Y. Li, M. Yang, Z. Xu, and Z. Zhang, “Learning with Marginalized Corrupted Features and Labels Together”, AAAI, vol. 30, no. 1, Feb. 2016.