advanced convolutional neural networks, Q1 2019

Student session walking through recent CNN papers — training regularization (dropout, shake-shake, cutout, mixup), image-recognition architectures (CondenseNet, SparseNet, ENAS, MnasNet) and 3D convolutions.