T400 mic not working - Lenovo Community:
'via Blog this'
Blog Archive
-
▼
2015
(226)
-
▼
April
(19)
- T400 mic not working - Lenovo Community
- Robots Bring Couple Together, Engagement Ensues - ...
- Audacity label format
- NICT - Top page | NICT-National Institute of Infor...
- 国内基金业绩亮瞎眼。。。小小试水。。 - 未名空间(mitbbs.com)
- 关于CS,一些你不知道的东西(附加一些东西让准备转专业的人看 - 未名空间(mitbbs.com)
- 2015年度苏州高新区科技创新创业领军人才计划申报公告-首页-中国驻旧金山总领馆教育组
- LIUM Speaker Diarization Wiki [i-vector]
- userguide LIA_SpkDe
- new trick tmux
- 10次左右独立引用, 语音识别研究员NSC中心EB-1B获批! - 未名空间(mitbbs.com)
- Writing shell scripts - Lesson 13: Positional Para...
- 大数据在工业界流行的黑暗真相 - 未名空间(mitbbs.com)
- CHANGE HOSTNAME PERMANENTLY USING THE COMMAND LINE
- How to install Teamviewer on 14.04?
- Snapchat年薪百万招聘毕业生 - 未名空间(mitbbs.com) - 海外华人第一门户
- LDL Cholesterol: The Test
- Speech Signal Processing Toolkit (SPTK)
- GMM on Hadoop
-
▼
April
(19)
Thursday, April 30, 2015
Audacity label format
Example:
Audacity_label.txt
29.473812 49.066307 text1
57.213861 96.885166 text2
Note:
column1 and column2 indicates starting and ending time in second, respectively
save the label file as txt
Reference:
http://manual.audacityteam.org/o/man/label_tracks.html
Audacity_label.txt
29.473812 49.066307 text1
57.213861 96.885166 text2
Note:
column1 and column2 indicates starting and ending time in second, respectively
save the label file as txt
Reference:
http://manual.audacityteam.org/o/man/label_tracks.html
Tuesday, April 28, 2015
Sunday, April 26, 2015
Saturday, April 25, 2015
关于CS,一些你不知道的东西(附加一些东西让准备转专业的人看 - 未名空间(mitbbs.com)
关于CS,一些你不知道的东西(附加一些东西让准备转专业的人看 - 未名空间(mitbbs.com): "为了H1?H1最多只有6年,满了你还是得滚蛋。如果你不喜欢CS,这条路真心没必要。
为了绿卡?之前我说过了CS拿绿卡的难度,你不要拿着“某几个"人可能有天赋转了来
忽悠所有人转。经"
'via Blog this'
为了绿卡?之前我说过了CS拿绿卡的难度,你不要拿着“某几个"人可能有天赋转了来
忽悠所有人转。经"
'via Blog this'
Tuesday, April 21, 2015
Sunday, April 19, 2015
Thursday, April 16, 2015
userguide LIA_SpkDe
http://mistral.univ-avignon.fr/doc/userguide_LIA_SpkDet.002.pdf
Wednesday, April 15, 2015
new trick tmux
tmux at -t 1
ctrl b + d
ctrl b + c: new windows
ctr b + num:
tmux list-sessions
ctrl b + d
ctrl b + c: new windows
ctr b + num:
tmux list-sessions
Monday, April 13, 2015
Writing shell scripts - Lesson 13: Positional Parameters
http://linuxcommand.org/wss0130.php
As we discussed before, many programs, particularly ones from the GNU Project, support both short and long command line options. For example, to display a help message for many of these programs, you may use either the "-h" option or the longer "--help" option. Long option names are typically preceded by a double dash. We will adopt this convention for our scripts.Here is the code we will use to process our command line:interactive=
filename=~/system_page.html
while [ "$1" != "" ]; do
case $1 in
-f | --file ) shift
filename=$1
;;
-i | --interactive ) interactive=1
;;
-h | --help ) usage
exit
;;
* ) usage
exit 1
esac
shift
done
interactive= filename=~/system_page.html while [ "$1" != "" ]; do case $1 in -f | --file ) shift filename=$1 ;; -i | --interactive ) interactive=1 ;; -h | --help ) usage exit ;; * ) usage exit 1 esac shift done
Saturday, April 11, 2015
大数据在工业界流行的黑暗真相 - 未名空间(mitbbs.com)
大数据在工业界流行的黑暗真相 - 未名空间(mitbbs.com):
这两天在faculty版上灌水纠结大数据概念,让我突然想明白了大数据在工业界兴起的
原因:
很多规律在数据量大了以后都会变得很显然,算法再傻B,在超大量数据作用下效果也
会变得很好。因为开发牛B算法开销很大风险很高,而获取数据对不少公司而言很容易。
所以会出现一个大数据潮流,大家都指着数据变成救命法宝。
比如大数据的源头Google,Pagerank其实就是个幌子,真正优化Google搜索结果的其实
是人民群众。结果被点得多了排名就上去了。算法够傻B吧。还有Google翻译也用到了
大数据技术。算法是否傻B是和系统复杂度相对的。比如linear regression,数据量小用
atlab三五行就写出来了。数据量大了就得上Hadoop, Spark,支持运算的平台比真正的
算法要复杂得多得多。
但真相果真如此吗?做过machine learning的人都知道,数据量增大对于结果改善的边
际效应是递减的,而且往往是指数递减的。需要获得10个百分点的提高,可能需要成千
上万,甚至亿万倍的数据量。真是有点愚公移山的感觉了。
大数据的驱动力背后还有更黑暗的真相:提高barrier to entry。这世上聪明人很多。
工业界忽悠算法其实是最危险的。因为算法的发明往往是单枪匹马干的。如果一个大公
司建立在一个算法的基础之上,岂不是有被但强匹马干掉的危险?特别是学术界虽然不
济,几百年才出一个牛顿爱因斯坦这样的人物,但方法论一直是正确的,难保不出一两
个真正管用的新算法。现在告诉你,你没个一千台机器就没法做研究,把智力门槛转换
成资本门槛,并且搞
一个虚无飘渺的概念,把funding标准从科研水准转换成忽悠水准。以此打击真正的科
学研究,试图多维持几年大公司的盈利。其用心不可不谓险恶。
此计一出,大合那些脑子已经使不动了的学霸的心意,于是大数据一夜之间就火了。
'via Blog this'
这两天在faculty版上灌水纠结大数据概念,让我突然想明白了大数据在工业界兴起的
原因:
很多规律在数据量大了以后都会变得很显然,算法再傻B,在超大量数据作用下效果也
会变得很好。因为开发牛B算法开销很大风险很高,而获取数据对不少公司而言很容易。
所以会出现一个大数据潮流,大家都指着数据变成救命法宝。
比如大数据的源头Google,Pagerank其实就是个幌子,真正优化Google搜索结果的其实
是人民群众。结果被点得多了排名就上去了。算法够傻B吧。还有Google翻译也用到了
大数据技术。算法是否傻B是和系统复杂度相对的。比如linear regression,数据量小用
atlab三五行就写出来了。数据量大了就得上Hadoop, Spark,支持运算的平台比真正的
算法要复杂得多得多。
但真相果真如此吗?做过machine learning的人都知道,数据量增大对于结果改善的边
际效应是递减的,而且往往是指数递减的。需要获得10个百分点的提高,可能需要成千
上万,甚至亿万倍的数据量。真是有点愚公移山的感觉了。
大数据的驱动力背后还有更黑暗的真相:提高barrier to entry。这世上聪明人很多。
工业界忽悠算法其实是最危险的。因为算法的发明往往是单枪匹马干的。如果一个大公
司建立在一个算法的基础之上,岂不是有被但强匹马干掉的危险?特别是学术界虽然不
济,几百年才出一个牛顿爱因斯坦这样的人物,但方法论一直是正确的,难保不出一两
个真正管用的新算法。现在告诉你,你没个一千台机器就没法做研究,把智力门槛转换
成资本门槛,并且搞
一个虚无飘渺的概念,把funding标准从科研水准转换成忽悠水准。以此打击真正的科
学研究,试图多维持几年大公司的盈利。其用心不可不谓险恶。
此计一出,大合那些脑子已经使不动了的学霸的心意,于是大数据一夜之间就火了。
'via Blog this'
Tuesday, April 7, 2015
CHANGE HOSTNAME PERMANENTLY USING THE COMMAND LINE
Ubuntu – Change Hostname Permanently Using the Command Line | Michael Lane's Blog: "#!/bin/bash
#Assign existing hostname to $hostn
hostn=$(cat /etc/hostname)
#Display existing hostname
echo "Existing hostname is $hostn"
#Ask for new hostname $newhost
echo "Enter new hostname: "
read newhost
#change hostname in /etc/hosts & /etc/hostname
sudo sed -i "s/$hostn/$newhost/g" /etc/hosts
sudo sed -i "s/$hostn/$newhost/g" /etc/hostname
#display new hostname
echo "Your new hostname is $newhost"
#Press a key to reboot
read -s -n 1 -p "Press any key to reboot"
sudo reboot"
https://pricklytech.wordpress.com/2013/04/24/ubuntu-change-hostname-permanently-using-the-command-line/
'via Blog this'
#Assign existing hostname to $hostn
hostn=$(cat /etc/hostname)
#Display existing hostname
echo "Existing hostname is $hostn"
#Ask for new hostname $newhost
echo "Enter new hostname: "
read newhost
#change hostname in /etc/hosts & /etc/hostname
sudo sed -i "s/$hostn/$newhost/g" /etc/hosts
sudo sed -i "s/$hostn/$newhost/g" /etc/hostname
#display new hostname
echo "Your new hostname is $newhost"
#Press a key to reboot
read -s -n 1 -p "Press any key to reboot"
sudo reboot"
https://pricklytech.wordpress.com/2013/04/24/ubuntu-change-hostname-permanently-using-the-command-line/
'via Blog this'
How to install Teamviewer on 14.04?
Reference:
http://askubuntu.com/questions/453157/how-to-install-teamviewer-on-14-04
First, install gdebi:
In the same directory you downloaded the .deb file just run:
Option 2:
http://askubuntu.com/questions/453157/how-to-install-teamviewer-on-14-04
First, install gdebi:
sudo dpkg --add-architecture i386;
sudo apt-get update;
sudo apt-get install gdebi
sudo apt-get -f install
In the same directory you downloaded the .deb file just run:
sudo gdebi teamviewer_linux.deb
#replace teamviewer_linux.deb
with your downloaded debOption 2:
- Download the official image from http://www.teamviewer.com/en/download/linux.aspx (always get 32bit) http://download.teamviewer.com/download/teamviewer_i386.deb [May 4,2015]
- Open a terminal
- Run command
cd ~/Downloads/
- Run command
sudo dpkg -i teamviewer_i386.deb
- Follow whatever steps it needs you to take.
Note:
In case you encounter any error and depends on your linux machine, you may need install
libjpeg62.
libsm6
Option 3:
Install Xmanager
'via Blog this'
Saturday, April 4, 2015
Friday, April 3, 2015
LDL Cholesterol: The Test
LDL Cholesterol: The Test: "LDL Cholesterol"
http://labtestsonline.org/understanding/analytes/ldl/tab/test/
https://www.pritikin.com/your-health/health-benefits/lower-cholesterol/1468-7-tips-for-improving-your-ldl-cholesterol.html#.VR9fbObF_9k
'via Blog this'
http://labtestsonline.org/understanding/analytes/ldl/tab/test/
https://www.pritikin.com/your-health/health-benefits/lower-cholesterol/1468-7-tips-for-improving-your-ldl-cholesterol.html#.VR9fbObF_9k
The Test
- How is it used?
- When is it ordered?
- What does the test result mean?
- Is there anything else I should know?
How is it used?
The test for low-density lipoprotein cholesterol (LDL-C) is used as part of a lipid profile to predict an individual's risk of developing heart disease. Of all the forms of cholesterol in the blood, the LDL cholesterol is considered the most important form in determining risk of heart disease. Since treatment decisions are often based on LDL values, this test may be used to monitor levels after the start of diet or exercise programs or to determine whether or not prescribing one of the lipid-lowering drugs, such as statins, would be useful.
In addition to measuring the amount of LDL-C in blood, a test that measures the number of LDL particles (LDL-P) and/or their size may be useful in helping to determine risk of heart disease in certain people, according to some recent studies. For more on this, see Common Questions #1 and the article on Lipoprotein Subfraction Testing.
When is it ordered?
LDL-C levels are ordered as part a lipid profile, along with total cholesterol, HDL-C, and triglycerides. This profile may be ordered as a screening profile for a healthy person as part of a routine physical exam. It is recommended that all adults be tested at least once every five years. A fasting lipid profile may be ordered more frequently for those who have one or more major risk factor for heart disease (see below). It may be ordered for someone who has had a high screening cholesterol result to see if the total cholesterol is high because of too much LDL-C.
For children and adolescents, routine lipid testing is recommended by the American Academy of Pediatrics (AAP) once between the ages of 9 and 11 and again between the ages of 17 and 21. Screening with a lipid profile is recommended for children and youth who are at increased risk of developing heart disease as adults. Some of the risk factors are similar to those in adults and include a family history of heart disease or health problems such as diabetes, high blood pressure, or being overweight. High-risk children should have their first lipid profile between 2 and 8 years of age. Children younger than 2 years old are too young to be tested.
LDL-C levels may also be ordered at regular intervals to evaluate the success of lipid-lowering lifestyle changes, such as diet and exercise, or to determine the effectiveness of drug therapy such as statins.
What does the test result mean?
Elevated levels of LDL cholesterol can indicate risk for heart disease, so an LDL-C result is evaluated with respect to the upper limits that are desired based on individual risk factors. According to the National Cholesterol Education Program, if a person has no other risk factors, an LDL-C level can be evaluated as follows:
- Less than 100 mg/dL (2.59 mmol/L) — Optimal
- 100-129 mg/dL (2.59-3.34 mmol/L) — Near optimal, above optimal
- 130-159 mg/dL (3.37-4.12 mmol/L) — Borderline high
- 160-189 mg/dL (4.15-4.90 mmol/L) — High
- Greater than 189 mg/dL (4.90 mmol/L) — Very high
Desired goals for LDL levels change based on individual risk factors (listed below). Lifestyle changes, such as altering diet and exercise, are always recommended as treatment to lower elevated levels of LDL to target LDL cholesterol values. Certain combinations of LDL-C levels and individual risk factors for heart disease may warrant treatment with cholesterol-lowering drugs, such as statins, in addition to lifestyle changes.
Target values based on risk factors are:
- LDL-C less than 100 mg/dL (2.59 mmol/L) with heart disease or diabetes.*
- LDL-C less than 130 mg/dL (3.37 mmol/L) with 2 or more risk factors (intermediate risk for heart disease).
- LDL-C less than 160 mg/dL (4.14 mmol/L) with 0 or 1 risk factor (low risk for heart disease).
*Some organizations recommend that LDL-C be less than 70 mg/dL (1.82 mmol/L) if a person has heart disease or has had a heart attack.
Major risk factors for heart disease include:
- Cigarette smoking
- Age (males 45 years or older or females 55 years or older)
- High blood pressure (blood pressure of 140/90 or higher or taking high blood pressure medications)
- Family history of premature heart disease (heart disease in a first degree male relative under age 55 or a first degree female relative under age 65)
- Pre-existing coronary heart disease
- Diabetes mellitus
[Note: High HDL-C (60 mg/dL or above) is considered a "negative risk factor" and its presence allows the removal of one risk factor from the total.]
According to the American Academy of Pediatrics, the LDL-C level can be evaluated for youth with no other risk factors as follows:
For children and teens:
- Less than 110 mg/dL (2.85 mmol/L) — Acceptable
- 110-129 mg/dL (2.85-3.34 mmol/L) — Borderline high
- Greater than 130 mg/dL (3.36 mmol/L) — High
For young adults:
- Less than 120 mg/dL (3.10 mmol/L) — Acceptable
- 120-159 mg/dL (3.10-4.11 mmol/L) — Borderline high
- Greater than 160 mg/dL (4.12 mmol/L) — High
Low levels of LDL cholesterol are not generally a concern and are not monitored. They may be seen in people with an inherited lipoprotein deficiency and in people with hyperthyroidism, infection, inflammation, or cirrhosis.
Is there anything else I should know?
LDL cholesterol should be measured when a person is not ill. LDL cholesterol is temporarily low during acute illness, immediately following a heart attack, or during stress (like from surgery or an accident). Wait at least six weeks after any illness to have LDL-C measured.
Certain types of prescription drugs may raise or lower LDL cholesterol levels. Inform your health care provider of any drugs or supplements that you are taking before testing.
In women, LDL cholesterol usually rises during pregnancy. Women should wait at least six weeks after having a baby to have LDL-C measured.
'via Blog this'
Wednesday, April 1, 2015
GMM on Hadoop
Source:
http://www.eie.polyu.edu.hk/~mwmak/Download.htm
1. GMM on Hadoop
This is a set of Java programs that I developed recently for teaching the subject "Distributed Systems and Cloud Computing". The package contains both sequential and parallel version of the EM algorithm for training Gaussian mixture models (GMM). The parallel version is designed to run on Hadoop clusters. For the source code, visithttps://github.com/enmwmak/GMM-on-Hadoop
'via Blog this'
http://www.eie.polyu.edu.hk/~mwmak/Download.htm
1. GMM on Hadoop
This is a set of Java programs that I developed recently for teaching the subject "Distributed Systems and Cloud Computing". The package contains both sequential and parallel version of the EM algorithm for training Gaussian mixture models (GMM). The parallel version is designed to run on Hadoop clusters. For the source code, visithttps://github.com/enmwmak/GMM-on-Hadoop
'via Blog this'
Subscribe to:
Posts (Atom)