Meta Reinforcement Learning and Hard-Coded Behaviors

This chapter explores the concept of meta reinforcement learning (meta RL) and the relevance of hard-coded behaviors in animals. It discusses examples of instinctive or heuristic behaviors in animals and describes two example cases in training neural network agents. The chapter also discusses the limitations of fruit flies in developing language and culture, as well as the concept of learning from culture and teachers.