Research Article
Leveraging Large Language Models for Enhanced Insights in Multi-Document Question Answering
@INPROCEEDINGS{10.4108/eai.28-4-2025.2357854, author={Juluri Shiva Sai and Vantala Swamynath and Bethi Mayookh and Undrakonda Ananth Lakshmi Srinivas and Hilly Gohain Baruah and Rakcinpha Hatibaruah}, title={Leveraging Large Language Models for Enhanced Insights in Multi-Document Question Answering}, proceedings={Proceedings of the 4th International Conference on Information Technology, Civil Innovation, Science, and Management, ICITSM 2025, 28-29 April 2025, Tiruchengode, Tamil Nadu, India, Part I}, publisher={EAI}, proceedings_a={ICITSM PART I}, year={2025}, month={10}, keywords={multi-document question answering large language models information retrieval natural language understanding evaluation metrics}, doi={10.4108/eai.28-4-2025.2357854} }- Juluri Shiva Sai
Vantala Swamynath
Bethi Mayookh
Undrakonda Ananth Lakshmi Srinivas
Hilly Gohain Baruah
Rakcinpha Hatibaruah
Year: 2025
Leveraging Large Language Models for Enhanced Insights in Multi-Document Question Answering
ICITSM PART I
EAI
DOI: 10.4108/eai.28-4-2025.2357854
Abstract
For data retrieval and comprehensive natural language understanding, answering multiple documents at the same time (MDQA) is essential with the basics of Large Language Models (LLMs). This paper is an attempted research work in formulating a MDQA system using the recent LLMs to solve questions from multiple passing documents. Moreover, we use F1-score, BLEU and ROUGE as evaluation metrics to assess the quality of the generated responses. Our results show that our approach improves significantly in answer correctness, coherence and contextual relevance over the baseline Answer extraction models. In doing so we present important insights into generation of LLMs for MDQA and pave ground towards further advancements in multi-document reasoning and knowledge incorporation.
