L. Sha and T. Lukasiewicz, “Multi-type Disentanglement without Adversarial Training”, AAAI, vol. 35, no. 11, pp. 9515-9523, May 2021.