<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title>Deep Learning - Tag - Xiaopeng Xu</title><link>https://xu-xp.com/tags/deep-learning/</link><description>Deep Learning - Tag - Xiaopeng Xu</description><generator>Hugo -- gohugo.io</generator><language>en</language><managingEditor>xiaopeng.xu@kaust.edu.sa (Xiaopeng Xu)</managingEditor><webMaster>xiaopeng.xu@kaust.edu.sa (Xiaopeng Xu)</webMaster><lastBuildDate>Mon, 01 Jan 2024 00:00:00 +0000</lastBuildDate><atom:link href="https://xu-xp.com/tags/deep-learning/" rel="self" type="application/rss+xml"/><item><title>深度学习在图上的应用</title><link>https://xu-xp.com/posts/dl_in_graph/</link><pubDate>Mon, 01 Jan 2024 00:00:00 +0000</pubDate><author>xiaopeng.xu@kaust.edu.sa (Xiaopeng Xu)</author><guid>https://xu-xp.com/posts/dl_in_graph/</guid><description>&lt;blockquote>
&lt;p>Zhang Z , Cui P , Zhu W . Deep Learning on Graphs: A Survey[J]. 2018.
深度学习在大量领域表现出明显的效果，无论是语音，图像，还是自然语言处理。但是由于图结构数据具有独特的属性，深度学习并不是自然的适用。最近，在这个方向进行了大量的研究极大地促进了图分析技术。调研了可以应用于图的不同种类深度学习方法，主要分为三大类：半监督学习，包括图神经网络和图卷积神经网络；无监督学习图自编码机；最新的进展，图循环神经网络和图强化学习。分析了不同方法的特点和联系。&lt;/p></description></item><item><title>Transformer 学习笔记</title><link>https://xu-xp.com/posts/transformer_notes/</link><pubDate>Mon, 13 Jun 2022 00:00:00 +0000</pubDate><author>xiaopeng.xu@kaust.edu.sa (Xiaopeng Xu)</author><guid>https://xu-xp.com/posts/transformer_notes/</guid><description><![CDATA[<h2 id="基础介绍">基础介绍</h2>
<p>参考：</p>
<ol>
<li>
<p><a href="https://lilianweng.github.io/posts/2020-04-07-the-transformer-family/" target="_blank" rel="noopener noreffer ">https://lilianweng.github.io/posts/2020-04-07-the-transformer-family/</a></p>
</li>
<li>
<p><a href="https://zhuanlan.zhihu.com/p/351558402" target="_blank" rel="noopener noreffer ">https://zhuanlan.zhihu.com/p/351558402</a></p>
</li>
</ol>
<h3 id="注意力机制-attention">注意力机制 attention</h3>
<p>Attention 是神经网络中的一种机制：模型可以通过选择性地关注给定的数据集来学习做出预测。Attention的个数是通过学习权重来量化的，输出则通常是一个加权平均值。</p>]]></description></item><item><title>DL 深度学习笔记</title><link>https://xu-xp.com/posts/deep_learning/</link><pubDate>Mon, 14 Jun 2021 00:00:00 +0000</pubDate><author>xiaopeng.xu@kaust.edu.sa (Xiaopeng Xu)</author><guid>https://xu-xp.com/posts/deep_learning/</guid><description><![CDATA[<h2 id="nn-基础">NN 基础</h2>
<h3 id="二分类问题">二分类问题</h3>
<ul>
<li>对 64x64x3  的图片，判断是否有猫（1/0）</li>
</ul>
<h4 id="基本标记">基本标记</h4>
<ul>
<li>
<p></p>
</li>
<li>
<p>X 的每一行对应的是一个样本，每一列对应的是一个特征。</p>
</li>
</ul>
<h3 id="逻辑回归-logistics-regression">逻辑回归 (Logistics regression)</h3>
<ul>
<li>
<p></p>]]></description></item><item><title>图深度学习：成果、挑战与未来</title><link>https://xu-xp.com/posts/graph_dl/</link><pubDate>Fri, 20 Nov 2020 00:00:00 +0000</pubDate><author>xiaopeng.xu@kaust.edu.sa (Xiaopeng Xu)</author><guid>https://xu-xp.com/posts/graph_dl/</guid><description><![CDATA[<p><strong>本文最初发表在 TowardsDataScience 博客，经原作者 Michael Bronstein 授权，InfoQ 中文站翻译并分享。</strong></p>
<p><strong>2020-01-15</strong></p>
<p><strong>来源：</strong><a href="https://towardsdatascience.com/deep-learning-on-graphs-successes-challenges-and-next-steps-7d9ec220ba8" target="_blank" rel="noopener noreffer ">https://towardsdatascience.com/deep-learning-on-graphs-successes-challenges-and-next-steps-7d9ec220ba8</a></p>
<p>来自一个系列：<a href="https://towardsdatascience.com/graph-deep-learning/home" target="_blank" rel="noopener noreffer ">https://towardsdatascience.com/graph-deep-learning/home</a></p>
<h2 id="引言--介绍"><strong>引言 &amp; 介绍</strong></h2>
<p></p>]]></description></item></channel></rss>