NLP Advances in Risk Analysis Context: Application of Quantum Computing

The models used for facilitating Risk Analysis have undergone notable changes in recent times. This can be attributed to the significant progress made in computing capabilities, as well as the breakthroughs achieved in the field of artificial intelligence. These developments have allowed for the effective extraction of information from complex and unstructured datasets. A particularly useful tool in this regard is Natural Language Processing (NLP), which facilitates the extraction, organization, and classification of information from textual sources. This enables the identification of patterns and trends in an automated fashion. The emergence of Quantum Computing (QC) over the last fifty years has made it one of the most significant fields in computational science. Although fully scalable QC remains elusive, the availability of intermediate noisy scalable QC devices has enabled the realization of near-term computation on quantum devices using quantum algorithms. QC can (i) solve classically intractable problems and (ii) problems that, while tractable, are classically infeasible. Given the significant computational resources required by state-of-the-art NLP models to extract meaning from text, exploring methods to address such problems using QC would be valuable. The primary aim of this study is to develop a model based on Quantum Natural Language Processing (QNLP) to classify whether human error caused aviation accidents, based on the accident narrative. The research utilizes a database of accident investigation reports conducted by the National Transportation Safety Board (NTSB) spanning from 1982 to 2022.

Keywords: Quantum natural language processing, Quantum computing, Aviation accidents, Accident investigation reports.

Download PDF