Nonparametric models which allow for data with unobservable heterogeneity are studied. The first publication introduces new estimators and their asymptotic properties for conditional mixture models. The second publication considers estimation of a function from noisy observations of its Radon transform in a Gaussian white noise model.
