What AI says about TextFooler

TextFooler is an open-source project that provides code to perform natural language adversarial attacks against text classification and entailment models, to evaluate model robustness. It hosts the source code for the paper 'Is BERT Really Robust? Natural Language Attack on Text Classification and Entailment' and includes datasets, prerequisites, and scripts to run attacks on classifiers and NLI models. The repository aims to help researchers and developers study vulnerabilities of NLP models and benchmark defense strategies using seven datasets and attack tools.