FinLLMs: A Framework for Financial Reasoning Dataset Generation with Large Language Models

TL;DR

FinLLMs introduces a novel framework that leverages large language models to automatically generate financial question-answering datasets, reducing manual annotation costs and improving model performance in financial reasoning tasks.

Contribution

The paper presents a new method for generating financial QA data using LLMs based on financial formulas and variable graphs, enhancing data resources efficiently.

Findings

Synthetic data improves financial reasoning model performance.

FinLLMs outperforms existing benchmark datasets.

Graph-based formula augmentation enhances data diversity.

Abstract

Large Language models (LLMs) usually rely on extensive training datasets. In the financial domain, creating numerical reasoning datasets that include a mix of tables and long text often involves substantial manual annotation expenses. To address the limited data resources and reduce the annotation cost, we introduce FinLLMs, a method for generating financial question-answering data based on common financial formulas using Large Language Models. First, we compile a list of common financial formulas and construct a graph based on the variables these formulas employ. We then augment the formula set by combining those that share identical variables as new elements. Specifically, we explore formulas obtained by manual annotation and merge those formulas with shared variables by traversing the constructed graph. Finally, utilizing GPT-3.5, we generate financial question-answering data that encompasses both tabular information and long textual content, building on the collected formula set. Our experiments demonstrate that synthetic data generated by FinLLMs effectively enhances the performance of several large-scale numerical reasoning models in the financial domain, outperforming two established benchmark financial question-answering datasets.