<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[Work Hard, Play Harder!]]></title><description><![CDATA[Hello Deep Learning! Hello MS!]]></description><link>https://blog.csdn.net/weixingstudio</link><language>zh-cn</language><generator>https://blog.csdn.net/</generator><copyright><![CDATA[Copyright &copy; weixingstudio]]></copyright><item><title><![CDATA[Understanding LSTM Networks]]></title><link>https://blog.csdn.net/weixingstudio/article/details/53066436</link><guid>https://blog.csdn.net/weixingstudio/article/details/53066436</guid><author>weixingstudio</author><pubDate>Mon, 07 Nov 2016 15:06:45 +0800</pubDate><description><![CDATA[本文转载于：http://colah.github.io/posts/2015-08-Understanding-LSTMs/

Understanding LSTM Networks


Posted on August 27, 2015



Recurrent Neural Networks

Humans don’t start their thinking from]]></description><category></category></item><item><title><![CDATA[Elasticlunr.js 支持其他语言 V0.9.5]]></title><link>https://blog.csdn.net/weixingstudio/article/details/52485447</link><guid>https://blog.csdn.net/weixingstudio/article/details/52485447</guid><author>weixingstudio</author><pubDate>Fri, 09 Sep 2016 11:19:03 +0800</pubDate><description><![CDATA[之前一直没有处理其他的语言的需求，所以没有测试elasticlunr.js对于其他的语言的支持。多亏了Github的网友，帮忙发现了elasticlunr.js对于其他语言支持的问题。昨天，elasticlunr.js发布了V0.9.5版本，利用lunr-languages已有的对于其他语言的支持功能，使得elasticlunr.js能够支持目前所有的lunr-languages支持的语言。目前lu]]></description><category></category></item><item><title><![CDATA[Evaluation of Deep Learning Toolkits]]></title><link>https://blog.csdn.net/weixingstudio/article/details/52297861</link><guid>https://blog.csdn.net/weixingstudio/article/details/52297861</guid><author>weixingstudio</author><pubDate>Wed, 24 Aug 2016 10:36:03 +0800</pubDate><description><![CDATA[本文转载于Evaluation of Deep Learning ToolkitsEvaluation of Deep Learning ToolkitsAbstract. In this study, I evaluate some popular deep learning toolkits. The candidates are listed in alphabetical order: Ca]]></description><category></category></item><item><title><![CDATA[learn word2vector from scratch -- part 4]]></title><link>https://blog.csdn.net/weixingstudio/article/details/52213799</link><guid>https://blog.csdn.net/weixingstudio/article/details/52213799</guid><author>weixingstudio</author><pubDate>Mon, 15 Aug 2016 19:01:16 +0800</pubDate><description><![CDATA[本文详细介绍了word2vector的算法原理，公式推导，参数更新的公式推导。

一共分为4个部分，为了保证每个网页不是太长，分为4个网页分别给出。]]></description><category></category></item><item><title><![CDATA[learn word2vector from scratch -- part 3]]></title><link>https://blog.csdn.net/weixingstudio/article/details/52213776</link><guid>https://blog.csdn.net/weixingstudio/article/details/52213776</guid><author>weixingstudio</author><pubDate>Mon, 15 Aug 2016 18:57:32 +0800</pubDate><description><![CDATA[本文详细介绍了word2vector的算法原理，公式推导，参数更新的公式推导。

一共分为4个部分，为了保证每个网页不是太长，分为4个网页分别给出。]]></description><category></category></item><item><title><![CDATA[learn word2vector from scratch -- part 2]]></title><link>https://blog.csdn.net/weixingstudio/article/details/52213761</link><guid>https://blog.csdn.net/weixingstudio/article/details/52213761</guid><author>weixingstudio</author><pubDate>Mon, 15 Aug 2016 18:52:32 +0800</pubDate><description><![CDATA[本文详细介绍了word2vector的算法原理，公式推导，参数更新的公式推导。
一共分为4个部分，为了保证每个网页不是太长，分为4个网页分别给出。]]></description><category></category></item><item><title><![CDATA[learn word2vector from scratch -- part 1]]></title><link>https://blog.csdn.net/weixingstudio/article/details/52213717</link><guid>https://blog.csdn.net/weixingstudio/article/details/52213717</guid><author>weixingstudio</author><pubDate>Mon, 15 Aug 2016 18:49:23 +0800</pubDate><description><![CDATA[本文详细介绍了word2vector的算法原理，公式推导，参数更新的公式推导。
一共分为4个部分，为了保证每个网页不是太长，分为4个网页分别给出。]]></description><category></category></item><item><title><![CDATA[javascript全文检索工具]]></title><link>https://blog.csdn.net/weixingstudio/article/details/52150052</link><guid>https://blog.csdn.net/weixingstudio/article/details/52150052</guid><author>weixingstudio</author><pubDate>Mon, 08 Aug 2016 11:05:31 +0800</pubDate><description><![CDATA[https://github.com/weixsong/elasticlunr.js


javascript全文检索工具，有兴趣的小伙伴请帮忙点赞。





Elasticlunr.js

Lightweight full-text search engine in Javascript for browser search and offline search.
El]]></description><category></category></item><item><title><![CDATA[S2S (sequence to sequence) 算法理解]]></title><link>https://blog.csdn.net/weixingstudio/article/details/52130236</link><guid>https://blog.csdn.net/weixingstudio/article/details/52130236</guid><author>weixingstudio</author><pubDate>Fri, 05 Aug 2016 16:18:53 +0800</pubDate><description><![CDATA[最近S2S火得一塌糊涂，不得不学习一下]]></description><category></category></item><item><title><![CDATA[使用深度学习打造智能聊天机器人]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51892027</link><guid>https://blog.csdn.net/weixingstudio/article/details/51892027</guid><author>weixingstudio</author><pubDate>Tue, 12 Jul 2016 19:54:55 +0800</pubDate><description><![CDATA[原文转载于：http://geek.csdn.net/news/detail/85091




作者：张俊林，中科院软件所博士，技术书籍《这就是搜索引擎：核心技术详解》、《大数据日知录：架构与算法》作者。曾担任阿里巴巴、百度、新浪微博资深技术专家，目前是用友畅捷通工智能相关业务负责人，关注深度学习在自然语言处理方面的应用。 
责编：周建丁（zhoujd@csdn.net） 
本文为]]></description><category></category></item><item><title><![CDATA[Understanding LSTM Networks]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51840967</link><guid>https://blog.csdn.net/weixingstudio/article/details/51840967</guid><author>weixingstudio</author><pubDate>Wed, 06 Jul 2016 17:28:40 +0800</pubDate><description><![CDATA[Understanding LSTM Networks


Posted on August 27, 2015



Recurrent Neural Networks

Humans don’t start their thinking from scratch every second. As you read this essay, you understand each ]]></description><category></category></item><item><title><![CDATA[Hinton's Dropout in 3 Lines of Python]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51790632</link><guid>https://blog.csdn.net/weixingstudio/article/details/51790632</guid><author>weixingstudio</author><pubDate>Thu, 30 Jun 2016 15:20:30 +0800</pubDate><description><![CDATA[Hinton's Dropout in 3 Lines of Python

How to install Dropout into a neural network by only changing 3 lines of python.
Posted by iamtrask on July 28, 2015






Summary: Dropout is a vital ]]></description><category></category></item><item><title><![CDATA[Learning RNN from scratch (ＲＮＮ神经网络参数推导)]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51773524</link><guid>https://blog.csdn.net/weixingstudio/article/details/51773524</guid><author>weixingstudio</author><pubDate>Tue, 28 Jun 2016 09:49:01 +0800</pubDate><description><![CDATA[从上一篇原创的文章到现在，已经有一年多了，目前终于有一些新的总结分享个大家。


本文主要讲了RNN神经网络的基本原理，并给出了RNN神经网络参数更新的详细推导过程（back propagation），对于想知道RNN的参数是如果推导的，可以仔细阅读本文。
因为时间有限，下面的总结难免有疏漏之处，请大家指正。


本文结合了一个非常经典的RNN的例子代码，进行了详细的说明，RNN的代]]></description><category></category></item><item><title><![CDATA[Cross entropy]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51755217</link><guid>https://blog.csdn.net/weixingstudio/article/details/51755217</guid><author>weixingstudio</author><pubDate>Fri, 24 Jun 2016 19:15:28 +0800</pubDate><description><![CDATA[In information theory, the cross entropy between two probability
 distributions {\displaystyle p} and {\displaystyle
 q} over
 the same underlying set of events measures the average number of bits ]]></description><category></category></item><item><title><![CDATA[RNN computation]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51647256</link><guid>https://blog.csdn.net/weixingstudio/article/details/51647256</guid><author>weixingstudio</author><pubDate>Sun, 12 Jun 2016 16:33:20 +0800</pubDate><description><![CDATA[link: http://karpathy.github.io/2015/05/21/rnn-effectiveness/RNN computation. So how do these things work? At the core, RNNs have a deceptively simple API: They accept an input vector x and give you an]]></description><category></category></item><item><title><![CDATA[The Max Trick when Computing Softmax]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51644808</link><guid>https://blog.csdn.net/weixingstudio/article/details/51644808</guid><author>weixingstudio</author><pubDate>Sun, 12 Jun 2016 11:16:21 +0800</pubDate><description><![CDATA[The softmax function appears in many machine learning algorithms. The idea is, if you have a set of values, to scale them so they sum to 1.0 and therefore can be interpreted as probabilities.For exampl]]></description><category></category></item><item><title><![CDATA[大端序与小端序]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51395511</link><guid>https://blog.csdn.net/weixingstudio/article/details/51395511</guid><author>weixingstudio</author><pubDate>Fri, 13 May 2016 15:06:22 +0800</pubDate><description><![CDATA[原文转载于：http://blog.csdn.net/shanshanpt/article/details/7365613其实 大小端序是针对于不同的硬件而言的，具体的就是不同的CPU 是不同的处理的。 对于 Intel x86 系列就是小端序。下面具体介绍：简单点说，就是字节的存储顺序，如果数据都是单字节的，那怎么存储无所谓了，但是对于多字节数据，比如int，double等，就要考虑存储的顺序了。]]></description><category></category></item><item><title><![CDATA[前端搜索JS库 Elasticlunr.js]]></title><link>https://blog.csdn.net/weixingstudio/article/details/51312176</link><guid>https://blog.csdn.net/weixingstudio/article/details/51312176</guid><author>weixingstudio</author><pubDate>Wed, 04 May 2016 09:12:52 +0800</pubDate><description><![CDATA[有Github账号的小伙伴们，帮忙点赞啦(Star)本人写了一个JS的离线搜索库，支持Field search, 采用BM25模型进行Ranking, Ranking的算法和Elasticsearch一样的，效果比lunr.js有很高提升， 希望小伙伴们能够给与肯定，帮忙点赞。Elasticlunr.js is a lightweight full-text search engine in Jav]]></description><category></category></item><item><title><![CDATA[UNDERSTANDING CONVOLUTIONAL NEURAL NETWORKS FOR NLP]]></title><link>https://blog.csdn.net/weixingstudio/article/details/50543679</link><guid>https://blog.csdn.net/weixingstudio/article/details/50543679</guid><author>weixingstudio</author><pubDate>Tue, 19 Jan 2016 17:04:50 +0800</pubDate><description><![CDATA[原文转载于:http://www.wildml.com/2015/11/understanding-convolutional-neural-networks-for-nlp/




When we hear about Convolutional Neural Network (CNNs), we typically think of Computer Vision. CNNs we]]></description><category></category></item><item><title><![CDATA[python 装饰器]]></title><link>https://blog.csdn.net/weixingstudio/article/details/50232403</link><guid>https://blog.csdn.net/weixingstudio/article/details/50232403</guid><author>weixingstudio</author><pubDate>Wed, 09 Dec 2015 09:54:02 +0800</pubDate><description><![CDATA[原文转载于：http://www.liaoxuefeng.com/wiki/001374738125095c955c1e6d8bb493182103fac9270762a000/001386819879946007bbf6ad052463ab18034f0254bf355000




由于函数也是一个对象，而且函数对象可以被赋值给变量，所以，通过变量也能调用该函数。
>>> def ]]></description><category></category></item></channel></rss>