Deriving photometric redshifts using fuzzy archetypes and self-organizing maps - I. Methodology

We propose a method to substantially increase the flexibility and powerof template fitting-based photometric redshifts by transforming a largenumber of galaxy spectral templates into a corresponding collection of'fuzzy archetypes' using a suitable set of perturbative priors designedto account for empirical variation in dust attenuation and emission-linestrengths. To bypass widely separated degeneracies in parameter space(e.g. the redshift-reddening degeneracy), we train self-organizing maps(SOMs) on large 'model catalogues' generated from Monte Carlo samplingof our fuzzy archetypes to cluster the predicted observables in atopologically smooth fashion. Subsequent sampling over the SOM thenallows full reconstruction of the relevant probability distributionfunctions (PDFs). This combined approach enables the multimodalexploration of known variation among galaxy spectral energydistributions with minimal modelling assumptions. We demonstrate thepower of this approach to recover full redshift PDFs using discreteMarkov chain Monte Carlo sampling methods combined with SOMs constructedfrom Large Synoptic Survey Telescope ugrizY and Euclid YJH mockphotometry.