In this chapter, we've learned about TAML for reducing the task bias. We saw two types of methods: entropy-based and inequality-based TAML. Then, we explored meta imitation learning, which combines meta learning with imitation learning. We saw how meta learning helps imitation learning to learn from fewer imitations. We also saw how to apply model agnostic meta learning in an unsupervised learning setting using CACTUS. Then, we explored a deep meta learning algorithm called learning to learn in concept space. We saw how meta learning can be boosted by the power of deep learning.

Meta learning is one of the most interesting branches in the field of AI; now that you've understood various meta learning algorithms, you can start building meta learning models that are generalizable across various tasks and contribute to meta learning research.