Imitation learning, also known as learning by demonstration, is an approach being used in an esports project. There are two approaches to imitation learning: behavioral cloning captures and reproduces human skills by recording a specialist's demonstrations of a task, while inverse reinforcement learning develops a reward function from observed agent behavior to understand the goal. Imitation learning is being used over reinforcement learning for the esports project because specifying an accurate reward function is difficult, and the goal is for an agent to perform like a human.