arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2109.05075cs.LGcs.AIeess.SP
收藏

利用 ℓ0 范数正则化与权重剪枝压缩神经网络

On the Compression of Neural Networks Using $\ell_0$-Norm Regularization and Weight Pruning

Federal University of Santa Catarina(圣卡塔琳娜联邦大学)

Felipe Dennis de Resende Oliveira, Eduardo Luiz Ortiz Batista, Rui Seara

AI总结:

本文提出结合新型 ℓ0 范数正则化、权重剪枝、ℓ2 正则化与微调的神经网络压缩方案,以训练稀疏性并获得小型高效网络。

摘要:

尽管高容量计算平台日益普及,但实现复杂度仍是神经网络在真实世界部署中的一大关切。这一关切不仅源于最先进网络架构的高昂成本,也源于近期对边缘智能以及在嵌入式应用中使用神经网络的推动。在此背景下,网络压缩技术因能够在将推理精度保持在令人满意水平的同时降低部署成本而受到关注。本文致力于开发一种新型神经网络压缩方案。为此,首先提出一种新的基于 ℓ0 范数的正则化形式,它能够在训练期间促使网络产生强稀疏性。随后,利用剪枝技术针对已训练网络中较小的权重进行处理,从而获得规模更小但高效的网络。所提压缩方案还结合使用 ℓ2 范数正则化以避免过拟合,并通过微调提升剪枝后网络的性能。文中给出实验结果,旨在展示该方案的有效性,并与竞争性方法进行比较。

英文摘要:

Despite the growing availability of high-capacity computational platforms, implementation complexity still has been a great concern for the real-world deployment of neural networks. This concern is not exclusively due to the huge costs of state-of-the-art network architectures, but also due to the recent push towards edge intelligence and the use of neural networks in embedded applications. In this context, network compression techniques have been gaining interest due to their ability for reducing deployment costs while keeping inference accuracy at satisfactory levels. The present paper is dedicated to the development of a novel compression scheme for neural networks. To this end, a new form of $\ell_0$-norm-based regularization is firstly developed, which is capable of inducing strong sparseness in the network during training. Then, targeting the smaller weights of the trained network with pruning techniques, smaller yet highly effective networks can be obtained. The proposed compression scheme also involves the use of $\ell_2$-norm regularization to avoid overfitting as well as fine tuning to improve the performance of the pruned network. Experimental results are presented aiming to show the effectiveness of the proposed scheme as well as to make comparisons with competing approaches.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑