I will discuss a research program aimed at building a Minecraft assistant (or minion, playmate, etc.) The scientific meta-goal of this program is to facilitate the study of agents that can complete tasks specified by dialogue, and eventually, to learn from interactions with players. I will describe the tools and platform we have built, allowing agents and people to play Minecraft together, and the data we have collected. In addition, I will describe an initial agent from which we (and hopefully others) can iterate.
Arthur Szlam is a research scientist at Facebook AI Research. He was previously on the mathematics faculty of the City College of New York, and more previously, earned his PhD in mathematics from Yale University.