receiving a source generative model for a source domain and a set of training images for a target domain;

adapting the source generative model using the set of training images to provide an adapted generative model for the target domain;

determining a plurality of interpretable factors for the source generative model and/or the adapted generative model, each interpretable factor comprising a direction in a latent space of the source generative model and/or a direction in a latent space of the adapted generative model;

receiving input regarding a user-selected interpretable factor from the plurality of interpretable factors, the user-selected interpretable factor corresponding to a first subset of weights in a weight matrix for the source generative model;

generating a user-adapted generative model for the target domain based on the user-selected interpretable factor by training the user-adapted generative model using a loss function minimizing a distance between the first subset of weights in the weight matrix for the source generative model and a corresponding first subset of weights in a weight matrix for the user-adapted generative model; and

using the user-adapted generative model to generate a new image in the target domain.