Phase Transitions in Neural Networks Pruning

Diego Pesce, Yang-Hui He, Guido Caldarelli

Deep neural networks are strongly over-parameterized, often containing far more weights than required for their task. Although such redundancy can aid optimization, it leads to ine…