This repository is a curated list of state-of-the-art research in embodied AI, focusing on vision-language-action (VLA), vision-language navigation (VLN), and related multimodal learning approaches. It serves as a comprehensive resource for researchers and practitioners in robotics and AI, aiming to track and organize advancements in the field, particularly those inspired by the "LLM moment" in natural language processing.

How It Works

The repository categorizes research into Vision-Language-Action (VLA) Models, Vision-Language Navigation (VLN) Models, Vision-Action (VA) Models, and other Multimodal Large Language Model (MLLM)-based embodied learning. It compiles papers, projects, and datasets, providing links to official resources for each entry. The organization aims to reflect the rapid evolution of multimodal AI in robotics.

Quick Start & Requirements

This is a curated list, not a runnable codebase. To engage with the research, users would typically need to clone individual project repositories, install their specific dependencies (often Python, PyTorch, TensorFlow, and specialized robotics libraries), and potentially set up simulation environments or access hardware.

Highlighted Details

• Extensive collection of papers and projects from 2022-2025, covering foundational models, diffusion policies, and LLM integration.
• Includes dedicated sections for Vision-Language-Action (VLA), Vision-Language Navigation (VLN), and Vision-Action (VA) models.
• Features links to relevant surveys, benchmarks, simulators, and related "awesome" lists for broader context.
• Actively seeks community contributions to maintain its comprehensiveness.

Maintenance & Community

The repository is maintained by "jonyzhang2023" and welcomes community contributions via pull requests or issues. It aims to be a continuously updated resource.

Licensing & Compatibility

As a curated list of external research, this repository itself does not have a specific license. The licensing of individual projects linked within the list would vary and must be checked on their respective repositories.

Limitations & Caveats

This repository is a passive index of research papers and projects; it does not provide any executable code or direct functionality. Users must individually locate, download, and set up the code for each research project they wish to explore.