Learning Montezuma’s Revenge from a single demonstration
OpenAI Blog
An AI agent learned to play Montezuma's Revenge from a single human demonstration, achieving a score of 74,500. The agent was trained by replaying games from states extracted from that single demonstration and optimizing performance using PPO reinforcement learning. This outperforms previous methods that required multiple demonstrations or other learning approaches to master the same game.
Why it matters
We’ve trained an agent to achieve a high score of 74,500 on Montezuma’s Revenge from a single human demonstration, better than any previously published result. Our algorithm is simple: the agent plays a sequence of games starting from carefully chosen states from the demonstration, and learns from them by optimizing the game score using PPO, the same reinforcement learning algorithm that underpins OpenAI Five.
