arXiv subjects

Omar Anwar

Publications and source records attributed to Omar Anwar.

2 recordsLinked to original sources

arXiv2026

SM-Net: Learning a Continuous Spectral Manifold from Multiple Stellar Libraries

We present SM-Net, a machine-learning model that learns a continuous spectral manifold from multiple high-resolution stellar libraries. SM-Net generates stellar spectra directly from the fundamental stellar parameters effective temperature (Teff), surface gravity (log g), and metallicity (log Z). It is trained on a combined grid derived from the PHOENIX-Husser, C3K-Conroy, OB-PoWR, and TMAP-Werner libraries. By combining their parameter spaces, we construct a composite dataset that spans a broader and more continuous region of stellar parameter space than any individual library. The unified grid covers Teff = 2,000-190,000 K, log g = -1 to 9, and log Z = -4 to 1, with spectra spanning 3,000-100,000 Angstrom. Within this domain, SM-Net provides smooth interpolation across heterogeneous library boundaries. Outside the sampled region, it can produce numerically smooth exploratory predictions, although these extrapolations are not directly validated against reference models. Zero or masked flux values are treated as unknowns rather than physical zeros, allowing the network to infer missing regions using correlations learned from neighbouring grid points. Across 3,538 training and 11,530 test spectra, SM-Net achieves mean squared errors of 1.47 x 10^-5 on the training set and 2.34 x 10^-5 on the test set in the transformed log1p-scaled flux representation. Inference throughput exceeds 14,000 spectra per second on a single GPU. We also release the model together with an interactive web dashboard for real-time spectral generation and visualisation. SM-Net provides a fast, robust, and flexible data-driven complement to traditional stellar population synthesis libraries.

astro-ph.IM↗

arXiv2025

GalProTE: Galactic Properties Mapping using Transformer Encoder

This work presents GalProTE, a proof-of-concept Machine Learning model utilizing a Transformer Encoder to determine stellar age, metallicity, and dust attenuation from optical spectra. Designed for large astronomical surveys, GalProTE significantly accelerates processing while maintaining accuracy. Using the E-MILES spectral library, we construct a dataset of 111,936 diverse templates by expanding 636 simple stellar population models with varying extinction, spectral combinations, and noise modifications. This ensures robust training over 4750 to 7100 Angstrom at 2.5 Angstrom resolution. GalProTE employs four parallel attention-based encoders with varying kernel sizes to capture spectral features. On synthetic test data, it achieves a mean squared error (MSE) of 0.27% between input and predicted spectra. Validation on PHANGS-MUSE galaxies NGC4254 and NGC5068 confirms its ability to extract physical parameters efficiently, with residuals averaging -0.02% and 0.28% and standard deviations of 4.3% and 5.3%, respectively. To contextualize these results, we compare GalProTE's age, metallicity, and dust attenuation maps with pPXF, a state-of-the-art spectral fitting tool. While pPXF requires approximately 11 seconds per spectrum, GalProTE processes one in less than 4 milliseconds, offering a 2750 times speedup and consuming 68 times less power per spectrum. The strong agreement between pPXF and GalProTE highlights the potential of machine learning to enhance traditional methods, paving the way for faster, energy-efficient, and scalable analyses of galactic properties in modern surveys.

astro-ph.IM↗