Towards Precise Observations of Neural Model Robustness in Classification

Wenchuan Mu, Kwan Hui Lim

In deep learning applications, robustness measures the ability of neural models that handle slight changes in input data, which could lead to potential safety hazards, especially i…