| Constraining
                  the properties of AGN by fitting model templates to
                  their broad-band Spectral Energy Distribution 
 Implementation of AGN model templates into the CIGALE Spectral Energy Distribution
              fitting code. Applying CIGALE to the observed
              multi-wavelength photometry of AGN provides constraints to
              both the properties of the active black hole (e.g.
              accretion luminosity) and the characteristics of the
              underlying host galaxy (e.g. stellar mass, star-formation
              rate). Simulations are used to produce mock Spectral Energy
              Distributions of galaxies with realistic star-formation
              histories and with variable levels of superimposed AGN
              emission. The plots on the left show the examples of two
              such simulations. The broad-band photometric data points
              are for galaxies hosting unobscured (type 1) AGN with
              accretion luminosities that represent 20% (top panel) and
              60% (middle panel) of the total infrared luminosity. The CIGALE is then used to fit the mock
              SEDs with model AGN+galaxy templates (curves on the plots)
              by taking into account dust extinction and redistribution
              of the absorbed radiation in the infrared part of the
              spectrum.
 These simulation can test how well AGN and galaxy
              properties can be constrained via template fits to
              broad-band multi-wavelength observations. The bottom
              figure shows that the stellar mass of AGN hosts can be
              robustly determined. It plots the relative accuracy of the
              stellar mass determination as a function of the fraction
              of AGN light relative to the galaxy. The different curves
              correspond to different parametrisations of the
              star-formation history of the galaxy. Stellar mass can be
              determined at better than the 40% level from the observed
              Spectral Energy Distribution. More details in Ciesla et al. (2015). 
 
 |