jonyzhang2023/awesome-embodied-vla-va-vln

What it solves

This project addresses the challenge of navigating the rapidly evolving field of Embodied AI. As research in robotics shifts toward large-scale foundation models, it can be difficult for researchers and developers to track the latest breakthroughs in Vision-Language-Action (VLA) models, world models, and robotic navigation.

How it works

The repository serves as a curated, community-driven index of state-of-the-art research. It organizes hundreds of academic papers, technical reports, and open-source codebases into logical categories such as VLA models, physics-aware policies, simulation platforms, and evaluation benchmarks. The list is maintained by the community, with new entries added via pull requests to ensure the collection remains current.

Who it’s for

This resource is designed for robotics researchers, AI engineers, and students who want to stay informed about the "LLM moment" in robotics. It is particularly useful for those looking to implement, study, or build upon the latest foundation models for robot learning and manipulation.

Highlights

• Comprehensive survey papers covering the landscape of embodied AI.
• Extensive lists of Vision-Language-Action (VLA) and World Action Models (WAM).
• Dedicated sections for Vision-Language Navigation (VLN) and Vision-Action (VA) models.
• Resources for Sim-to-Real transfer, physics-aware policies, and evaluation benchmarks.
• Active maintenance with reverse-chronological ordering to highlight the newest developments.