Procedural Graphs: Self-Evolving Execution Structures for LLM Agents
Researchers introduce Procedural Graphs, self-evolving (procedure, relation, procedure) structures guiding LLM agent tool use and planning.
Procedural Graphs organize procedural knowledge into (procedure, relation, procedure) triplets to guide LLM agent actions, addressing drift such as lost objectives, out-of-order tool calls, and repeated unproductive steps. At each decision step the framework localizes the active node and a guidance model translates the surrounding subgraph into step-level situational guidance. An LLM refiner edits graph topology by contrasting failed with successful trajectories, and across datasets, task types and LLMs the approach outperforms memory-based baselines and matches or surpasses hand-designed graphs.
- Organizes procedural knowledge into (procedure, relation, procedure) triplets guiding step-level agent decisions.
- Self-evolving LLM refiner contrasts failed and successful trajectories to edit graph topology.
- Consistently outperforms memory-based baselines and can repair flawed expert priors.
- Builds graphs matching or surpassing hand-designed ones starting from a minimal skeleton.
From the original article
1 min read at arxiv.orgLarge language models are increasingly deployed as agents that plan over long horizons and act through external tools. Most agents select actions through unconstrained generation over an accumulating history, leaving implicit the procedural knowledge of what to do, in what order, and under which conditions.
By Yuxing Lu · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.