XLand-MiniGrid

'XLand-MiniGrid' is a benchmark dataset that contains a variety of grid-based environments used to evaluate the exploration and adaptation capabilities of reinforcement learning agents in the context of unsupervised learning and goal-setting.

Papers using XLand-MiniGrid (4)

Unsupervised Learning of Efficient Exploration: Pre-training Adaptive Policies via Self-Imposed Goals2026NiceWebRL: a Python library for human subject experiments with reinforcement learning environments2025Yes, Q-learning Helps Offline In-Context RL2025XLand-100B: A Large-Scale Multi-Task Dataset for In-Context Reinforcement Learning2024