Exploring Security Vulnerabilities of Deep Learning Models by Adversarial Attacks

Figure 2

Levenshtein distance.