mbreuss/diffusion-literature-for-robotics

Organizes diffusion model papers across nine robotics application areas—from imitation learning and video prediction to task planning and pose estimation—with curated foundational theory resources and links to open-source implementations. Covers both classical diffusion theory (score-based modeling, denoising probabilistic models) and robotics-specific adaptations including equivariant policies, offline reinforcement learning, and multimodal behavior learning. Serves as a structured entry point for researchers implementing diffusion-based robot controllers, with guided progression from theoretical foundations through domain-specific applications.

905 stars. No commits in the last 6 months.