Learning Controllable 3D Level Generators
This work addresses procedural content generation for 3D environments like Minecraft, which is incremental as it extends PCGRL to more complex domains.
The paper tackled the problem of generating 3D levels in Minecraft using Procedural Content Generation via Reinforcement Learning (PCGRL), resulting in an agent that produces relatively complex and diverse levels and generalizes to random initial states and control targets.
Procedural Content Generation via Reinforcement Learning (PCGRL) foregoes the need for large human-authored data-sets and allows agents to train explicitly on functional constraints, using computable, user-defined measures of quality instead of target output. We explore the application of PCGRL to 3D domains, in which content-generation tasks naturally have greater complexity and potential pertinence to real-world applications. Here, we introduce several PCGRL tasks for the 3D domain, Minecraft (Mojang Studios, 2009). These tasks will challenge RL-based generators using affordances often found in 3D environments, such as jumping, multiple dimensional movement, and gravity. We train an agent to optimize each of these tasks to explore the capabilities of previous research in PCGRL. This agent is able to generate relatively complex and diverse levels, and generalize to random initial states and control targets. Controllability tests in the presented tasks demonstrate their utility to analyze success and failure for 3D generators.