Con-Detect: Detecting Adversarially Perturbed Natural Language Inputs to Deep Classifiers Through Holistic Analysis

This repository is an open-source code to reproduce the results of our paper titled, Con-Detect: Detecting Adversarially Perturbed Natural Language Inputs to Deep Classifiers Through Holistic Analysis.

Citation

If you use this code in your work, please cite using the following BibTeX entry:

@article{ali2023detect,
  author       = {Hassan Ali and
                  Muhammad Suleman Khan and
                  Amer AlGhadhban and
                  Meshari Alazmi and
                  Ahmed Alzamil and
                  Khaled Al{-}Utaibi and
                  Junaid Qadir},
  title        = {Con-Detect: Detecting adversarially perturbed natural language inputs
                  to deep classifiers through holistic analysis},
  journal      = {Computers & Security},
  volume       = {132},
  pages        = {103367},
  year         = {2023},
  url          = {https://doi.org/10.1016/j.cose.2023.103367},
  doi          = {10.1016/J.COSE.2023.103367}
}