Skip to content

Latest commit

 

History

History
28 lines (22 loc) · 1.42 KB

readme.md

File metadata and controls

28 lines (22 loc) · 1.42 KB
Raw

TRIESTE: TRanslatIon basEd defenSe for Text classifiErs

Codes and implementation files to be used as a reference to the paper: TRIESTE: translation based defense for text classifiers.

Status - Published in Springer's Journal of Ambient Intelligence and Humanized Computing

Overview

This repository contains the following files:

  • run_attack.sh - For generating attack files using textattack library.
  • Translation_EN-DE.py - For running proposed defense using EN-DE translation scheme.
  • Translation_EN-FR.py - For running proposed defense using EN-FR translation scheme.
  • Timer_Trans_DE.py - For inference time comparision with and without the application of the proposed defense using EN-DE scheme.
  • Timer_Trans_FR.py - For inference time comparision with and without the application of the proposed defense using EN-DE scheme.

Further details about the methodology may be directly referred to from the published study.

Citation

If you intend to use this work, kindly cite us as follows:

@ARTICLE{gupta2022trieste,  
  author={Gupta, Anup K. and Paliwal, Vardhan and Rastogi, Aryan and Gupta, Puneet},  
  journal={Journal of Ambient Intelligence and Humanized Computing},   
  title={TRIESTE: translation based defense for text classifiers},   
  year={2022},  
  doi={10.1007/s12652-022-03859-07}
  }