ESPN: Extremely Sparse Pruned Networks
This work addresses the challenge of deploying deep neural networks in resource-constrained environments, representing an incremental improvement over existing pruning methods.
The paper tackles the problem of deep neural networks being overparameterized for compute-limited systems by developing an iterative mask discovery method that achieves state-of-the-art compression of very deep networks, outperforming existing pruning approaches in test accuracy and compression ratio.
Deep neural networks are often highly overparameterized, prohibiting their use in compute-limited systems. However, a line of recent works has shown that the size of deep networks can be considerably reduced by identifying a subset of neuron indicators (or mask) that correspond to significant weights prior to training. We demonstrate that an simple iterative mask discovery method can achieve state-of-the-art compression of very deep networks. Our algorithm represents a hybrid approach between single shot network pruning methods (such as SNIP) with Lottery-Ticket type approaches. We validate our approach on several datasets and outperform several existing pruning approaches in both test accuracy and compression ratio.