Abstract
Modern organizations often struggle to identify systemic structural issues because post-mortem reviews, lessons-learned records, and risk logs are isolated in unstructured narrative text files across disparate storage repositories. A system is described that utilizes a multi-stage language model processing pipeline to autonomously aggregate, structure, and cross-reference completely unstructured, multi-source operational narrative text documents to expose global systemic risks and organizational blind spots. The system continuously connects to raw text silos, extracts semantic entities, and standardizes unstructured content into a globally normalized taxonomy to map risk correlations. Keywords: unstructured data, semantic analysis, natural language processing, language models, risk management, data clustering, organizational taxonomy, and anomaly detection.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Vianana, Michael; Kayande, Tanmay; Nageer, Nigel; and Murlidharan, Nupur, "Automated Multi-Stage Semantic Analysis and Synthesis of Heterogeneous Unstructured Operational Repositories", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/10860