Automated System for Generating Multilingual Insights from Global News Content

The huge volume of digital news makes it difficult to extract useful insights because of redundancy and information overflow. Although Natural Language Processing (NLP)-based summarization improved efficiency, most techniques lag when it comes to contextual coherence and multilingual support. This study introduces an Interactive News Summary Explorer using the Bidirectional and Auto-Regressive Transformers (BART-Large-CNN) model for abstractive summarization. The system produces compact, context-aware summaries with integrated translation capabilities. In contrast to existing methods, the suggested approach ranks and organizes articles by measuring their semantic similarity to the main topic, ensuring that the most contextually relevant content is presented first. Comparative analysis highlights that the model achieves higher relevance scores (~0.22–0.31) than Term Frequency-Inverse Document Frequency (TF-IDF) and comparable scores to the abstractive Text-to-Text Transfer Transformer (T5) model, while maintaining significantly lower execution time than T5. Although TF-IDF is faster, it lacks semantic understanding. Thus, the proposed method strikes a practical balance between relevance and efficiency, enabling accurate and timely summarization with multilingual support. Future enhancements plan to incorporate multimedia summarization capabilities to create a more flexible framework with expanded functionality and scope.