Over the last few years, deep learning-based methods have achieved impressive results on image understanding problems, such as image classification, object detection, or semantic segmentation. However, real-word computer vision applications often require models that are able to (a) learn with few annotated examples, and (b) continually adapt to new data without forgetting prior knowledge. Unfortunately, classic supervised deep learning methods have not been designed with these requirements in mind. As a result, one of the next big challenges in computer vision is to develop learning approaches that are capable of addressing the important shortcomings of existing methods in this regard. This tutorial will cover possible approaches for achieving this goal. Specifically, the tutorial consists of three parts that cover the subjects of (1) incremental learning, (2) few-shot learning, and (3) leveraging self-supervised learning.