Showing papers by "Gregor Traven published in 2023"
Peer Review•
08 Jun 2023
TL;DR: In this paper , a hybrid convolutional neural network (CNN) was proposed for supercharged stellar parametrization by combining Gaia RVS spectra with the full set of Gaia products and high-resolution, high-quality spectroscopic reference data sets.
Abstract: Gaia DR3 has provided the community with about one million RVS spectra covering the CaII triplet region. In the next Gaia data releases, we anticipate the number of RVS spectra to successively increase from several 10 million spectra to eventually more than 200M spectra. Thus, stellar spectra are produced on an"industrial scale"with numbers well above those for current and anticipated ground based surveys. However, many of these spectra have low S/N (from 15 to 25 per pixel), such that they pose problems for classical spectral analysis pipelines and therefore alternative ways to tap into these large datasets need to be devised. We aim to leverage the versatility/capabilities of machine learning techniques for supercharged stellar parametrization, by combining Gaia RVS spectra with the full set of Gaia products and high-resolution, high-quality spectroscopic reference data sets. We develop a hybrid Convolutional Neural-Network (CNN) which combines the Gaia DR3 RVS spectra, photometry (G, Bp, Rp), parallaxes, and XP coefficients to derive atmospheric parameters (Teff, log(g), and overall [M/H]) and chemical abundances ([Fe/H] and [$\alpha$/M]). We trained the CNN with a high-quality training sample based on APOGEE DR17 labels. With this CNN, we derived homogeneous atmospheric parameters and abundances for 841300 stars, that remarkably compared to external data-sets. The CNN is robust against noise in the RVS data, and very precise labels are derived down to S/N=15. We managed to characterize the [$\alpha$/M]-[M/H] bimodality from the inner regions to the outer parts of the Milky Way, which has never been done using RVS spectra or similar datasets. This work is the first to combine machine-learning with such diverse datasets (spectroscopy, astrometry, and photometry), and paves the way for the large scale machine-learning analysis of Gaia-RVS spectra from future data releases.
TL;DR: In this article , the authors used more than 872,000 mid-to-high resolution (R$20,000) spectra of stars from the GALAH survey to determine the spectrum of diffuse interstellar bands (DIBs).
Abstract: We use more than 872,000 mid-to-high resolution (R $\sim$ 20,000) spectra of stars from the GALAH survey to discern the spectra of diffuse interstellar bands (DIBs). We use four windows with the wavelength range from 4718 to 4903, 5649 to 5873, 6481 to 6739, and 7590 to 7890 \AA, giving a total coverage of 967 \AA. We produce $\sim$400,000 spectra of interstellar medium (ISM) absorption features and correct them for radial velocities of the DIB clouds. Ultimately, we combine the 33,115 best ISM spectra into six reddening bins with a range of $0.1 \,\mathrm{mag}