A Reinforced Generation of Adversarial Examples for Neural Machine Translation
This work addresses the reliability of neural machine translation for industrial applications by exposing system pitfalls, though it is incremental in applying reinforcement learning to this domain.
The paper tackles the problem of neural machine translation systems failing on suboptimal inputs by generating adversarial examples using a reinforcement learning paradigm, which efficiently produces stable, meaning-preserving attacks on architectures like RNN-search and Transformer.
Neural machine translation systems tend to fail on less decent inputs despite its significant efficacy, which may significantly harm the credibility of this systems-fathoming how and when neural-based systems fail in such cases is critical for industrial maintenance. Instead of collecting and analyzing bad cases using limited handcrafted error features, here we investigate this issue by generating adversarial examples via a new paradigm based on reinforcement learning. Our paradigm could expose pitfalls for a given performance metric, e.g., BLEU, and could target any given neural machine translation architecture. We conduct experiments of adversarial attacks on two mainstream neural machine translation architectures, RNN-search, and Transformer. The results show that our method efficiently produces stable attacks with meaning-preserving adversarial examples. We also present a qualitative and quantitative analysis for the preference pattern of the attack, demonstrating its capability of pitfall exposure.