Combating Adversarial Misspellings with Robust Word Recognition

Combating Adversarial Misspellings with Robust Word Recognition 來源：ACL2019 鏈接： https://arxiv.org/pdf/1905.11268.pdf 代碼：https://github.com/danishpruthi/Adversarial-Misspellings 動機 文本攻防逐漸發展成爲一個研究方向。如圖所示

>>阅读原文<<