Title: Artificial Intelligence Across Disciplines: Research, Innovation, and Intelligent Solutions
Editors: Dr. Subita Bhagat, Dr. A. Balamurugan, Dr. P. Krishna Kumar, and Mrs. S. Nandhini Devi
ISBN: 978-81-69857-83-3
Chapter: 25
DOI: https://doi.org/10.59646/815/25
Authors: Ms. Hemavathi S, Ms. Annapoorani T, and Ms. Sindhuja S
Abstract
Contemporary digital enterprise architectures produce vast quantities of multi-modal, unstructured content, creating substantial bottlenecks for traditional lexical and static information retrieval mechanisms. This chapter presents an end-to-end framework for artificial intelligence-powered intelligent search and autonomous information discovery, uniting sparse probabilistic matching, dense multi-vector semantic embeddings, reciprocal rank fusion, and autonomous agentic query-decomposition loops. Utilizing a rigorous multi-domain case study comprising enterprise clinical research archives and regulatory filings, the proposed architecture orchestrates autonomous iterative discovery cycles that parse multi-hop research hypotheses, perform targeted retrieval, dynamically assess passage utility, and synthesize verified knowledge. Empirical results demonstrate that the hybrid autonomous approach outperforms isolated lexical and dense retrieval paradigms across precision, recall, mean reciprocal rank, and normalized discounted cumulative gain, while cutting manual query formulation overhead. By integrating dynamic query reformation with cross-encoder re-ranking, this chapter establishes a scalable, explainable, and fault-tolerant paradigm for enterprise knowledge intelligence and autonomous scientific discovery.
Keywords: Autonomous Discovery Agents; Dense Passage Retrieval; Hybrid Search Architectures; Multi-Hop Information Retrieval; Reciprocal Rank Fusion; Semantic Knowledge Discovery.