# FinBERT: A Pretrained Language Model for Financial Communications

**Authors:** Yi Yang, Mark Christopher Siy UY, Allen Huang  
**arXiv:** [2006.08097](https://arxiv.org/abs/2006.08097)  
**Date:** 2020-06-15  
**Summary page:** https://www.summarizepaper.com/en/paper/2006.08097v2/

## Key points

- Authors Yi Yang, Mark Christopher Siy UY, and Allen Huang introduce FinBERT, a financial domain-specific BERT model trained on financial communication text.
- FinBERT aims to address challenges in NLP tasks specific to the financial sector due to the lack of pretrained finance-specific language models.
- The authors conducted experiments showing FinBERT's superiority over generic domain BERT models in three financial sentiment classification tasks.
- FinBERT outperforms generic BERT models in financial NLP tasks and provides valuable resources for practitioners and researchers with publicly available code and pretrained models on GitHub.
- By bridging the gap between general-purpose language models and industry-specific requirements, FinBERT enhances understanding of financial communications and sets a benchmark for future developments in domain-specific pretrained models.

## Summary

In their work titled "FinBERT: A Pretrained Language Model for Financial Communications," authors Yi Yang, Mark Christopher Siy UY, and Allen Huang introduce FinBERT, a financial domain-specific BERT model trained on a large corpus of financial communication text. This model aims to address the unique challenges posed by the financial sector in NLP tasks due to the lack of pretrained finance-specific language models. The authors conducted experiments on three financial sentiment classification tasks and demonstrated FinBERT's superiority over generic domain BERT models. Their study confirms that FinBERT outperforms generic BERT models in financial NLP tasks and provides a valuable resource for practitioners and researchers with its publicly available code and pretrained models on GitHub. By bridging the gap between general-purpose language models and industry-specific requirements, FinBERT not only enhances the capabilities of understanding financial communications but also sets a benchmark for future developments in domain-specific pretrained models. opens up new possibilities for more accurate and efficient analysis of financial texts.