Lexical-Based Adversarial Reinforcement Training for Robust Sentiment Classification

Published in EMNLP, 2019

In this paper, we proposed an adversarial reinforcement training method that can borrow the power of external knowledge base to enhance robustness of sentiment classifier.

Download paper here
Jingjing Xu#, Liang Zhao#(equal contribution), Hanqi Yan, Qi Zeng, Yun Liang, Xu Sun*
In EMNLP19, Full paper