<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[github_34349558的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/github_34349558</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; github_34349558]]></copyright><item><title><![CDATA[论文解读：How Powerful are Graph Neural Networks?]]></title><link>https://blog.youkuaiyun.com/github_34349558/article/details/120424154</link><guid>https://blog.youkuaiyun.com/github_34349558/article/details/120424154</guid><author>github_34349558</author><pubDate>Wed, 22 Sep 2021 22:11:33 +0800</pubDate><description><![CDATA[How Powerful are Graph Neural Networks?
[Blog : Paper Notes sorted out by paper]
???? About This Paper

Keyulu Xu, Weihua Hu, Jure Leskovec, Stefanie Jegelka
MIT and Stanford University
https://arxiv.org/pdf/1810.00826.pdf
???? GNNs

一个图神经网络的更新准则一般可以归纳为这样：]]></description><category></category></item><item><title><![CDATA[论文解读：Structural Optimization Makes Graph Classification Simpler and Better]]></title><link>https://blog.youkuaiyun.com/github_34349558/article/details/120402724</link><guid>https://blog.youkuaiyun.com/github_34349558/article/details/120402724</guid><author>github_34349558</author><pubDate>Tue, 21 Sep 2021 16:22:48 +0800</pubDate><description><![CDATA[Structural Optimization Makes Graph Classification Simpler and Better
About This Paper
Junran Wu, Jianhao Li, Yicheng Pan, Ke Xu
State Key Lab of Software Development Environment, Beihang University
http://arxiv.org/abs/2109.02027
Preliminary
我们要做的是 图分类任务 ]]></description><category></category></item><item><title><![CDATA[图的结构熵]]></title><link>https://blog.youkuaiyun.com/github_34349558/article/details/120397460</link><guid>https://blog.youkuaiyun.com/github_34349558/article/details/120397460</guid><author>github_34349558</author><pubDate>Tue, 21 Sep 2021 00:25:45 +0800</pubDate><description><![CDATA[图的结构熵
What Is The Problem ?
衡量图的熵值有好多种方式，比如吉布斯熵、香农熵、冯·诺依曼熵等。
如果我们通过一个网络模型去优化得到一个图模型，那么这个图模型的熵可以这么算：
HS=−∑G∈Sp(G)logp(G)H_S=-\sum_{G\in S}p(G)logp(G)HS​=−∑G∈S​p(G)logp(G), 称为：网络模型的结构熵(Structural Entropy of Models of Networks)
其中：p(G)p(G)p(G) 是图 GGG (更准确地说，是]]></description><category></category></item><item><title><![CDATA[《自然语言处理-基于预训练模型的方法》笔记]]></title><link>https://blog.youkuaiyun.com/github_34349558/article/details/119683135</link><guid>https://blog.youkuaiyun.com/github_34349558/article/details/119683135</guid><author>github_34349558</author><pubDate>Fri, 13 Aug 2021 20:06:26 +0800</pubDate><description><![CDATA[《自然语言处理-基于预训练模型的方法》笔记
文章目录《自然语言处理-基于预训练模型的方法》笔记@[toc]〇.写在前面一、绪论(一) NLP 任务体系I.任务层级II.任务类别III.研究层次(二) 预训练的时代二、NLP 基础(一) 文本表示I.独热向量II.分布式表示III.词嵌入表示IV.词袋表示(二) NLP 任务I.语言模型II. 基础任务III. 应用任务(三) 基本问题I. 文本分类问题II. 结构预测问题III. 序列到序列问题(四) 评价指标I. 标准答案明确的情况II. 标准答案不明确的]]></description><category></category></item></channel></rss>