FinGen: A Dataset for Argument Generation in Finance

TL;DR

This work addresses the gap in finance NLP for forward-looking argument generation by introducing FinGen, a dataset supporting three tasks: Evidence2Claim, Chart2Argument, and News2Argument. It evaluates multiple pretrained models and introduces a keyterm-guided method to emphasize numerals and financial terms in financial narratives. Findings show the tasks remain challenging, with fluent outputs but limitations in numeral accuracy, term chaining, and scenario planning; performance varies by task and input modality. The authors discuss future directions—forward-looking claim generation templates and scenario planning—and acknowledge limitations such as multilingual scope and dataset sizes, outlining a roadmap for advancing finance-oriented argument generation.

Abstract

Thinking about the future is one of the important activities that people do in daily life. Futurists also pay a lot of effort into figuring out possible scenarios for the future. We argue that the exploration of this direction is still in an early stage in the NLP research. To this end, we propose three argument generation tasks in the financial application scenario. Our experimental results show these tasks are still big challenges for representative generation models. Based on our empirical results, we further point out several unresolved issues and challenges in this research direction.

Figures (1)

• Figure 1: Example input of Chart2Argument task.