A question-answering system for aircraft pilots' documentation

Abstract

The aerospace industry relies on massive collections of complex and technical documents covering system descriptions, manuals or procedures. This paper presents a question answering (QA) system that would help aircraft pilots access information in this documentation by naturally interacting with the system and asking questions in natural language. After describing each module of the dialog system, we present a multi-task based approach for the QA module which enables performance improvement on a Flight Crew Operating Manual (FCOM) dataset. A method to combine scores from the retriever and the QA modules is also presented.

arXiv comment11 pages, 8 figures

References in corpus (8)

• A Survey on Multi-Task Learning
• TAPAS: Weakly Supervised Table Parsing via Pre-training
• End-to-End Open-Domain Question Answering with BERTserini
• A Survey of Available Corpora for Building Data-Driven Dialogue Systems
• Rasa: Open Source Language Understanding and Dialogue Management
• A BERT Baseline for the Natural Questions
• Risk and Ambiguity in Information Seeking: Eye Gaze Patterns Reveal Contextual Behaviour in Dealing with Uncertainty
• TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data