What is awesome-wam?

Awesome-WAM is a curated reading list and resource hub for researchers following a specific area of AI research called World Action Models, or WAMs. The broader field it covers is embodied AI, the study of AI systems that can perceive an environment, predict what will happen next, and take actions in the real world, like robotic arms or autonomous agents. A World Action Model is a system that combines two capabilities: predicting future states of the world, called a world model, and deciding what actions to take, called an action model. This repository organizes and summarizes the academic papers exploring how to build these systems, covering different architectural approaches, for example, whether the prediction and action parts are built separately or jointly, and whether they use autoregressive or diffusion-based generation methods. Beyond just listing papers, the repository includes structured blog-style summaries of each paper, so readers can quickly understand the key ideas without reading the full academic text. The summarization method used to produce those write-ups is also included in the repository. You would use this resource if you are a researcher, student, or practitioner trying to keep up with developments in embodied AI and robotic learning. The repository is community-driven and continuously updated as new work is published. The list organizes papers under a set of tags that describe how each system is built, such as whether it uses an explicit pixel-space representation or an implicit latent representation, and whether the action generation is autoregressive or diffusion-based. It also covers work on using world models to support a related area called vision-language-action learning, including imitation learning, reinforcement learning, and evaluation methods, along with a section on the training data used across these approaches. Contributors are invited to open an issue or a pull request if they find a paper missing from the list. The full README is longer than what was shown.