In this lecture, we will talk about different ways in which possibilistic modelling can be used to solve learning problems. The lecture will be divided into two main parts: the first part will explore how possibilistic tool can be used to learn statistical model from data, with a focus on the profile likelihood as an instrumental tool; the second part will explore how possibilistic modelling can be used to model uncertain data, and how such uncertain data can be integrated into learning strategies. Provided we have sufficient time, a final part will give some elements as to how these two aspects can be merged together.