The data for the Malmo reinforcement learning environment includes the image frames that the agent is moving in. A sample gaming window for Malmo will look like the following. Here, the agent dies if they step over the lava:
Malmo requires developers to specify the XML schema in order to generate the mission. We will need to create mission data for both the agent and the server to create blocks in the world (that is, the gaming environment). In this recipe, we will create an XML schema to specify the mission data.