Adversarial Example Generation with Syntactically Controlled Paraphrase Networks
| Autor: | Mohit Iyyer, Luke Zettlemoyer, Kevin Gimpel, John Wieting |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
FOS: Computer and information sciences
Computer Science - Computation and Language Parsing Computer science business.industry 020207 software engineering 02 engineering and technology computer.software_genre Syntax Paraphrase Adversarial system TheoryofComputation_MATHEMATICALLOGICANDFORMALLANGUAGES 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence business Computation and Language (cs.CL) computer Sentence Natural language processing |
| Zdroj: | NAACL-HLT |
| DOI: | 10.18653/v1/n18-1170 |
| Popis: | We propose syntactically controlled paraphrase networks (SCPNs) and use them to generate adversarial examples. Given a sentence and a target syntactic form (e.g., a constituency parse), SCPNs are trained to produce a paraphrase of the sentence with the desired syntax. We show it is possible to create training data for this task by first doing backtranslation at a very large scale, and then using a parser to label the syntactic transformations that naturally occur during this process. Such data allows us to train a neural encoder-decoder model with extra inputs to specify the target syntax. A combination of automated and human evaluations show that SCPNs generate paraphrases that follow their target specifications without decreasing paraphrase quality when compared to baseline (uncontrolled) paraphrase systems. Furthermore, they are more capable of generating syntactically adversarial examples that both (1) "fool" pretrained models and (2) improve the robustness of these models to syntactic variation when used to augment their training data. NAACL 2018 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|