Go-Explore: a New Approach for Hard-Exploration Problems


    A grand challenge in reinforcement learning is intelligent exploration, especially when rewards are sparse or deceptive. Two Atari games serve as benchmarks for such hard-exploration domains: Montezuma’s Revenge and Pitfall. On both games, current RL algorithms perform poorly, even those with intrinsic motivation, which is the dominant method to improve performance on hard-exploration domains. To address this shortfall, we introduce a new algorithm called Go-Explore. It exploits the following principles: (1) remember previously visited states, (2) first return to a promising state (without exploration), then explore from it, and (3) solve simulated environments through any available means (including by introducing determinism), then robustify via imitation learning. The combined effect of these principles is a dramatic performance improvement on hard-exploration problems. On Montezuma’s Revenge, Go-Explore scores a mean of over 43k points, almost 4 times the previous state of the art. Go-Explore can also harness human-provided domain knowledge and, when augmented with it, scores a mean of over 650k points on Montezuma’s Revenge. Its max performance of nearly 18 million surpasses the human world record, meeting even the strictest definition of “superhuman” performance. On Pitfall, Go-Explore with domain knowledge is the first algorithm to score above zero. Its mean score of almost 60k points exceeds expert human performance. Because Go-Explore produces high-performing demonstrations automatically and cheaply, it also outperforms imitation learning work where humans provide solution demonstrations. Go-Explore opens up many new research directions into improving it and weaving its insights into current RL algorithms. It may also enable progress on previously unsolvable hard-exploration problems in many domains, especially those that harness a simulator during training (e.g. robotics).


    Uber AI


    Adrien Ecoffet, Joost Huizinga, Joel Lehman, Kenneth O. Stanley, Jeff Clune



    Full Paper

    ‘Go-Explore: a New Approach for Hard-Exploration Problems’ (PDF)

    Previous articleAn Atari Model Zoo for Analyzing, Visualizing, and Comparing Deep Reinforcement Learning Agents
    Adrien Ecoffet
    Adrien Ecoffet is a research scientist with Uber AI Labs.
    Joost Huizinga
    Joost Huizinga is a research scientist with Uber AI.
    Joel Lehman
    Joel Lehman was previously an assistant professor at the IT University of Copenhagen, and researches neural networks, evolutionary algorithms, and reinforcement learning.
    Kenneth O. Stanley
    Before joining Uber AI Labs full time, Ken was an associate professor of computer science at the University of Central Florida (he is currently on leave). He is a leader in neuroevolution (combining neural networks with evolutionary techniques), where he helped invent prominent algorithms such as NEAT, CPPNs, HyperNEAT, and novelty search. His ideas have also reached a broader audience through the recent popular science book, Why Greatness Cannot Be Planned: The Myth of the Objective.
    Jeff Clune
    Jeff is on leave from the University of Wyoming, where he is the Loy and Edith Harris Associate Professor in Computer Science and directs the Evolving AI Lab (http://EvolvingAI.org). He researches robotics and creating artificial intelligence in neural networks, either via deep learning or evolutionary algorithms. In the last three years a robotics paper he coauthored was on the cover of Nature, he won an NSF CAREER award, he received the Distinguished Young Investigator Award from the International Society for Artificial Life, and deep learning papers he coauthored were awarded oral presentations at NIPS, CVPR, ICLR, and an ICML workshop.