Y. Guo, Y. Li, L. Wang, and T. Rosing, “Depthwise Convolution Is All You Need for Learning Multiple Visual Domains”, AAAI, vol. 33, no. 01, pp. 8368-8375, Jul. 2019.