Uber AI ‘reliably’ completes all stages in Montezuma’s Revenge

Montezuma’s Revenge is a notoriously difficult video game for humans, much less artificial intelligence (AI), to beat — the first level alone consists of 24 rooms filled with traps, ropes, ladders, enemies, and concealed keys. But recently, AI systems from OpenAI, Google’s DeepMind, and others have managed to make impressive gains. And this week, new research from Uber raises the bar higher still.

In a blog post and forthcoming paper, AI scientists at Uber describe Go-Explore, a family of so-called quality diversity AI models capable of achieving maximum scores of over 2,000,000 on Montezuma’s Revenge and average scores over 400,000. (That’s compared to the current state-of-the-art model’s average and maximum score of 10,070 and 17,500, respectively.) Furthermore, in testing, the models were able to “reliably” solve the entire game up to level 159.