Reinforcement Learning with Sparse Rewards for Procedural Game Content Generation #073295816061 (2025-02-03)