ER-AE: Differentially Private Text Generation for Authorship Anonymization

2 code implementations • NAACL 2021 • Haohan Bo, Steven H. H. Ding, Benjamin C. M. Fung, Farkhund Iqbal

By augmenting the semantic information through a REINFORCE training reward function, the model can generate differentially private text that has a close semantic and similar grammatical structure to the original text while removing personal traits of the writing style.

Privacy Preserving Text Generation