AGNFormer I: Reconstruction of AGN spectra using a probabilistic transformer model

概要

We explore how an uncertainty-aware transformer-based architecture can leverage information embedded across the entire observed optical spectra of AGN, focusing on the algorithm's ability to predict unseen or masked parts of luminous AGN spectra. This provides a direct probe of the learnable correlations between AGN continua and broad lines. We introduce AGNFormer, a transformer model trained to predict the mean expected flux and variance in masked spectral regions (major broad lines to ${\pm}10…