Self-Organizing Maps (SOM), an artificial neural network trained by unsupervised learning for dimensionality reduction, has shown to be a powerful tool to extract knowledge from multiband photometric observations of galaxies including to calibrate the color-redshift relation. In this talk, I will present my new results suggesting that SOMs can also be a powerful tool for estimating the spectroscopic properties of large samples of galaxies from their broadband photometry. I have trained SOMs with near ultraviolet to infrared photometry from the COSMOS 2020 photometric catalogue and used measured spectroscopic properties for a subsample of galaxies from LEGA-C and zCOSMOS surveys to estimate spectroscopic features such as D4000 for the parent sample. I will discuss potential applications of this method including using calibrated SOMs to measure galaxy physical properties in future massive photometric surveys with Euclid and Rubin where millions of galaxies will be observed with a limited number of filters. Also, we can use trained SOMs to identify regions of parameter space where we lack enough spectroscopic measurements to create the necessary complete sample for a comprehensive study of stellar population of galaxies, including for chemical evolution studies.