FinToolSyn: A forward synthesis Framework for Financial Tool-Use Dialogue Data with Dynamic Tool Retrieval

2026-03-25

AI Summary

This work addresses limitations in existing financial dialogue data synthesis methods, which rely on backward generation paradigms that produce overly explicit queries, lack real-world event-driven dynamics, and struggle to simulate tool retrieval in large-scale tool spaces. To overcome these challenges, we propose a forward synthesis framework that leverages role-guided instructions, atomic tool composition, and dynamic tool retrieval to generate dialogues more representative of authentic financial scenarios. We construct a comprehensive tool library comprising 43,066 functions and synthesize 148,000 high-quality dialogue instances. Furthermore, we establish the first benchmark specifically designed for evaluating financial tool usage. Experimental results demonstrate that models trained on our synthesized data achieve a 21.06% improvement in tool-calling accuracy.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Tool-use capabilities are vital for Large Language Models (LLMs) in finance, a domain characterized by massive investment targets and data-intensive inquiries. However, existing data synthesis methods typically rely on a reverse synthesis paradigm, generating user queries from pre-sampled tools. This approach inevitably introduces artificial explicitness, yielding queries that fail to capture the implicit, event-driven nature of real-world needs. Moreover, its reliance on static tool sets overlooks the dynamic retrieval process required to navigate massive tool spaces. To address these challenges, we introduce \textit{FinToolSyn}, a forward synthesis framework designed to generate high-quality financial dialogues. Progressing from persona instruction and atomic tool synthesis to dynamic retrieval dialogue generation, our pipeline constructs a repository of 43,066 tools and synthesizes over 148k dialogue instances, incorporating dynamic retrieval to emulate the noisy candidate sets typical of massive tool spaces. We also establish a dedicated benchmark to evaluate tool-calling capabilities in realistic financial scenarios. Extensive experiments demonstrate that models trained on FinToolSyn achieve a 21.06\% improvement, providing a robust foundation for tool learning in financial scenarios.

Problem

• tool-use
• financial dialogue
• data synthesis
• dynamic retrieval
• large language models

Innovation

• forward synthesis
• dynamic tool retrieval
• financial tool-use dialogue
• tool-calling benchmark
• LLM tool learning