Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Large Language Models in Finance: A Survey

Yinheng Li, Shaofei Wang, Han Ding, Hang Chen

TL;DR

This survey addresses how LLMs can transform finance by surveying existing approaches, such as zero-shot/few-shot use, fine-tuning, and domain-specific pretraining, and by offering a decision framework that guides practitioners from lightweight experimentation to full-scale customization. It highlights improvements in finance classification tasks from fine-tuned LLMs and strong performance in finance-specific pretraining like BloombergGPT, while noting limitations in generative capabilities and the need for high-quality domain data. The paper also discusses risk, governance, and ethical considerations, proposing retrieval augmentation and tool integration as practical mitigations. Overall, it provides a roadmap for responsibly applying LLMs to finance, with concrete guidance on costs, data requirements, and evaluation.

Abstract

Recent advances in large language models (LLMs) have opened new possibilities for artificial intelligence applications in finance. In this paper, we provide a practical survey focused on two key aspects of utilizing LLMs for financial tasks: existing solutions and guidance for adoption. First, we review current approaches employing LLMs in finance, including leveraging pretrained models via zero-shot or few-shot learning, fine-tuning on domain-specific data, and training custom LLMs from scratch. We summarize key models and evaluate their performance improvements on financial natural language processing tasks. Second, we propose a decision framework to guide financial professionals in selecting the appropriate LLM solution based on their use case constraints around data, compute, and performance needs. The framework provides a pathway from lightweight experimentation to heavy investment in customized LLMs. Lastly, we discuss limitations and challenges around leveraging LLMs in financial applications. Overall, this survey aims to synthesize the state-of-the-art and provide a roadmap for responsibly applying LLMs to advance financial AI.

Large Language Models in Finance: A Survey

TL;DR

This survey addresses how LLMs can transform finance by surveying existing approaches, such as zero-shot/few-shot use, fine-tuning, and domain-specific pretraining, and by offering a decision framework that guides practitioners from lightweight experimentation to full-scale customization. It highlights improvements in finance classification tasks from fine-tuned LLMs and strong performance in finance-specific pretraining like BloombergGPT, while noting limitations in generative capabilities and the need for high-quality domain data. The paper also discusses risk, governance, and ethical considerations, proposing retrieval augmentation and tool integration as practical mitigations. Overall, it provides a roadmap for responsibly applying LLMs to finance, with concrete guidance on costs, data requirements, and evaluation.

Abstract

Recent advances in large language models (LLMs) have opened new possibilities for artificial intelligence applications in finance. In this paper, we provide a practical survey focused on two key aspects of utilizing LLMs for financial tasks: existing solutions and guidance for adoption. First, we review current approaches employing LLMs in finance, including leveraging pretrained models via zero-shot or few-shot learning, fine-tuning on domain-specific data, and training custom LLMs from scratch. We summarize key models and evaluate their performance improvements on financial natural language processing tasks. Second, we propose a decision framework to guide financial professionals in selecting the appropriate LLM solution based on their use case constraints around data, compute, and performance needs. The framework provides a pathway from lightweight experimentation to heavy investment in customized LLMs. Lastly, we discuss limitations and challenges around leveraging LLMs in financial applications. Overall, this survey aims to synthesize the state-of-the-art and provide a roadmap for responsibly applying LLMs to advance financial AI.
Paper Structure (21 sections, 1 equation, 1 figure, 3 tables)

This paper contains 21 sections, 1 equation, 1 figure, 3 tables.

Table of Contents

  1. Introduction
  2. Basics of Language Models
  3. Overview of AI Applications in Finance
  4. Current AI Applications in Finance
  5. Advancements of LLMs in Finance
  6. LLM Solutions for Finance
  7. Utilizing Few-shot/Zero-shot Learning in Finance Applications
  8. Fine-tuning a Model
  9. Common Techniques for LLM Fine-tuning
  10. Fine-tuned finance LLM evaluation
  11. Pretrain from Scratch
  12. Decision Process in Applying LLM to Financial Applications
  13. Determining the Need for a LLM
  14. A general decision guidance for applying LLMs on finance tasks
  15. Level 1: Zero-shot Applications
  16. ...and 6 more sections

Figures (1)

  • Figure 1: Decision process flow chart