AI-Driven Summarization of Academic Literature using Transformer Model

This research study describes an innovative method for automated academic research article summarization that uses a transformer-based model in conjunction with natural language processing techniques. The proposed approach uses Named Entity Recognition, Latent Dirichlet Allocation for topic modelling, Term Frequency Analysis, Term Frequency-Inverse Document Frequency, and extraction of quantitatively significant sentences to identify critical features of a document. The summarization is based on an optimized Text-To-Text Transfer Transformer model that can provide short summaries that include important conclusions, techniques, and outcomes from research articles. This work not only contributes to the domain of automated text summarization but also has the potential to help researchers and academics quickly grasp knowledge from large research literature. The proposed model has achieved a F1 score of 93.5% for text summarization.