Monte Carlo Tree Diffusion: A Scalable AI Framework for Long-Horizon Planning

Monte Carlo Tree Diffusion is a scalable AI framework that combines the structured search of Monte Carlo Tree Search with the generative flexibility of diffusion models. This hybrid approach optimizes long-horizon planning by integrating tree-structured rollouts, adaptive exploration-exploitation trade-offs, and efficient denoising methods. It demonstrates superior performance in complex tasks such as maze navigation, robotic cube manipulation, and image-based planning, outperforming existing diffusion and search-based models.