Optimized Text Summarization Using BART

Web article summarization is a challenging task in approaches natural language processing (NLP) that aims to generate a condensed summarization and informative summary of a long web article. Recently, It emerged as a powerful pre-trained deep learning model in NLP tasks. In this study, the effectiveness of BERT-based approach on three datasets: SAMSum, CNN dailymail and was pre-trained Scientific Paper. BERT model was pre-trained on a large corpus of text data and fine-tune it on the target datasets for extractive summarization task. The performance of the proposed comprehension approach using standard evaluation metrics such as ROUGE-1. ROUGE-2, and ROUGE-L. The experimental results show that proposed approach outperforms several existing state-of-the-art summarization models on all three datasets. Specifically, best performance on SAMSum dataset with ROUGE-1, ROUGE-2, and ROUGE-L scores of 44.51, 20.90, and 37.30, respectively. These results demonstrate the potential of BART-based approach for web article summarization and its applicability on different types of datasets.