Adversarial Example Generation with Syntactically Controlled Paraphrase Networks (NAACL 2018)

Iyyer et al. propose syntactically controlled paraphrase networks (SCPNs) and use them to generate adversarial examples. They train SCPNs to produce a paraphrase of a sentence with a specified syntax (constituent tree). They generate training data for SCPNs via back-translated sentences augmented with parses and show that the generated paraphrases better fool existing models and improve robustness.


Want to receive more content like this in your inbox?