Learning Private Representations through Entropy-based Adversarial Training

Tassilo Klein, Moin Nabi

How can we learn a representation with high predictive power while preserving user privacy? We present an adversarial representation learning method for sanitizing sensitive conten…