• 文献检索
  • 文档翻译
  • 深度研究
  • 学术资讯
  • Suppr Zotero 插件Zotero 插件
  • 邀请有礼
  • 套餐&价格
  • 历史记录
应用&插件
Suppr Zotero 插件Zotero 插件浏览器插件Mac 客户端Windows 客户端微信小程序
定价
高级版会员购买积分包购买API积分包
服务
文献检索文档翻译深度研究API 文档MCP 服务
关于我们
关于 Suppr公司介绍联系我们用户协议隐私条款
关注我们

Suppr 超能文献

核心技术专利:CN118964589B侵权必究
粤ICP备2023148730 号-1Suppr @ 2026

文献检索

告别复杂PubMed语法,用中文像聊天一样搜索,搜遍4000万医学文献。AI智能推荐,让科研检索更轻松。

立即免费搜索

文件翻译

保留排版,准确专业,支持PDF/Word/PPT等文件格式,支持 12+语言互译。

免费翻译文档

深度研究

AI帮你快速写综述,25分钟生成高质量综述,智能提取关键信息,辅助科研写作。

立即免费体验

受限玻尔兹曼机和广义霍普菲尔德网络的相图,具有任意先验。

Phase diagram of restricted Boltzmann machines and generalized Hopfield networks with arbitrary priors.

机构信息

Dipartimento di Matematica e Fisica Ennio De Giorgi, Università del Salento, 73100 Lecce, Italy.

Institut für Mathematik, Universität Zürich, CH-8057 Zürich, Switzerland.

出版信息

Phys Rev E. 2018 Feb;97(2-1):022310. doi: 10.1103/PhysRevE.97.022310.

DOI:10.1103/PhysRevE.97.022310
PMID:29548112
Abstract

Restricted Boltzmann machines are described by the Gibbs measure of a bipartite spin glass, which in turn can be seen as a generalized Hopfield network. This equivalence allows us to characterize the state of these systems in terms of their retrieval capabilities, both at low and high load, of pure states. We study the paramagnetic-spin glass and the spin glass-retrieval phase transitions, as the pattern (i.e., weight) distribution and spin (i.e., unit) priors vary smoothly from Gaussian real variables to Boolean discrete variables. Our analysis shows that the presence of a retrieval phase is robust and not peculiar to the standard Hopfield model with Boolean patterns. The retrieval region becomes larger when the pattern entries and retrieval units get more peaked and, conversely, when the hidden units acquire a broader prior and therefore have a stronger response to high fields. Moreover, at low load retrieval always exists below some critical temperature, for every pattern distribution ranging from the Boolean to the Gaussian case.

摘要

受限玻尔兹曼机由二部自旋玻璃的吉布斯测度描述,而自旋玻璃又可以看作广义的霍普菲尔德网络。这种等价关系使我们能够根据其检索能力来描述这些系统的状态,无论是在低负载还是高负载下,都可以对纯态进行检索。我们研究了顺磁-自旋玻璃和自旋玻璃-检索相变,因为模式(即权重)分布和自旋(即单元)先验从高斯实变量到布尔离散变量平滑地变化。我们的分析表明,检索相的存在是稳健的,并且不仅限于具有布尔模式的标准霍普菲尔德模型。当模式条目和检索单元变得更加尖锐时,检索区域会变大,而当隐藏单元获得更广泛的先验时,它们对高场的响应会更强。此外,在低负载下,对于从布尔到高斯的每种模式分布,在某个临界温度以下总是存在检索。

相似文献

1
Phase diagram of restricted Boltzmann machines and generalized Hopfield networks with arbitrary priors.受限玻尔兹曼机和广义霍普菲尔德网络的相图,具有任意先验。
Phys Rev E. 2018 Feb;97(2-1):022310. doi: 10.1103/PhysRevE.97.022310.
2
Phase transitions in restricted Boltzmann machines with generic priors.具有通用先验的受限玻尔兹曼机的相变。
Phys Rev E. 2017 Oct;96(4-1):042156. doi: 10.1103/PhysRevE.96.042156. Epub 2017 Oct 27.
3
On the equivalence of Hopfield networks and Boltzmann Machines.Hopfield 网络与 Boltzmann 机的等价性。
Neural Netw. 2012 Oct;34:1-9. doi: 10.1016/j.neunet.2012.06.003. Epub 2012 Jun 23.
4
On the effective initialisation for restricted Boltzmann machines via duality with Hopfield model.通过对偶性与 Hopfield 模型实现受限 Boltzmann 机的有效初始化。
Neural Netw. 2021 Nov;143:314-326. doi: 10.1016/j.neunet.2021.06.017. Epub 2021 Jun 23.
5
Parallel retrieval of correlated patterns: from Hopfield networks to Boltzmann machines.并行关联模式检索:从霍普菲尔德网络到玻尔兹曼机。
Neural Netw. 2013 Feb;38:52-63. doi: 10.1016/j.neunet.2012.11.010. Epub 2012 Nov 28.
6
Reconstructing the Hopfield network as an inverse Ising problem.将霍普菲尔德网络重构为逆伊辛问题。
Phys Rev E Stat Nonlin Soft Matter Phys. 2010 Mar;81(3 Pt 2):036104. doi: 10.1103/PhysRevE.81.036104. Epub 2010 Mar 10.
7
Boltzmann Machines as Generalized Hopfield Networks: A Review of Recent Results and Outlooks.作为广义霍普菲尔德网络的玻尔兹曼机:近期成果与展望综述
Entropy (Basel). 2020 Dec 29;23(1):34. doi: 10.3390/e23010034.
8
Mean-field message-passing equations in the Hopfield model and its generalizations.霍普菲尔德模型及其推广中的平均场消息传递方程。
Phys Rev E. 2017 Feb;95(2-1):022117. doi: 10.1103/PhysRevE.95.022117. Epub 2017 Feb 14.
9
Attention in a Family of Boltzmann Machines Emerging From Modern Hopfield Networks.源自现代霍普菲尔德网络的玻尔兹曼机家族中的注意力机制
Neural Comput. 2023 Jun 20:1-18. doi: 10.1162/neco_a_01597.
10
Multitasking attractor networks with neuronal threshold noise.具有神经元阈值噪声的多重任务吸引子网络。
Neural Netw. 2014 Jan;49:19-29. doi: 10.1016/j.neunet.2013.09.008. Epub 2013 Sep 30.

引用本文的文献

1
Inverse problems for structured datasets using parallel TAP equations and restricted Boltzmann machines.使用并行TAP方程和受限玻尔兹曼机求解结构化数据集的逆问题。
Sci Rep. 2021 Oct 7;11(1):19990. doi: 10.1038/s41598-021-99353-2.
2
Boltzmann Machines as Generalized Hopfield Networks: A Review of Recent Results and Outlooks.作为广义霍普菲尔德网络的玻尔兹曼机:近期成果与展望综述
Entropy (Basel). 2020 Dec 29;23(1):34. doi: 10.3390/e23010034.
3
Statistical Physics for Medical Diagnostics: Learning, Inference, and Optimization Algorithms.
用于医学诊断的统计物理学:学习、推理与优化算法
Diagnostics (Basel). 2020 Nov 19;10(11):972. doi: 10.3390/diagnostics10110972.
4
A high-bias, low-variance introduction to Machine Learning for physicists.面向物理学家的机器学习高偏差、低方差入门介绍。
Phys Rep. 2019 May 30;810:1-124. doi: 10.1016/j.physrep.2019.03.001. Epub 2019 Mar 14.