Bad Characters - Imperceptible NLP Attacks

Explore imperceptible NLP attacks in this 20-minute IEEE conference talk presented by researchers from the University of Cambridge, Vector Institute, University of Edinburgh, and University of Toronto. Delve into the concept of "Bad Characters" and their potential impact on natural language processing systems, gaining insights into cutting-edge cybersecurity challenges in the field of artificial intelligence and machine learning.