35th Congress of the International Council of the Aeronautical Sciences

11 - Operations and Sustainment

EXPLAINABLE NATURAL LANGUAGE PROCESSING FOR AIRCRAFT STRUCTURAL INTEGRITY DEFECT CLASSIFICATION

V. Kekoc, Defence Science and Technology Group, Australia; M. Scott¹, W.J.C. Verhagen¹, P. Marzocca¹; ¹RMIT University, Australia

Explainable NLP framework to classify structural integrity issues in aircraft defect reports, using transformer-based language models & XAI methods for transparency. Preliminary analysis on aviation safety datasets is promising, w/ explainability techniques identifying domain-relevant indicators. Further validation on ADF data & comparison to human reasoning within sustainment workflows to follow.


View Paper

DOI: 10.71945/icas2026_0738