The Malmo Collaborative AI Challenge

Please complete the registration form. Before submitting your registration, you will need to set up a repository on GitHub because you will have to provide a valid link during the registration process.

Make sure that your team’s registration form is complete before submitting. Once submitted to Microsoft via your registration form, any changes to your GitHub repository and registration details will be accepted at the sole discretion of Microsoft. 

Your Challenge submission will be the version of the code on your GitHub repository branch master at the time of the entry submission deadline. Your submission needs to contain code in working order and be accompanied by a one-page description of your approach and a link to a one-to-three-minute video that shows off your agent.

The example code is provided as an optional starting point, and you do not need to use it in your submission. However, note that the Challenge task definition cannot be changed. 

Malmo clients can be run on Windows, Linux, and Mac. We also provide ready-to-build Docker images that can be run on any platform supported by docker.io. The Challenge has been developed around Python—both 2.7 and >= 3.4 are compatible.  

Yes, we encourage academic publications relating to the Malmo Collaborative AI Challenge. If you do decide to publish results, please cite the Malmo Platform as:

Johnson M., Hofmann K., Hutton T., Bignell D. (2016) The Malmo Platform for Artificial Intelligence Experimentation. Proc. 25th International Joint Conference on Artificial Intelligence, Ed. Kambhampati S., p. 4246. AAAI Press, Palo Alto, California USA.

Microsoft is offering the following prizes: (1) three winning teams will be awarded placements at its AI Summer School at Microsoft Research Cambridge, UK, and (2) three winning teams will be awarded Microsoft Azure for Research grants with a maximum value of $20,000 USD. Please read the Challenge Rules on this page for further details. 

You can direct administrative questions to [email protected]. For technical questions, we encourage you to use the Microsoft Malmo forum. We endeavor to respond within two business days.

Microsoft Azure is the Microsoft cloud computing service. Azure provides a wide range of services and supports a wide variety of popular platforms, tools and programming languages. It is ideal for deep learning using GPUs and for running machine learning models. As part of the Malmo Collaborative AI Challenge we provide example code for running experiments using Docker on Azure.

The Microsoft Azure for Research program awards cloud computing time, training, and resources for the purpose of conducting cloud-based research. Qualified applicants must be affiliated with a university or nonprofit research organization.

During registration, you’ll be asked “Do you want to use a FREE Microsoft Azure Pass?” If you choose “Yes,” we will contact you within two business days to set this up. Note that these passes are available only to the first 150 entries under the precondition that your team’s registration form is complete and the terms and conditions stated in the Challenge Rules apply.

The EEA (European Economic Area) includes EU countries and also Iceland, Liechtenstein, and Norway. The EU countries are: Austria, Belgium, Bulgaria, Croatia, Republic of Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, and the UK.

During registration, you will be asked whether you wish to apply to win the respective prize. Please answer accordingly. Note that the AI Summer School prize is restricted to participants from EEA countries plus Switzerland and the Azure for Research Grant prize is restricted to participants who are affiliated with a university or nonprofit research organization. Please check the Challenge Rules tab on this page.

At the close of the Entry Period, the judges will select winning teams from eligible entries based upon the following criteria:

1. Innovativeness/originality of the solution; and
2. Overall performance on the Challenge tasks (i.e. game score); and
3. Quality of code developed; and
4. Public voting on GitHub (i.e. GitHub “stars”).

The judges will give equal weighting to each of the above criteria. The judges will be employees of Microsoft Corporation or one of its affiliated companies. For further details please check the Challenge Rules.