Exploring Security Vulnerabilities of Deep Learning Models by Adversarial Attacks

Table 4

Accuracy of deep learning models on the Amazon review dataset.

Model	LSTM	BiLSTM
Accuracy	88.483%	88.550%