra observed by the Dark Energy Spectroscopic Instrument (DESI). Building on the model-driven Payne framework, we train an artificial neural network on ~370,000 synthetic spectra spanning FGK stars, and fit the network directly to the flux-calibrated, un-normalised spectral energy distribution (SED) of each star, rather than to a continuum-normalised spectrum as done by DESI's official Stellar Parameter (SP) pipeline. Internal accuracy tests show a median interpolation error of less than 0.3% for 90% of a synthetic verification sample, and the full fitting workflow recovers input labels to high accuracy even at low signal-to-noise ratio (S/N). Validating our method on cross-matched samples of 6719 APOGEE and 3455 GALAH stars observed by DESI with S/N > 20, we recover Teff and logg with smaller systematic offsets than the DESI SP pipeline, an improvement we attribute to fitting the full SED shape rather than a normalised spectrum. We further recover 12 elemental abundances (Na, Mg, Al, Si, Ca, Ti, V, Cr, Mn, Ni, Ba, and Y) to an accuracy that matches or exceeds SP where a comparison is possible, and provide four abundances, Mn, V, Ba, and Y, that SP does not report at all. Applying this pipeline to the DESI Data Release 1 stellar sample, we deliver a catalogue of stellar parameters and abundances for 3.2 million stars.
The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.
Read it on ReadSearch
Each opens after you create an account, and brings you back to this paper.
More in Solar & Stellar Astrophysics
AI Agents as a Research Team: A Human-Supervised Case Study in Stellar Spectroscopy
Cross-Modal Solar Image Synthesis: Adapting the Surya Foundation Model from He I 10830 Å to EUV Translation and Coronal Hole Segmentation
A Framework for Automated Multi-Source Satellite Data Analytics and LLM-Based Report Generation
Measuring the Age and Initial Composition of Interstellar Objects
Armillary: A Label-Free Coordinate System for Stellar Spectra
Get the few papers worth your time in Solar & Stellar Astrophysics every morning.
ReadSearch reads the new papers in your field each day and brings you the few worth reading, each with its key finding in a sentence.