Joint Regularization on Activations and Weights for Efficient Neural Network Pruning
This work addresses the need for more efficient neural network pruning for deployment in resource-constrained environments, representing an incremental advance over existing weight-only pruning methods.
The paper tackles the problem of improving deployment efficiency in deep neural networks by proposing joint regularization on both activations and weights, achieving up to 98.8% computation cost savings with only 0.4% accuracy degradation.
With the rapid scaling up of deep neural networks (DNNs), extensive research studies on network model compression such as weight pruning have been performed for improving deployment efficiency. This work aims to advance the compression beyond the weights to neuron activations. We propose the joint regularization technique which simultaneously regulates the distribution of weights and activations. By distinguishing and leveraging the significance difference among neuron responses and connections during learning, the jointly pruned network, namely \textit{JPnet}, optimizes the sparsity of activations and weights for improving execution efficiency. The derived deep sparsification of JPnet reveals more optimization space for the existing DNN accelerators dedicated for sparse matrix operations. We thoroughly evaluate the effectiveness of joint regularization through various network models with different activation functions and on different datasets. With $0.4\%$ degradation constraint on inference accuracy, a JPnet can save $72.3\% \sim 98.8\%$ of computation cost compared to the original dense models, with up to $5.2\times$ and $12.3\times$ reductions in activation and weight numbers, respectively.