Source record

Published by arXiv cs.MA (T1 source). The original is at https://arxiv.org/abs/2610.07663.

Pipeline notes

The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.

SWORD is a framework that jointly optimizes multi-agent workflow topology and natural-language prompts for user behavior simulation, guided only by a scalar task metric. It outperforms prompt-only, workflow-only, and staged-optimization baselines at low cost ($4–$6 per dataset) and surfaces domain-relevant features from error feedback alone.

A single-metric joint optimizer over workflow topology and prompts, with a documented side effect of unsupervised feature discovery — useful pattern for agent pipeline tuning.

Cited by

No citations on record.