{"title":"用于结构稀疏贝叶斯神经网络的 \"尖峰-板块收缩先验\"(Spike-and-Slab Shrinkage Priors for Structurally Sparse Bayesian Neural Networks)。","authors":"Sanket Jantre, Shrijita Bhattacharya, Tapabrata Maiti","doi":"10.1109/TNNLS.2024.3485529","DOIUrl":null,"url":null,"abstract":"<p><p>Network complexity and computational efficiency have become increasingly significant aspects of deep learning. Sparse deep learning addresses these challenges by recovering a sparse representation of the underlying target function by reducing heavily overparameterized deep neural networks. Specifically, deep neural architectures compressed via structured sparsity (e.g., node sparsity) provide low-latency inference, higher data throughput, and reduced energy consumption. In this article, we explore two well-established shrinkage techniques, Lasso and Horseshoe, for model compression in Bayesian neural networks (BNNs). To this end, we propose structurally sparse BNNs, which systematically prune excessive nodes with the following: 1) spike-and-slab group Lasso (SS-GL) and 2) SS group Horseshoe (SS-GHS) priors, and develop computationally tractable variational inference, including continuous relaxation of Bernoulli variables. We establish the contraction rates of the variational posterior of our proposed models as a function of the network topology, layerwise node cardinalities, and bounds on the network weights. We empirically demonstrate the competitive performance of our models compared with the baseline models in prediction accuracy, model compression, and inference latency.</p>","PeriodicalId":13303,"journal":{"name":"IEEE transactions on neural networks and learning systems","volume":null,"pages":null},"PeriodicalIF":10.2000,"publicationDate":"2024-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Spike-and-Slab Shrinkage Priors for Structurally Sparse Bayesian Neural Networks.\",\"authors\":\"Sanket Jantre, Shrijita Bhattacharya, Tapabrata Maiti\",\"doi\":\"10.1109/TNNLS.2024.3485529\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Network complexity and computational efficiency have become increasingly significant aspects of deep learning. Sparse deep learning addresses these challenges by recovering a sparse representation of the underlying target function by reducing heavily overparameterized deep neural networks. Specifically, deep neural architectures compressed via structured sparsity (e.g., node sparsity) provide low-latency inference, higher data throughput, and reduced energy consumption. In this article, we explore two well-established shrinkage techniques, Lasso and Horseshoe, for model compression in Bayesian neural networks (BNNs). To this end, we propose structurally sparse BNNs, which systematically prune excessive nodes with the following: 1) spike-and-slab group Lasso (SS-GL) and 2) SS group Horseshoe (SS-GHS) priors, and develop computationally tractable variational inference, including continuous relaxation of Bernoulli variables. We establish the contraction rates of the variational posterior of our proposed models as a function of the network topology, layerwise node cardinalities, and bounds on the network weights. We empirically demonstrate the competitive performance of our models compared with the baseline models in prediction accuracy, model compression, and inference latency.</p>\",\"PeriodicalId\":13303,\"journal\":{\"name\":\"IEEE transactions on neural networks and learning systems\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":10.2000,\"publicationDate\":\"2024-10-31\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IEEE transactions on neural networks and learning systems\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://doi.org/10.1109/TNNLS.2024.3485529\",\"RegionNum\":1,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE transactions on neural networks and learning systems","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.1109/TNNLS.2024.3485529","RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0
摘要
网络复杂性和计算效率已成为深度学习越来越重要的方面。稀疏深度学习通过减少严重过参数化的深度神经网络,恢复底层目标函数的稀疏表示,从而应对这些挑战。具体来说,通过结构稀疏性(如节点稀疏性)压缩的深度神经架构可提供低延迟推理、更高的数据吞吐量和更低的能耗。在本文中,我们探讨了两种成熟的收缩技术--Lasso 和 Horseshoe,用于贝叶斯神经网络(BNN)的模型压缩。为此,我们提出了结构稀疏的贝叶斯神经网络(BNN),通过以下方法系统性地剪除过多的节点:1) 穗-片组拉索(SS-GL)和 2) SS 组马蹄(SS-GHS)先验,并开发出计算上可控的变分推理,包括伯努利变量的连续松弛。我们建立了我们提出的模型的变分后验收缩率,它是网络拓扑结构、层节点心数和网络权重边界的函数。我们通过实证证明,与基线模型相比,我们的模型在预测准确性、模型压缩和推理延迟方面都具有竞争力。
Spike-and-Slab Shrinkage Priors for Structurally Sparse Bayesian Neural Networks.
Network complexity and computational efficiency have become increasingly significant aspects of deep learning. Sparse deep learning addresses these challenges by recovering a sparse representation of the underlying target function by reducing heavily overparameterized deep neural networks. Specifically, deep neural architectures compressed via structured sparsity (e.g., node sparsity) provide low-latency inference, higher data throughput, and reduced energy consumption. In this article, we explore two well-established shrinkage techniques, Lasso and Horseshoe, for model compression in Bayesian neural networks (BNNs). To this end, we propose structurally sparse BNNs, which systematically prune excessive nodes with the following: 1) spike-and-slab group Lasso (SS-GL) and 2) SS group Horseshoe (SS-GHS) priors, and develop computationally tractable variational inference, including continuous relaxation of Bernoulli variables. We establish the contraction rates of the variational posterior of our proposed models as a function of the network topology, layerwise node cardinalities, and bounds on the network weights. We empirically demonstrate the competitive performance of our models compared with the baseline models in prediction accuracy, model compression, and inference latency.
期刊介绍:
The focus of IEEE Transactions on Neural Networks and Learning Systems is to present scholarly articles discussing the theory, design, and applications of neural networks as well as other learning systems. The journal primarily highlights technical and scientific research in this domain.