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.
Back to articles
AIOpenAI News
Learning Montezuma’s Revenge from a single demonstration
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 game...
This source only provides an excerpt in its RSS feed. FlowMarket displays all content available from the feed and keeps the original publication link for attribution.
Need an n8n workflow or help installing it?
After the briefing, move to execution: find an n8n template or a creator who can adapt it to your tools.