Invariant Representations without Adversarial Training

Open in new window