A field guide to diffusion models for roboticists

What it does This is a curated bibliography and reading list that collects foundational diffusion theory papers, introductory blog posts, and video tutorials alongside a categorized index of robotics research. The robotics section splits work into nine areas—including imitation learning, video diffusion, offline RL, world models, and tactile sensing—and links out to relevant codebases. It is essentially a literature review maintained as a living markdown document.

The interesting bit Instead of dumping every paper into a single list, the maintainer organizes them by where they sit in a robot stack: policy learning, motion planning, even robot hardware design. That structure turns a noisy, fast-moving niche into something a practitioner can actually browse without drowning in arXiv abstracts.

Key highlights

• Foundational theory (Sohl-Dickstein 2015, Ho 2020, Song 2020) sits next to robotics application papers.
• Robotics papers are grouped into nine practical categories, from inverse RL to task and motion planning.
• Includes a section linking to code implementations for starting diffusion robotics projects.
• The maintainer notes the literature is still “small, but growing rapidly” and aims to capture most published conference and arXiv work.
• Direct paper links with venue names are provided throughout.

Caveats

• The README source provided is truncated mid-entry, suggesting the list may be incomplete or actively expanding.
• At least one duplicate paper entry appears in the imitation-learning section, hinting at manual curation without automated checks.
• Discovery is limited to scrolling a long markdown file; there is no search or filtering beyond the table of contents.

Verdict Bookmark this if you are a robotics researcher or engineer trying to map where diffusion models fit into control, planning, or manipulation. Look elsewhere if you need a software library; this is pure curation, not code.