Rethinking Targeted Adversarial Attacks For Neural Machine Translation

Open in new window