# Fine-tune BERT for Extractive Summarization

## TL;DR

This paper introduces BERTSUM, a simple BERT-based model for extractive summarization, achieving state-of-the-art results on CNN/Dailymail with a significant ROUGE-L improvement.

## Contribution

It presents BERTSUM, a novel BERT variant tailored for extractive summarization, setting new performance benchmarks.

## Key findings

- Outperforms previous models by 1.65 ROUGE-L on CNN/Dailymail
- Achieves state-of-the-art extractive summarization results
- Provides reproducible code for the proposed method

## Abstract

BERT, a pre-trained Transformer model, has achieved ground-breaking performance on multiple NLP tasks. In this paper, we describe BERTSUM, a simple variant of BERT, for extractive summarization. Our system is the state of the art on the CNN/Dailymail dataset, outperforming the previous best-performed system by 1.65 on ROUGE-L. The codes to reproduce our results are available at https://github.com/nlpyang/BertSum