Human Perceptibility
task
ID: humanperceptibility
Related Papers
How do humans perceive adversarial text? A reality check on the validity and naturalness of word-based adversarial attacks
AbstractNatural Language Processing (NLP) models based on Machine Learning (ML) are susceptible to adversarial attacks – malicious algorithms that imperceptibly modify input text to force models into making incorrect predictions. …...