这里的一些东西只是将过去已有的东西用PGM解释了一遍,但优势还是明显的,对整体认识有帮助。

Video: https://www.youtube.com/watch?v=ogs4Oj8KahQ&index=13&list=PL50E6E80E8525B59C

3 - 1 - Overview of Template Models

3 - 2 - Temporal Models

3 - 3 - Temporal Models - HMMs

3 - 4 - Plate Models (以下主要是此内容)

DBM

动态BM其实就是加入了时间这个变脸,然后随着时间的变化,原有的节点会产生状态转移这样的过程。

Nested Plates & Overlapping Plates

  

左:一个courses框框可以包含很多个students框框。

右:如此一来,Courses框框的Difficulty就不是共享模式了。

结合后的效果如下:

意义在于:

Parameters and structure are reused within a BN and across different BNs.

Collective Inference

一个推断思维的例子:

学生选择第二第三课程,分低;但他的第一课程,分高;

那么,第一课程可能真的简单。

Plate notation

In Bayesian inferenceplate notation is a method of representing variables that repeat in a graphical model.

Instead of drawing each repeated variable individually, a plate or rectangle is used to group variables into a subgraph that repeat together, and a number is drawn on the plate to represent the number of repetitions of the subgraph in the plate.

The assumptions are that

    • the subgraph is duplicated that many times,
    • the variables in the subgraph are indexed by the repetition number, and
    • any links that cross a plate boundary are replicated once for each subgraph repetition.

In this example, we consider Latent Dirichlet allocation, a Bayesian network that models how documents in a corpus are topically related. There are two variables not in any plate:

    • α is the parameter of the uniform Dirichlet prior on the per-document topic distributions,
    • β is the parameter of the uniform Dirichlet prior on the per-topic word distribution.

The outermost plate represents all the variables related to a specific document, including , the topic distribution for document i.

The M in the corner of the plate indicates that the variables inside are repeated M times, once for each document.

The inner plate represents the variables associated with each of the  words in document i:  is the topic for the jth word in document i, and  is the actual word used.

The N in the corner represents the repetition of the variables in the inner plate  times, once for each word in document i.

    • The circle representing the individual words is shaded, indicating that each  is observable, and
    • the other circles are empty, indicating that the other variables are latent variables.

The directed edges between variables indicate dependencies between the variables: for example, each  depends on  and β.

[PGM] Temporal Models的更多相关文章

  1. 本人AI知识体系导航 - AI menu

    Relevant Readable Links Name Interesting topic Comment Edwin Chen 非参贝叶斯   徐亦达老板 Dirichlet Process 学习 ...

  2. {ICIP2014}{收录论文列表}

    This article come from HEREARS-L1: Learning Tuesday 10:30–12:30; Oral Session; Room: Leonard de Vinc ...

  3. A Statistical View of Deep Learning (IV): Recurrent Nets and Dynamical Systems

    A Statistical View of Deep Learning (IV): Recurrent Nets and Dynamical Systems Recurrent neural netw ...

  4. 论文笔记:语音情感识别(三)手工特征+CRNN

    一:Emotion Recognition from Human Speech Using Temporal Information and Deep Learning(2018 InterSpeec ...

  5. Machine Learning Basic Knowledge

    常用的数据挖掘&机器学习知识(点) Basis(基础): MSE(MeanSquare Error 均方误差),LMS(Least MeanSquare 最小均方),LSM(Least Squ ...

  6. stanford推荐阅读目录

    stanford deep learning 网站上推荐的阅读目录: UFLDL Recommended Readings   If you're learning about UFLDL (Unsu ...

  7. [PGM] What is Probabalistic Graphical Models

    学术潜规则: 概率图模型提出的意义在于将过去看似零散的topic/model以一种统一的方式串联了起来,它便于从整体上看待这些问题,而非具体解决了某个细节. 举个例子:梯度下降,并非解决神经网络收敛问 ...

  8. 使用 LaTeX 绘制 PGM(Probabilistic Graphical Models)中的贝叶斯网络(bayesian networks)

    Software for drawing bayesian networks (graphical models) 这里需要调用 latex 中的绘图库:TikZ and PGF. 注意,下述 tex ...

  9. PRML读书会第八章 Graphical Models(贝叶斯网络,马尔科夫随机场)

    主讲人 网神 (新浪微博: @豆角茄子麻酱凉面) 网神(66707180) 18:52:10 今天的内容主要是: 1.贝叶斯网络和马尔科夫随机场的概念,联合概率分解,条件独立表示:2.图的概率推断in ...

随机推荐

  1. fixed、absolute

    <!DOCTYPE html><html> <head> <meta charset="utf-8"> <meta name= ...

  2. Mac 安装配置nexus2.6 搭建Maven的中央仓库

    今天配置java 环境,安装nexus 百度了好久才安装好,所以特别写下来 分享给同样遇到问题的你.废话不多说,直接上步骤 前置条件 :已经安装了JDK 下载nexus(http://www.sona ...

  3. 系统wmiprvse.exe占用CPU非常高,求解决

    1.wmiprvse.exe是微软Windows操作系统的一部分.用于通过WinMgmt.exe程序处理WMI操作.文件位置有二处: C:\WINDOWS\system32\wbem\wmiprvse ...

  4. android: 碎片的demo

    现在你已经将关于碎片的重要知识点都掌握得差不多了,不过在灵活运用方面可能还有 些欠缺,因此又该进入我们本章的最佳实践环节了. 前面有提到过,碎片很多时候都是在平板开发当中使用的,主要是为了解决屏幕空间 ...

  5. cordova 开发笔记

    1.安装 Node.js Cordova需要Node.js环境,访问https://nodejs.org 下载安装, LTS版本即可,不要最新版. 2.安装 Cordova 执行下述命令把Cordov ...

  6. java中四舍五入——double转BigDecimal的精度损失问题

    代码: double d = -123456789012345.3426;//5898895455898954895989; NumberFormat nf = new DecimalFormat(& ...

  7. LINK:fatal error LNK1123: 转换到 COFF 期间失败: 文件无效或损坏 (转)

    很多伙伴在更新VS2010,或者卸载VS2012安装2010后,建立Win32 Console Project/MFC项目时会出现"LINK : fatal error LNK1123: 转 ...

  8. [leetcode]Rotate List @ Python

    原题地址:https://oj.leetcode.com/problems/rotate-list/ 题意: Given a list, rotate the list to the right by ...

  9. 解锁scott账户方法

    装完了数据库,忘了给scott账户解锁.这时可以在sql plus工具里,也可以在控制台通过命令行给scott账户解锁. 在第一种情况下,以system账户+自己安装时设置的密码,登录SQL Plus ...

  10. iOS 几种加密方法

    iOS常见的几种加密方法 普通加密方法是讲密码进行加密后保存到用户偏好设置中 钥匙串是以明文形式保存,但是不知道存放的具体位置 1.base64加密 base64 编码是现代密码学的基础 基本原理: ...