<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[AugustMoore的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/AugustMoore</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; AugustMoore]]></copyright><item><title><![CDATA[Error using sumInvalid option.]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/126647612</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/126647612</guid><author>AugustMoore</author><pubDate>Thu, 01 Sep 2022 17:21:30 +0800</pubDate><description><![CDATA[Error using sumInvalid option. ]]></description><category></category></item><item><title><![CDATA[Linux程序安装创建桌面图标]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/106209727</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/106209727</guid><author>AugustMoore</author><pubDate>Tue, 19 May 2020 10:54:24 +0800</pubDate><description><![CDATA[以FileZilla为例，从安装程序到创建快捷方式。
安装程序
冲官方网站找到相应版本的下载链接。
FileZilla Linux
下载完成后，打开终端，输入安装指令：
sudo apt-get install filezilla
检查安装是否成功可以看显示所有安装的程序栏。
创建桌面快捷方式和图标

在桌面新建文件， 命名为 Filezilla.desktop
查看 FileZilla 安装位置：

which filezilla
返回位置：
/usr/bin/filezilla
查看 FileZilla]]></description><category></category></item><item><title><![CDATA[Moreau Decomposition 和 共轭次梯度定理]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/104896801</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/104896801</guid><author>AugustMoore</author><pubDate>Mon, 16 Mar 2020 15:58:17 +0800</pubDate><description><![CDATA[在学习临近梯度(proximal gradient)中遇到的问题。证明 Moreau Decompositon。


]]></description><category></category></item><item><title><![CDATA[抽象代数入门（四）]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/94567597</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/94567597</guid><author>AugustMoore</author><pubDate>Wed, 03 Jul 2019 16:08:55 +0800</pubDate><description><![CDATA[向量空间(vector space)

Definition of vector space: a vector space is a set closed under vector additon and scalar multiplication.

满足加法交换律、结合律、分配律，存在加法单位元
	满足标量乘法的交换律、结合律、分配律，存在乘法单位元
Cyclic Subspace(循环子空...]]></description><category></category></item><item><title><![CDATA[抽象代数入门（三）]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/94560032</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/94560032</guid><author>AugustMoore</author><pubDate>Wed, 03 Jul 2019 14:49:22 +0800</pubDate><description><![CDATA[一些域上多项式的概念

prime(素)

Dfn:Then p is prime.



Unit(单位)

Dfn： If is unit, and, then:. A given field may have infinite units. 表示minimal and monoic polynomial of g.



Relative prime(相对素)

Dfn: is...]]></description><category></category></item><item><title><![CDATA[抽象代数入门（二）]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/94452817</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/94452817</guid><author>AugustMoore</author><pubDate>Tue, 02 Jul 2019 21:48:01 +0800</pubDate><description><![CDATA[Field(域)

Dfn: ”Field“ is a commutative ring, s.t. every non-zero element has inverse.

这个交换环对于乘法和加法都满足交换性，而且对于所有非零元素都存在逆。

对加法逆，即
	对乘法逆，即
这两个都满足，u是对应元素的逆。加法的逆写成-v, 乘法的逆写成都是可以的，都是符号表示。

Polynomials o...]]></description><category></category></item><item><title><![CDATA[抽象代数入门（一）]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/94384032</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/94384032</guid><author>AugustMoore</author><pubDate>Mon, 01 Jul 2019 12:23:22 +0800</pubDate><description><![CDATA[一、群论

幺半群(monoid)

之前看老师讲的叫Abelian monoid（阿贝尔幺半群），但是搜不到。

A monoid is a set closed under an associative binary operation and has an identity element such that.

Binary operation(二元运算): f(x, y) an op...]]></description><category></category></item><item><title><![CDATA[凸优化笔记（非常零碎）]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/89431349</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/89431349</guid><author>AugustMoore</author><pubDate>Sun, 21 Apr 2019 13:33:02 +0800</pubDate><description><![CDATA[1. Lipschitz Funcion（李比西斯函数）

对于不光滑(nonsomooth)的凸函数的分析引入了Lipschitz functions:

定义：

2.Nonexpansiveness of Projection(投影的非延展性）

Projection operator is non-expansive means:



也就是说， 投影操作是L小于1的L-Lipschi...]]></description><category></category></item><item><title><![CDATA[python对于文件的读取和重建]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/89036277</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/89036277</guid><author>AugustMoore</author><pubDate>Thu, 04 Apr 2019 22:11:46 +0800</pubDate><description><![CDATA[这里我想把关于对各类型文件的读取方法以及对其中数据的操作，只要是我遇到的就纪录下来

一.读取

通常我会以open函数开头，以读取txt文件为例：


data = [] # I will store the data in this list.
with open(file, mode='r') as f:
    for line in f:
        line = line.repl...]]></description><category></category></item><item><title><![CDATA[Bandit UCB推导]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/86590197</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/86590197</guid><author>AugustMoore</author><pubDate>Tue, 22 Jan 2019 11:31:40 +0800</pubDate><description><![CDATA[推导Reinforcement Learning Richard S.Sutton and Andrew G. Barto 第二章Bandit算法中的Upper-Confidence-Bound Action Selection.

预备知识

Markov Inequality

对于任意r.v. (random variable) X and constant a &amp;gt; 0,



Prf...]]></description><category></category></item><item><title><![CDATA[pickle模块]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/86097091</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/86097091</guid><author>AugustMoore</author><pubDate>Tue, 08 Jan 2019 21:13:25 +0800</pubDate><description><![CDATA[pickle 实现序列化和反序列化。

protocol = 0, 用ASCII，protocol=1, 用 二进制。


&amp;gt;&amp;gt;&amp;gt; import pickle as P  # python 3
	    

1.pickle.dump(obj, file, protocol=None, *, fix_imports=True)

write a pickled represent...]]></description><category></category></item><item><title><![CDATA[PIL Image]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/86094134</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/86094134</guid><author>AugustMoore</author><pubDate>Tue, 08 Jan 2019 19:56:53 +0800</pubDate><description><![CDATA[
from PIL import Image
import numpy as np




1. Image.open(fp, mode='r')

根据路径 fp 读取图片，这里默认是只读格式。


&amp;gt;&amp;gt;&amp;gt; p = '/home/hushch/图片/th.jpeg'
&amp;gt;&amp;gt;&amp;gt; im = Image.open(p)
&amp;gt;&amp;gt;&amp;gt; type(im)
&amp;l...]]></description><category></category></item><item><title><![CDATA[os模块]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/86092492</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/86092492</guid><author>AugustMoore</author><pubDate>Tue, 08 Jan 2019 19:28:57 +0800</pubDate><description><![CDATA[1.os.getcwd()

Return an unicode string representing the current working directory.


&amp;gt;&amp;gt;&amp;gt; os.getcwd()
'/home/hushch'

2.os.path.join(a, *p)

Join two or more path components, inserting '/' as...]]></description><category></category></item><item><title><![CDATA[torch]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/86089514</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/86089514</guid><author>AugustMoore</author><pubDate>Tue, 08 Jan 2019 18:53:36 +0800</pubDate><description><![CDATA[
from PIL import Image
import numpy as np
import torch
import torchvision.transforms as T

1.From Tensor to Numpy:


&amp;gt;&amp;gt;&amp;gt; a=torch.Tensor([[1,2],[3,2]])
&amp;gt;&amp;gt;&amp;gt; a
tensor([[1., 2.],
       ...]]></description><category></category></item><item><title><![CDATA[Softmax Derivation]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/84671177</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/84671177</guid><author>AugustMoore</author><pubDate>Sat, 01 Dec 2018 11:03:46 +0800</pubDate><description><![CDATA[讨论最简单情况：

     以神经网络为例：

     假设在softmax层，输入的数据是N维的一维数组，输出结果是多分类的各个概率，假设为C类。

--1. input: x --&amp;gt; import data with dimension N, can be writen as ,  in neural network, means the last hidden layer outp...]]></description><category></category></item><item><title><![CDATA[Conditional Expectation & Entropy]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/84478984</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/84478984</guid><author>AugustMoore</author><pubDate>Sun, 25 Nov 2018 11:21:30 +0800</pubDate><description><![CDATA[Conditional Expectation(条件期望)

1.Derivation

假设有两个离散随机变量X，Y。他们的概率密度分别用 表示。则条件期望E(Y|X)为：



And:



 

 

Remark:

 是关于X的函数。有：

 

 ...]]></description><category></category></item><item><title><![CDATA[Linux安装指南]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/82925635</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/82925635</guid><author>AugustMoore</author><pubDate>Tue, 02 Oct 2018 19:19:59 +0800</pubDate><description><![CDATA[从Windows 的Microsoft Store下载Ubuntu

      直接在应用商店搜索即可。（好像才200M)

Linux 基本操作： http://man.linuxde.net/sudo

安装

安装前操作

在Micorsoft 搜索PowerShell.然后输入：

Enable-WindowsOptionalFeature -Online -FeatureName Mi...]]></description><category></category></item><item><title><![CDATA[np.unique()]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/81608906</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/81608906</guid><author>AugustMoore</author><pubDate>Sun, 12 Aug 2018 18:59:50 +0800</pubDate><description><![CDATA[np.unique(ar, return_index=False, return_inverse=False, return_counts=False,axis=None)


&amp;gt;&amp;gt;&amp;gt; import numpy as np
&amp;gt;&amp;gt;&amp;gt; a=np.random.normal(size=(10))
&amp;gt;&amp;gt;&amp;gt; a
array([-0.11357244,  ...]]></description><category></category></item><item><title><![CDATA[set函数的用法]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/81219832</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/81219832</guid><author>AugustMoore</author><pubDate>Thu, 26 Jul 2018 14:47:49 +0800</pubDate><description><![CDATA[set()是python中的一个内置函数。功能强大。可以用于数据的筛选。

1. set().add()  : 往set的后面加数据

2.set().pop()  ： 删除set的最后一个数据，并返回该值

3.set().discard()  ： 删除set中的某个值，可以指定。


a=[1,3,2,3,0]
b=set(a)  # {0,1,2,3}
b.add(0)  # {0,1,2,...]]></description><category></category></item><item><title><![CDATA[python 中decorator和property]]></title><link>https://blog.youkuaiyun.com/AugustMoore/article/details/81005267</link><guid>https://blog.youkuaiyun.com/AugustMoore/article/details/81005267</guid><author>AugustMoore</author><pubDate>Wed, 11 Jul 2018 20:31:12 +0800</pubDate><description><![CDATA[1.装饰器decorator实质是一个迭代器，把定义的函数作为一个属性来调用。具体来看：def hello(fn):
    print(1)
    def wrapper():
        print(2)
        print(&quot;hello, %s&quot; % fn.__name__)
        fn()  # 在这里会调用foo()
        print(&quot;goodby, ...]]></description><category></category></item></channel></rss>