Transformer-based descriptive question answering system using a fine-tuned BERT model and sentence-ranking pipeline, published at an IEEE conference.
Highlights
Built a transformer-based architecture for descriptive question answering, fine-tuning BERT to retrieve and rank relevant passages from document collections.
Implemented a sentence-ranking pipeline that scores candidate answers using transfer learning and semantic similarity for accurate response generation.
Published the work at the IEEE Pune Section International Conference (Dec 2021) as part of a bachelor's final-year project.