Katja Hofmann, Senior Researcher at Microsoft Research Cambridge

Katja Hofman

Dr Katja Hofmann is a Senior Researcher, and leads a research team that focuses on Gaming and AI at Microsoft Research Cambridge. Her team focuses on reinforcement learning with applications in video games, as she believes that games will drive a transformation of how people interact with AI technology. She is the research lead of Project Malmo, which uses the popular game Minecraft as an experimentation platform for developing intelligent technology – most recently as the platform for the MineRL NeurIPS competition on sample efficient reinforcement learning with human priors. Her long-term goal is to develop AI systems that learn to collaborate with people, to empower their users and help solve complex real-world problems.

Minecraft as AI Playground and Laboratory

Modern video games provide exciting challenges and opportunities for pushing the state of the art in machine learning and other research areas, and, in turn, stand to first benefit from research advances. For driving research, games provide rich data that can be used to tackle hard problems, from complex decision making to collaboration. If and when these are successfully tackled, new algorithms and insights have the potential to enable entirely new game experiences.

This talk focuses on opportunities in the setting of the game Minecraft, one of the most popular video games of all time. Minecraft is an open-world game, where players explore, create, and continuously find new ways to play and engage with each other. This open-ended nature both make the game appealing to its human fan-base, and uniquely challenging to AI algorithms. To unlock the potential of Minecraft for AI experimentation, my team has developed Project Malmo – an open source experimentation platform built on top of Minecraft to enable a wide range of research. Here, I will illustrate the capabilities of the platform with recent examples that I find particularly exciting.

I will highlight our most recent collaboration, led by a team of PhD students at Carnegie Mellon University: the MineRL competition. This ambitious competition is designed to drive advances in sample efficient reinforcement learning with human priors. Sample efficient learning is a key challenge, with current algorithms often requiring millions of samples to learn to perform individual narrow tasks, limiting the scope and applicability of these approaches. This competition is built around a complex task, large-scale demonstration data, and an evaluation setup that requires and rewards sample efficient learning and effective generalization.

Looking out into the future, I will conclude by highlight selected open questions and challenges that have high potential for impact in video games and raise key questions for current state-of-the-art AI approaches.


Adrián Cuevas, Co Founder and Technical Director at Nomada Studio

Adrián Cuevas started working as a programmer for a bio-mechanical company in Barcelona. After a few years he decided to try a new path in the video game industry. His career brought him to companies such as Relentless Software in Brighton, Ubisoft Montreal, Square Enix Copenhagen and Ubisoft Barcelona. During this period he worked in games like FarCry 3, Buzz! The ultimate Music Quiz, Hitman (2016) and Rainbow Six: Siege among others. After many years working in AAA games he decided to partner with Conrad Roset and Roger Mendoza to start new adventure in the indie industry. That is when they started Nomada Studio, where they developed GRIS, a very artistic 2D platformer.

Adrián Cuevas

The creativity process behind GRIS

GRIS is the journey of a girl dealing with the most painful experience in her life. Explaining this story with no words or dialogue has been a great challenge for Nomada Studio. Adrian Cuevas will explain the process of creation for GRIS, showing the different tools, references and ideas that were used during the production of the game. Art, Music, Animation, Game Design, … everything is extremely important to have a consistent narrative and engage the player with a story that will not be told. On top of that GRIS targets all kind of audience, even people that are not used to play video games. This will brought more challenges to the design and construction of the game. Adrian will talk about some of the decisions that were made to make the game more appealing to non-gamers.

Close Menu