Towards enhancing the reproducibility of deep learning bugs: an empirical study

MB Shah, MM Rahman, F Khomh - Empirical Software Engineering, 2025 - Springer
Context Deep learning has achieved remarkable progress in various domains. However,
like any software system, deep learning systems contain bugs, some of which can have …

Improved Detection and Diagnosis of Faults in Deep Neural Networks Using Hierarchical and Explainable Classification

S Jahan, MB Shah, P Mahbub, MM Rahman - arxiv preprint arxiv …, 2025 - arxiv.org
Deep Neural Networks (DNN) have found numerous applications in various domains,
including fraud detection, medical diagnosis, facial recognition, and autonomous driving …

Predicting the Reliability of an Image Classifier under Image Distortion

D Nguyen, S Gupta, K Do, S Venkatesh - arxiv preprint arxiv:2412.16881, 2024 - arxiv.org
In image classification tasks, deep learning models are vulnerable to image distortions ie
their accuracy significantly drops if the input images are distorted. An image-classifier is …