For a large fraction of transiting hot Jupiters, spectroscopic observations
contain evidence of aerosol in their atmospheres. Generally, transit spectra
are interpreted using retrieval methods, which combine simple parameterised
atmospheric models with a comparison algorithm such as MultiNest. The physical
processes that produce clouds and hazes are complex; however, in a retrieval
model these phenomena must be represented using a minimal number of
parameters. I will discuss the challenges of exoplanet atmosphere retrieval,
with an emphasis on the different approaches to modelling clouds on hot
Jupiters, and how these influence our understanding of their atmospheres.