Newest papers

Separate Field Scoring Protects Unused Tools When Agents Pool Tool-Use Experience

Abhijit Chakraborty · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.30867 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

Los agentes que llaman herramientas del mismo catálogo compartido pueden agrupar lo que aprenden sobre qué herramienta sirve a qué solicitud. Medimos qué ejemplos de uso compartido hacen para la recuperación de herramientas, reportando por separado las herramientas retenidas, para las que ningún agente tiene uso, y todas las otras herramientas. Agregando ejemplos de uso al índice como entradas adicionales de texto para las herramientas usadas, se eleva Hit@5 (si una herramienta de oro está en el top cinco) para esas herramientas y se baja para las herramientas retenidas, por un gran margen de trabajo en dos de los tres catálogos y con ambos embedders que hemos probado. Dando a las herramientas usadas el mismo número de entradas igualmente largas llenadas con otras herramientas, las entradas de texto de las herramientas

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

Abstract

Agents that call tools from a shared catalog can pool what they learn about which tool serves which request. We measure what pooled usage examples do to tool retrieval, reporting separately the held-out tools, for which no agent has usage, and all other tools. Adding usage examples to the index as extra entries for used tools raises Hit@5 (whether a gold tool is in the top five) for those tools and lowers it for held-out tools, by a large margin on two of three catalogs and with both embedders we tested. Giving used tools the same number of equally long entries filled with other tools' text lowers held-out Hit@5 about as much, but gives little or none of the gain. We call our method of scoring usage as its own field SYNAPSE. On StableToolBench, the held-out gold tool appears in its top five more often than with a router that reserves slots for usage (0.785 against 0.683), and this held-out lead holds on all three catalogs with both embedders. This protection persisted when we replaced each tool's usage with other tools' text, and a constant score offset for tools with usage gives similar held-out Hit@5 in most settings. The gain for tools with usage, in contrast, needs usage about the right tool. The router keeps more of the gain on labeled queries on TOUCAN with both embedders and on ToolRet with one, so the better choice depends on the workload. Splitting the usage among ten simulated agents shows that pooling helps exactly as much as giving one agent the same records.

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

Topics: Pooling, Computer science, Offset (computer science), Router, Field (mathematics), Margin (machine learning)

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