MazeBase: A Sandbox for Learning from Games

Abstract

This paper introduces MazeBase: an environment for simple 2D games, designed as a sandbox for machine learning approaches to reasoning and planning. Within it, we create 10 simple games embodying a range of algorithmic tasks (e.g. if-then statements or set negation). A variety of neural models (fully connected, convolutional network, memory network) are deployed via reinforcement learning on these games, with and without a procedurally generated curriculum. Despite the tasks' simplicity, the performance of the models is far from optimal, suggesting directions for future development. We also demonstrate the versatility of MazeBase by using it to emulate small combat scenarios from StarCraft. Models trained on the MazeBase version can be directly applied to StarCraft, where they consistently beat the in-game AI.

References in corpus (2)

• Inferring Algorithmic Patterns with Stack-Augmented Recurrent Nets
• Learning Simple Algorithms from Examples

Cited by in corpus (27)

• Learning Multiagent Communication with Backpropagation
• Control of Memory, Active Perception, and Action in Minecraft
• Zero-Shot Task Generalization with Multi-Task Deep Reinforcement Learning
• TextWorld: A Learning Environment for Text-based Games
• Deep Successor Reinforcement Learning
• Dialog-based Language Learning
• Modeling Others using Oneself in Multi-Agent Reinforcement Learning
• Deep Hierarchical Reinforcement Learning Algorithm in Partially Observable Markov Decision Processes
• ShapeWorld - A new test methodology for multimodal language understanding
• Revisiting the Master-Slave Architecture in Multi-Agent Deep Reinforcement Learning
• Disentangling the independently controllable factors of variation by interacting with the world
• Composable Planning with Attributes
• ACCNet: Actor-Coordinator-Critic Net for "Learning-to-Communicate" with Deep Multi-agent Reinforcement Learning
• Hierarchical Decision Making by Generating and Following Natural Language Instructions
• Decoupling Dynamics and Reward for Transfer Learning
• Interactive Grounded Language Acquisition and Generalization in a 2D World
• Virtual Embodiment: A Scalable Long-Term Strategy for Artificial Intelligence Research
• Value Propagation Networks
• Qualitative Numeric Planning: Reductions and Complexity
• Why Build an Assistant in Minecraft?
• SoundSpaces: Audio-Visual Navigation in 3D Environments
• Language Expansion In Text-Based Games
• Not All Memories are Created Equal: Learning to Forget by Expiring
• Mastering the Dungeon: Grounded Language Learning by Mechanical Turker Descent
• Continual and Multi-task Reinforcement Learning With Shared Episodic Memory
• APES: a Python toolbox for simulating reinforcement learning environments
• Planning with Arithmetic and Geometric Attributes