Comparative Analysis of CNN-Based Image Denoising Techniques

Abstract

For this project, it aims to implement and enhance an automatic text summarization system using the recently popular BART (Bidirectional and Auto-Regressive Transformers) model. The system is trained on the CNN/DailyMail dataset and is also examined with the help of the ROUGE score where the system performs quite well in ROUGE-1, ROUGE-2, and the ROUGE-L test set. It was observed that the summaries generated were both logically connected and summarized data points were pertinent and hence, the capability of the BART model to work well in the field of abstractive summarization was proved. Altogether, the results shed more light on how the deep learning techniques can be used in improving the Information Retrieval and documents Management System. There is still more work that can be done in the future; considering to refine and enrich the dataset to enhance the model. The BART model integrates bidirectional context comprehension with autoregressive text generation to produce coherent and contextually relevant summaries. Performance was evaluated using ROUGE metrics, demonstrating the model's effectiveness in generating precise and informative summaries. This work underscores the potential of deep learning techniques in advancing summarisation technologies.