Efficient Variational Inference for Sparse Deep Learning with Theoretical Guarantee

–Neural Information Processing Systems 

Dense neural network (DNN) may face various problems despite its huge successes in AI fields.