Procedural Graphs: Self-Evolving Execution Structures for LLM Agents
Researchers from Google, Georgia Tech, and Peking University have introduced 'Procedural Graphs,' a method for LLM agents to store and evolve execution steps. This approach allows agents to learn from failed trajectories and improve their planning capabilities over time.
Why it matters
It addresses a core limitation in current AI agents—the inability to reliably plan and adapt over long horizons—by providing a structured, self-correcting memory mechanism.
Yuxing Lu, Yicheng Chen, Shanchan Wu, Sercan Ö. Arık
Chat with Paper First page The curator’s take Yuxing Lu, Yicheng Chen, Shanchan Wu and Sercan Arik at Google, with Georgia Tech and Peking University, introduce the Procedural Graph, an explicit store of (procedure, relation, procedure) triplets that supplies step-level guidance to an agent and rewrites itself from the difference between failed and successful trajectories.
Ask this paper Question about this paper Ask in Paper Chat Key points 01 What the graph stores: A knowledge graph holds facts as entity-relation-entity triplets; a Procedural Graph holds procedures as procedure-relation-procedure triplets, so the agent has a queryable representation of what to do next and in what order rather than leaving that implicit in a growing history.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in