LATIDIA · Investigación

Jev-LiteMAS: Separar las decisiones de la generación en flujos de trabajo multiagente

arXiv: 2610.21411v1Tipo de anuncio: nuevo Resumen: Los sistemas LLM multiagente se basan en modelos generativos para la ejecución experta, la planificación del flujo de trabajo y la revisión de artefactos. Sin embargo, muchas operaciones de coordinación solo requieren decisiones limitadas

La noticia

arXiv:2610.21411v1 Announce Type: new Abstract: Multi-agent LLM systems rely on generative models for expert execution, workflow planning, and artifact review. Yet many coordination operations require only bounded decisions, such as selecting the next expert or judging artifact quality. Profiling three representative MAS frameworks, we find that routing and review account for 31.7-43.5% of generative-LLM tokens. However, replacing these operations with low-cost decision models introduces two interface gaps: generative planners produce execution plans and handoff instructions (planning gap), while reviewers provide actionable feedback for repairing flawed artifacts (review gap), which bounded decisions alone cannot provide. We introduce Jev-LiteMAS, a multi-agent framework that bridges both gaps by combining a lightweight decision model