Block pruning residual networks using Multi-Armed Bandits

Recently, deep neural networks have been adopted in many application domains showing impressive performances that sometimes even exceed those of humans. However, the considerable size of deep learning models and their computational cost limits their use in mobile and embedded systems. Pruning is an important technique that has been used in the literature to compress a trained model by removing some of its elements (weights, neurons, filters, etc.) that do not notably contribute to its performance. In this paper, we propose a new block pruning method for residual networks that removes holes layers of the network based on the Multi-Armed Bandits (MAB) framework. The proposed method takes advantage of the existence of multiple trainable paths in residual models to identify and prune the less important residual blocks for inference. To evaluate our approach, we conducted extensive experiments using three different MAB algorithms, namely $\varepsilon $e-greedy, Win Stay Lose Shift (WSLS), and the Upper Confidence Bound (UCB). The obtained results on different datasets show that our block pruning method can considerably reduce the model size and FLOPs with little to no impact on overall accuracy. We conducted a comparative analysis of our method with three existing techniques to demonstrate that our approach is highly competitive in terms of accuracy, while significantly reducing both model size and FLOPs. Moreover, our experimental results demonstrate that when our block pruning method is combined with a filter-level pruning technique, it leads to even greater reductions in model size and FLOPs while maintaining a relatively high accuracy.

Automated summary

This paper proposes a new block pruning method for residual networks based on the Multi-Armed Bandits framework. The method identifies and prunes less important residual blocks to reduce model size and computational cost. Experimental results show the effectiveness of the method.