In this chapter, we looked at the field of meta learning, which can be described as learning to learn. We started with an introduction to meta learning. More specifically, we talked about zero-shot and few-shot learning, as well as meta training and meta testing. Then, we focused on several metric-based learning approaches. We looked at matching networks, implemented an example of a Siamese network, and we introduced prototypical networks. Next, we focused on optimization-based learning, where we introduced the MAML algorithm.
In the next chapter, we'll learn about an exciting topic: automated vehicles.
