Citi interview question

What is batch normalisation? What are ways to optimise a reinforcement learning algorithm?