Deep Learning

Is it possible to induce sparseness in neural networks while preserving its performances? An overview of latest advances in making neural approaches more parsimonious