Newest papers

Beyond Accuracy: Disentangling Decomposition from Typed-State Quality Assurance in Multi-Agent LLM Pipelines

Maria Gabriela Alves Zuppardo, Vinícius Cardoso Garcia · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31280 · cited by 0 works (OpenAlex) · open-access PDF

French

Les pipelines d'architecture multi-agent sont souvent créditées avec des gains dans les tâches d'ingénierie logicielle, mais il est rarement clair si un gain provient du contrat de type entre agents ou des prompts spécialisés que la décomposition permet. Nous étudions ce problème d'attribution dans Requirements Engineering, en comparant trois architectures pour classer les exigences en tant que fonctionnelles ou non fonctionnelles, puis en les priorisant avec MoSCoW: un modèle de petite langue hébergé localement; une variante à deux appels qui réutilise les prompts spécialisés de la pipeline mais passe un intermédiaire non validé; et un pipeline avec un état de type validé par Pydantic. Les 18 conditions couvrent l'étude PROMISE NFR

Machine translation (Meta M2M100); the English original follows.

Abstract

Multi-agent LLM pipelines are often credited with gains in software engineering tasks, but it is rarely clear whether a gain stems from the typed contract between agents or from the specialized prompts that decomposition enables. We study this attribution problem in Requirements Engineering, comparing three architectures for classifying requirements as functional or non-functional and then prioritizing them with MoSCoW: a single-call baseline; a two-call variant that reuses the pipeline's specialized prompts but passes an unvalidated intermediate; and a pipeline with a Pydantic-validated typed state. The 18 conditions cover PROMISE NFR+ (625 requirements), three locally hosted small language models, and two languages. The F1 gain over the baseline is associated with the two-call, specialized-prompt variant (up to +0.075), whereas the typed state adds a small difference (-0.003 to +0.013) that is significant in one of six comparisons. Beyond accuracy, the typed pipeline contains out-of-range model outputs at the agent boundary, while the unvalidated variant fails later and drops the requirement, and it shows a small, reproducible increase in inter-model MoSCoW consistency whose cause our design cannot isolate. The study is associative: the contrasts bundle prompts, call count, and staged context, and the results come from three 7--8B models on one dataset. The implementation and replication package are archived on Zenodo (https://doi.org/10.5281/zenodo.23380444).

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Pipeline transport, Pipeline (software), Computer science, Consistency (knowledge bases), Quality assurance, Decomposition

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange