The Problem

Current methods in Deep RL are sample inefficient, especially as we move to more and more complex domains: OpenAI Five collected 900 years of experience per day and AlphaGoZero played 4.9 million games of Go. Furthermore, there has been recent success in leveraging imitation learning to solve older benchmarks like Atari, as well as real-world problems such as robotic manipulation and self driving cars. More recently, AlphaStar was able to achieve Gold/Platinum MMR (~50% percentile human performance) using pretraining alone. We believe that leveraging human data will be an important piece of the puzzle as we tackle sample efficiency in more and more complex problems.

However, despite the increasing number of benchmark RL environments, there has been a lack of large-scale imitation learning datasets. Thus, we release MineRL: a large-scale dataset on Minecraft of seven different tasks, which highlight a variety of research challenges including open-world multi-agent interactions, long-term planning, vision, control, navigation, and explicit and implicit subtask hierarchies. We also release a platform on which to collect more data: a Minecraft mod that records human trajectories, a public Minecraft server with minigames/Survival play with this mod, and a flexible framework to define new Minecraft tasks.

Why Minecraft?

Minecraft is a rich environment to do RL on: it is an open-world environment, has sparse rewards, and has many innate task hierarchies and subgoals. Furthermore, it encompasses many of the problems that we must solve as we move towards more general AI (for example, what is the reward structure of “building a house”?). Besides all this, Minecraft has more than 90 million monthly active users, making it a good environment on which to collect a large-scale dataset.


MineRL: A Large-Scale Dataset of Minecraft Demonstrations

William H Guss, Brandon Houghton, Nicholay Topin, Phillip Wang, Cayden Codel, Manuela Veloso, Ruslan Salakhutdinov

Twenty-Eighth International Joint Conference on Artificial Intelligence


[BibTex] [Dataset Details]

The MineRL Competition on Sample Efficient Reinforcement Learning using Human Priors

William H. Guss, Cayden Codel, Katja Hofmann, Brandon Houghton, Noboru Kuno, Stephanie Milani, Sharada Mohanty, Diego Perez Liebana, Ruslan Salakhutdinov, Nicholay Topin, Manuela Veloso, Phillip Wang

NeurIPS 2019 Competition Track


[BibTex] [Competition Details]