Deep learning for text summarization using NLP for automated news digest.

Area of Science:

• Natural Language Processing (NLP)
• Artificial Intelligence (AI)
• Machine Learning

Background:

• Text summarization is crucial for condensing information while preserving meaning.
• Deep learning models face challenges like semantic understanding and handling long documents.
• Existing deep learning approaches offer benefits such as time savings and personalization.

Purpose of the Study:

• To propose and evaluate deep learning models for text summarization.
• To compare the performance of T5-base, T5-large, BART CNN, and PEGASUS models.
• To identify the model yielding the highest quality summaries based on evaluation metrics.

Main Methods:

• Data preprocessing and exploratory data analysis (EDA).
• Implementation and training of T5-base, T5-large, BART CNN, and PEGASUS models.
• Evaluation using Rouge and BLUE scores to assess summarization quality.

Main Results:

• The study quantifies the performance of each deep learning model.
• Rouge and BLUE scores were calculated post-training to measure effectiveness.
• Comparative analysis focused on identifying the model with the highest Rouge score.

Conclusions:

• Deep learning models show promise in advancing text summarization techniques.
• Model performance varies, with PEGASUS identified as a strong performer.
• Further research is ongoing to address challenges in deep learning-based summarization.