Mohammad Tooneer

Leakage-Robust Evaluation and Data-Scale Sensitivity of Attention-Enhanced Multi-Task Learning for Joint Fault Diagnosis and Remaining Useful Life Estimation

Md Mahamudur Rahaman Shamim, Md. Nuruzzaman, Zannatul Ferdus, Md Rajib Ahmed, Abieer Nwshad Anward, Mohammad Tooneer, et al.

Multi-task deep learning models that jointly perform fault classification and remaining useful life (RUL) regression are increasingly used in predictive maintenance, yet reported performance can be strongly affected by how sliding-window sequences are split into training and test…Read more