深度 | 朴素贝叶斯模型算法研究与实例分析

深度 | 朴素贝叶斯模型算法研究与实例分析

(白宁超 2018年9月3日15: 56:20)

导读:朴素贝叶斯模型是机器学习常用的模型算法之一,其在文本分类方面简单易行,且取得不错的分类效果。所以很受欢迎,对于朴素贝叶斯的学习,本文首先介绍理论知识即朴素贝叶斯相关概念和公式推导,为了加深理解,采用一个维基百科上面性别分类例子进行形式化描述。然后通过编程实现朴素贝叶斯分类算法,并在屏蔽社区言论、垃圾邮件、个人广告中获取区域倾向等几个方面进行应用,包括创建数据集、数据预处理、词集模型和词袋模型、朴素贝叶斯模型训练和优化等。然后结合复旦大学新闻语料进行朴素贝叶斯的应用。最后,大家熟悉其原理和实现之后,采用机器学习sklearn包进行实现和优化。由于篇幅较长,采用理论理解、案例实现、sklearn优化三个部分进行学习。(本文原创,转载必须注明出处:朴素贝叶斯模型算法研究与实例分析)

继续阅读

实现 | 朴素贝叶斯模型算法研究与实例分析

实现 | 朴素贝叶斯模型算法研究与实例分析

(白宁超 2018年9月2日 11: 16:31)

 

导读:朴素贝叶斯模型是机器学习常用的模型算法之一,其在文本分类方面简单易行,且取得不错的分类效果。所以很受欢迎,对于朴素贝叶斯的学习,本文首先介绍理论知识即朴素贝叶斯相关概念和公式推导,为了加深理解,采用一个维基百科上面性别分类例子进行形式化描述。然后通过编程实现朴素贝叶斯分类算法,并在屏蔽社区言论、垃圾邮件、个人广告中获取区域倾向等几个方面进行应用,包括创建数据集、数据预处理、词集模型和词袋模型、朴素贝叶斯模型训练和优化等。然后结合复旦大学新闻语料进行朴素贝叶斯的应用。最后,大家熟悉其原理和实现之后,采用机器学习sklearn包进行实现和优化。由于篇幅较长,采用理论理解、案例实现、sklearn优化三个部分进行学习。(本文原创,转载必须注明出处:朴素贝叶斯模型算法研究与实例分析)

继续阅读

理论 | 朴素贝叶斯模型算法研究与实例分析

理论 | 朴素贝叶斯模型算法研究与实例分析

(白宁超 2018年9月4日10:00:31)

导读:朴素贝叶斯模型是机器学习常用的模型算法之一,其在文本分类方面简单易行,且取得不错的分类效果。所以很受欢迎,对于朴素贝叶斯的学习,本文首先介绍理论知识即朴素贝叶斯相关概念和公式推导,为了加深理解,采用一个维基百科上面性别分类例子进行形式化描述。然后通过编程实现朴素贝叶斯分类算法,并在屏蔽社区言论、垃圾邮件、个人广告中获取区域倾向等几个方面进行应用,包括创建数据集、数据预处理、词集模型和词袋模型、朴素贝叶斯模型训练和优化等。然后结合复旦大学新闻语料进行朴素贝叶斯的应用。最后,大家熟悉其原理和实现之后,采用机器学习sklearn包进行实现和优化。由于篇幅较长,采用理论理解、案例实现、sklearn优化三个部分进行学习。(本文原创,转载必须注明出处:朴素贝叶斯模型算法研究与实例分析)

继续阅读

KNN模型算法研究与案例分析

KNN模型算法研究与案例分析

(白宁超 2018年8月30日11:46:14)

导读:机器学习算法中KNN属于比较简单的典型算法,既可以做聚类又可以做分类使用。本文通过一个模拟的实际案例进行讲解。整个流程包括:采集数据、数据格式化处理、数据分析、数据归一化处理、构造算法模型、评估算法模型和算法模型的应用。(本文原创,转载必须注明出处)

继续阅读

决策树模型算法研究与案例分析

决策树模型算法研究与案例分析

(白宁超 2018年8月30日11:46:14)

导读:决策树算法是一种基本的分类与回归方法,是最经常使用的算法之一。决策树模型呈树形结构,在分类问题中,表示基于特征对实例进行分类的过程。它可以认为是基于规则的集合。本文首先介绍决策树定义、工作原理、算法流程、优缺点等,然后结合案例进行分析。(本文原创,转载必须注明出处)

继续阅读

Coursera上数据科学相关课程(公开课)汇总推荐

Coursera上的数据科学课程有很多,这里汇总一批。

1、 Introduction to Data Science Specialization

IBM公司推出的数据科学导论专项课程系列(Introduction to Data Science Specialization),这个系列包括4门子课程,涵盖数据科学简介,面向数据科学的开源工具,数据科学方法论,SQL基础,感兴趣的同学可以关注:Launch your career in Data Science。Data Science skills to prepare for a career or further advanced learning in Data Science.

1) What is Data Science?
2) Open Source tools for Data Science
3) Data Science Methodology
4) Databases and SQL for Data Science

2、Applied Data Science Specialization

IBM公司推出的 应用数据科学专项课程系列(Applied Data Science Specialization),这个系列包括4门子课程,涵盖面向数据科学的Python,Python数据可视化,Python数据分析,数据科学应用毕业项目,感兴趣的同学可以关注:Get hands-on skills for a Career in Data Science。Learn Python, analyze and visualize data. Apply your skills to data science and machine learning.

1) Python for Data Science
2) Data Visualization with Python
3) Data Analysis with Python
4) Applied Data Science Capstone

3、Applied Data Science with Python Specialization

密歇根大学的Python数据科学应用专项课程系列(Applied Data Science with Python),这个系列的目标主要是通过Python编程语言介绍数据科学的相关领域,包括应用统计学,机器学习,信息可视化,文本分析和社交网络分析等知识,并结合一些流行的Python工具包进行讲授,例如pandas, matplotlib, scikit-learn, nltk以及networkx等Python工具。感兴趣的同学可以关注:Gain new insights into your data-Learn to apply data science methods and techniques, and acquire analysis skills.

1) Introduction to Data Science in Python
2) Applied Plotting, Charting & Data Representation in Python
3) Applied Machine Learning in Python
4) Applied Text Mining in Python
5) Applied Social Network Analysis in Python

4、Data Science Specialization

约翰霍普金斯大学的数据科学专项课程系列(Data Science Specialization),这个系列课程有10门子课程,包括数据科学家的工具箱,R语言编程,数据清洗和获取,数据分析初探,可重复研究,统计推断,回归模型,机器学习实践,数据产品开发,数据科学毕业项目,感兴趣的同学可以关注: Launch Your Career in Data Science-A nine-course introduction to data science, developed and taught by leading professors.

1) The Data Scientist’s Toolbox
2) R Programming
3) Getting and Cleaning Data
4) Exploratory Data Analysis
5) Reproducible Research
6) Statistical Inference
7) Regression Models
8) Practical Machine Learning
9) Developing Data Products
10) Data Science Capstone

5、Data Science at Scale Specialization

华盛顿大学的大规模数据科学专项课程系列(Data Science at Scale ),这个系列包括3门子课程和1个毕业项目课程,包括大规模数据系统和算法,数据分析模型与方法,数据科学结果分析等,感兴趣的同学可以关注: Tackle Real Data Challenges-Master computational, statistical, and informational data science in three courses.

1) Data Manipulation at Scale: Systems and Algorithms
2) Practical Predictive Analytics: Models and Methods
3) Communicating Data Science Results
4) Data Science at Scale – Capstone Project

6、Advanced Data Science with IBM Specialization

IBM公司推出的高级数据科学专项课程系列(Advanced Data Science with IBM Specialization),这个系列包括4门子课程,涵盖数据科学基础,高级机器学习和信号处理,结合深度学习的人工智能应用等,感兴趣的同学可以关注:Expert in DataScience, Machine Learning and AI。Become an IBM-approved Expert in Data Science, Machine Learning and Artificial Intelligence.

1) Fundamentals of Scalable Data Science
2) Advanced Machine Learning and Signal Processing
3) Applied AI with DeepLearning
4) Advanced Data Science Capstone

7、Data Mining Specialization

伊利诺伊大学香槟分校的数据挖掘专项课程系列(Data Mining Specialization),这个系列包含5门子课程和1个毕业项目课程,涵盖数据可视化,信息检索,文本挖掘与分析,模式发现和聚类分析等,感兴趣的同学可以关注:Data Mining Specialization-Analyze Text, Discover Patterns, Visualize Data. Solve real-world data mining challenges.

1) Data Visualization
2) Text Retrieval and Search Engines
3) Text Mining and Analytics
4) Pattern Discovery in Data Mining
5) Cluster Analysis in Data Mining
6) Data Mining Project

8、Data Analysis and Interpretation Specialization

数据分析和解读专项课程系列(Data Analysis and Interpretation Specialization),该系列包括5门子课程,分别是数据管理和可视化,数据分析工具,回归模型,机器学习,毕业项目,感兴趣的同学可以关注:Learn Data Science Fundamentals-Drive real world impact with a four-course introduction to data science.

1) Data Management and Visualization
2) Data Analysis Tools
3) Regression Modeling in Practice
4) Machine Learning for Data Analysis
5) Data Analysis and Interpretation Capstone

9、Executive Data Science Specialization

可管理的数据科学专项课程系列(Executive Data Science Specialization),这个系列包含4门子课程和1门毕业项目课程,涵盖数据科学速成,数据科学小组建设,数据分析管理,现实生活中的数据科学等,感兴趣的同学可以关注:Be The Leader Your Data Team Needs-Learn to lead a data science team that generates first-rate analyses in four courses.

1)A Crash Course in Data Science
2)Building a Data Science Team
3)Managing Data Analysis
4)Data Science in Real Life
5)Executive Data Science Capstone

10、其他相关的数据科学课程

1) Data Science Math Skills
2) Data Science Ethics
3) How to Win a Data Science Competition: Learn from Top Kagglers

注:本文首发“课程图谱博客”:http://blog.coursegraph.com

同步发布到这里, 本本文链接地址:http://blog.coursegraph.com/coursera上数据科学相关课程数据科学公开课汇总推荐 http://blog.coursegraph.com/?p=851

专利文本数据挖掘之AIpatent

这两年,我花了很多时间在专利文本数据挖掘上,这是一件很好玩的事情。目前我们的产品陆续上线了,感兴趣的朋友可以关注:

AIpatent专利翻译引擎http://t.aipatent.com

AIpatent专利科技词典http://d.aipatent.com/

AIpatent专利情报信息http://x.aipatent.com/

接下来,还有好玩的AIpatent专利检索产品,敬请期待。

从零开始搭建深度学习服务器: 1080TI四卡并行(Ubuntu16.04+CUDA9.2+cuDNN7.1+TensorFlow+Keras)

这个系列写了好几篇文章,这是相关文章的索引,仅供参考:

最近公司又弄了一套4卡1080TI机器,配置基本上和之前是一致的,只是显卡换成了技嘉的伪公版1080TI:技嘉GIGABYTE GTX1080Ti 涡轮风扇108TTURBO-11GD

部件	型号	价格	链接	备注CPU	英特尔(Intel)酷睿六核i7-6850K 盒装CPU处理器 	4599	http://item.jd.com/11814000696.html	散热器	美商海盗船 H55 水冷	449	https://item.jd.com/10850633518.html	主板	华硕(ASUS)华硕 X99-E WS/USB 3.1工作站主板	4759	内存	美商海盗船(USCORSAIR) 复仇者LPX DDR4 3000 32GB(16Gx4条)  	2799 * 2	https://item.jd.com/1990572.html	SSD	三星(SAMSUNG) 960 EVO 250G M.2 NVMe 固态硬盘	599	https://item.jd.com/3739097.html		硬盘	希捷(SEAGATE)酷鱼系列 4TB 5900转 台式机机械硬盘 * 2 	629 * 2	https://item.jd.com/4220257.html	电源	美商海盗船 AX1500i 全模组电源 80Plus金牌	3699	https://item.jd.com/10783917878.html机箱	美商海盗船 AIR540 USB3.0 	949	http://item.jd.com/12173900062.html显卡	技嘉(GIGABYTE) GTX1080Ti 11GB 非公版高端游戏显卡深度学习涡轮 * 4 7400 * 4    https://item.jd.com/10583752777.html

这台深度学习主机大概是这样的:

深度学习主机

安装完Ubuntu16.04之后,我又开始了CUDA、cuDnn等深度学习环境和工具的安装之旅,时隔大半年,又有了很多变化,特别是CUDA9.x和cuDnn7.x已经成了标配,这里记录一下。

安装CUDA9.x

注:如果还需要安装Tensorflow1.8,建议这里安装CUDA9.0,我在另一台机器上遇到了一点问题,怀疑和我这台机器先安装CUDA9.0,再安装CUDA9.2有关。

依然从英伟达官方下载当前的CUDA版本,我选择了最新的CUDA9.2:

点选完对应Ubuntu16.04的CUDA9.2 deb版本之后,英伟达官方主页会给出安装提示:

Installation Instructions:
`sudo dpkg -i cuda-repo-ubuntu1604-9-2-local_9.2.88-1_amd64.deb`
`sudo apt-key add /var/cuda-repo-/7fa2af80.pub`
`sudo apt-get update`
`sudo apt-get install cuda`

在下载完大概1.2G的cuda deb版本之后,实际安装命令是这样的:

sudo dpkg -i cuda-repo-ubuntu1604-9-2-local_9.2.88-1_amd64.debsudo apt-key add /var/cuda-repo-9-2-local/7fa2af80.pubsudo apt-get updatesudo apt-get install cuda

官方CUDA下载下载页面还附带了一个cuBLAS 9.2 Patch更新,官方强烈建议安装:

This update includes fix to cublas GEMM APIs on V100 Tensor Core GPUs when used with default algorithm CUBLAS_GEMM_DEFAULT_TENSOR_OP. We strongly recommend installing this update as part of CUDA Toolkit 9.2 installation.

可以用如下方式安装这个Patch更新:

sudo dpkg -i cuda-repo-ubuntu1604-9-2-local-cublas-update-1_1.0-1_amd64.deb sudo apt-get update  sudo apt-get upgrade cuda

CUDA9.2安装完毕之后,1080TI的显卡驱动也附带安装了,可以重启机器,然后用 nvidia-smi 命令查看一下:

最后在在 ~/.bashrc 中设置环境变量:

export PATH=/usr/local/cuda/bin${PATH:+:${PATH}}export LD_LIBRARY_PATH=/usr/local/cuda/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}export CUDA_HOME=/usr/local/cuda

运行 source ~/.bashrc 使其生效。

安装cuDNN7.x

同样去英伟达官网的cuDNN下载页面:https://developer.nvidia.com/rdp/cudnn-download,最新版本是cuDNN7.1.4,有三个版本可以选择,分别面向CUDA8.0, CUDA9.0, CUDA9.2:

cudnn7.1.4 cuda9.2 ubuntu16.04

下载完cuDNN7.1的压缩包之后解压,然后将相关文件拷贝到cuda的系统路径下即可:

tar -zxvf cudnn-9.2-linux-x64-v7.1.tgzsudo cp cuda/include/cudnn.h /usr/local/cuda/include/sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64/ -d sudo chmod a+r /usr/local/cuda/include/cudnn.hsudo chmod a+r /usr/local/cuda/lib64/libcudnn*

安装TensorFlow 1.8

TensorFlow的安装变得越来越简单,现在TensorFlow的官网也有中文安装文档了:https://www.tensorflow.org/install/install_linux?hl=zh-cn , 我们Follow这个文档,用Virtualenv的安装方式进行TensorFlow的安装,不过首先要配置一下基础环境。

首先在Ubuntu16.04里安装 libcupti-dev 库:

这是 NVIDIA CUDA 分析工具接口。此库提供高级分析支持。要安装此库,请针对 CUDA 工具包 8.0 或更高版本发出以下命令:

$ sudo apt-get install cuda-command-line-tools
并将其路径添加到您的 LD_LIBRARY_PATH 环境变量中:

$ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/extras/CUPTI/lib64
对于 CUDA 工具包 7.5 或更低版本,请发出以下命令:

$ sudo apt-get install libcupti-dev

然而我运行“sudo apt-get install cuda-command-line-tools”命令后得到的却是:

E: 无法定位软件包 cuda-command-line-tools

Google后发现其实在安装CUDA9.2的时候,这个包已经安装了,在CUDA的路径下这个库已经有了:

/usr/local/cuda/extras/CUPTI/lib64$ lslibcupti.so  libcupti.so.9.2  libcupti.so.9.2.88

现在只需要将其加入到环境变量中,在~/.bashrc中添加如下声明并令source ~/.bashrc另其生效即可:

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/extras/CUPTI/lib64

剩下的就更简单了:

sudo apt-get install python-pip python-dev python-virtualenv virtualenv --system-site-packages tensorflow1.8source tensorflow1.8/bin/activateeasy_install -U pippip install -i https://pypi.tuna.tsinghua.edu.cn/simple --upgrade tensorflow-gpu

强烈建议将清华的pip源写到配置文件里,这样就更方便快捷了。

最后测试一下TensorFlow1.8:

Python 2.7.12 (default, Dec  4 2017, 14:50:18) 
[GCC 5.4.0 20160609] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
>>> sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
2018-06-17 12:15:34.158680: I tensorflow/core/platform/cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2018-06-17 12:15:34.381812: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1356] Found device 0 with properties: 
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:05:00.0
totalMemory: 10.91GiB freeMemory: 5.53GiB
2018-06-17 12:15:34.551451: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1356] Found device 1 with properties: 
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:06:00.0
totalMemory: 10.92GiB freeMemory: 5.80GiB
2018-06-17 12:15:34.780350: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1356] Found device 2 with properties: 
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:09:00.0
totalMemory: 10.92GiB freeMemory: 5.80GiB
2018-06-17 12:15:34.959199: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1356] Found device 3 with properties: 
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:0a:00.0
totalMemory: 10.92GiB freeMemory: 5.80GiB
2018-06-17 12:15:34.966403: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1435] Adding visible gpu devices: 0, 1, 2, 3
2018-06-17 12:15:36.373745: I tensorflow/core/common_runtime/gpu/gpu_device.cc:923] Device interconnect StreamExecutor with strength 1 edge matrix:
2018-06-17 12:15:36.373785: I tensorflow/core/common_runtime/gpu/gpu_device.cc:929]      0 1 2 3 
2018-06-17 12:15:36.373798: I tensorflow/core/common_runtime/gpu/gpu_device.cc:942] 0:   N Y Y Y 
2018-06-17 12:15:36.373804: I tensorflow/core/common_runtime/gpu/gpu_device.cc:942] 1:   Y N Y Y 
2018-06-17 12:15:36.373808: I tensorflow/core/common_runtime/gpu/gpu_device.cc:942] 2:   Y Y N Y 
2018-06-17 12:15:36.373814: I tensorflow/core/common_runtime/gpu/gpu_device.cc:942] 3:   Y Y Y N 
2018-06-17 12:15:36.374516: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1053] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5307 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:05:00.0, compute capability: 6.1)
2018-06-17 12:15:36.444426: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1053] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 5582 MB memory) -> physical GPU (device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:06:00.0, compute capability: 6.1)
2018-06-17 12:15:36.506340: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1053] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:2 with 5582 MB memory) -> physical GPU (device: 2, name: GeForce GTX 1080 Ti, pci bus id: 0000:09:00.0, compute capability: 6.1)
2018-06-17 12:15:36.614736: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1053] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:3 with 5582 MB memory) -> physical GPU (device: 3, name: GeForce GTX 1080 Ti, pci bus id: 0000:0a:00.0, compute capability: 6.1)
Device mapping:
/job:localhost/replica:0/task:0/device:GPU:0 -> device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:05:00.0, compute capability: 6.1
/job:localhost/replica:0/task:0/device:GPU:1 -> device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:06:00.0, compute capability: 6.1
/job:localhost/replica:0/task:0/device:GPU:2 -> device: 2, name: GeForce GTX 1080 Ti, pci bus id: 0000:09:00.0, compute capability: 6.1
/job:localhost/replica:0/task:0/device:GPU:3 -> device: 3, name: GeForce GTX 1080 Ti, pci bus id: 0000:0a:00.0, compute capability: 6.1
2018-06-17 12:15:36.689345: I tensorflow/core/common_runtime/direct_session.cc:284] Device mapping:
/job:localhost/replica:0/task:0/device:GPU:0 -> device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:05:00.0, compute capability: 6.1
/job:localhost/replica:0/task:0/device:GPU:1 -> device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:06:00.0, compute capability: 6.1
/job:localhost/replica:0/task:0/device:GPU:2 -> device: 2, name: GeForce GTX 1080 Ti, pci bus id: 0000:09:00.0, compute capability: 6.1
/job:localhost/replica:0/task:0/device:GPU:3 -> device: 3, name: GeForce GTX 1080 Ti, pci bus id: 0000:0a:00.0, compute capability: 6.1

安装Keras2.1.x

Keras的后端支持TensorFlow, Theano, CNTK,在安装完TensorFlow GPU版本之后,继续安装Keras非常简单,在TensorFlow的虚拟环境中,直接"pip install keras"即可,安装的版本是Keras2.1.6:

Installing collected packages: h5py, scipy, pyyaml, keras
Successfully installed h5py-2.7.1 keras-2.1.6 pyyaml-3.12 scipy-1.1.0

测试一下:

Python 2.7.12 (default, Dec  4 2017, 14:50:18) 
[GCC 5.4.0 20160609] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import keras
Using TensorFlow backend.

注:原创文章,转载请注明出处及保留链接“我爱自然语言处理”:

本文链接地址:从零开始搭建深度学习服务器: 1080TI四卡并行(Ubuntu16.04+CUDA9.2+cuDNN7.1+TensorFlow+Keras) /?p=10334

AI Challenger 2017 奇遇记

本文记录一下去年下半年参加的AI Challenger比赛的过程,有那么一点意思,之所以说是奇遇,看完文章就明白了。

去年8月,由创新工场、搜狗、今日头条联合举办的“AI challenger全球AI挑战赛”首届比赛正式开赛。比赛共设6个赛道,包括英中机器同声传译、英中机器文本翻译、场景分类、图像中文描述、人体骨骼关键点预测以及虚拟股票趋势预测,一时汇集了众多关注的目光:

“AI Challenger 全球AI挑战赛”是面向全球人工智能(AI)人才的开放数据集和编程竞赛平台,致力于打造大型、全面的科研数据集与世界级竞赛平台,从科研角度出发,满足学术界对高质量数据集的需求,推进人工智能在科研与商业领域的结合,促进世界范围内人工智能研发人员共同探索前沿领域的技术突破及应用创新。在2017年的首届大赛中,AI Challenger发布了千万量级的机器翻译数据集、百万量级的计算机视觉数据集,一系列兼具学术前沿性和产业应用价值的竞赛以及超过200万人民币的奖金,吸引了来自全球65个国家的8892支团队参赛,成为目前国内规模最大的科研数据集平台、以及最大的非商业化竞赛平台。 AI Challenger以服务、培养AI高端人才为使命,打造良性可持续的AI科研新生态。

不过AI Challenger 最吸引我的不是每项比赛数十万元的奖金(这个掂量一下也拿不到),而是英中机器翻译提供的高达1千万的中英双语句对语料,这个量级,在开放的中英语料里仅次于联合国平行语料库,相当的有诱惑力:

简介
英中机器文本翻译作为此次比赛的任务之一,目标是评测各个团队机器翻译的能力。本次机器翻译语言方向为英文到中文。测试文本为口语领域数据。参赛队伍需要根据评测方提供的数据训练机器翻译系统,可以自由的选择机器翻译技术。例如,基于规则的翻译技术、统计机器翻译及神经网络机器翻译等。参赛队伍可以使用系统融合技术,但是系统融合系统不参与排名。需要指出,神经网络机器翻译常见的Ensemble方法,本次评测不认定为系统融合技术。

数据说明
我们将所有数据分割成为训练集、验证集和测试集合。我们提供了超过1000万的英中对照的句子对作为数据集合。其中,训练集合占据绝大部分,验证集合8000对,测试集A 8000条,测试集B 8000条。训练数据主要来源于英语学习网站和电影字幕,领域为口语领域。所有双语句对经过人工检查,数据集从规模、相关度、质量上都有保障。一个英中对照的句子对,包含一句英文和一句中文文本,中文句子由英文句子人工翻译而成。中英文句子分别保存到两个文件中,两个文件中的中英文句子以行号形成一一对应的关系。验证集和测试集最终是以标准的XML格式发布给参赛方。

训练条件
本次评测只允许参赛方使用使用评测方指定的数据训练机器翻译系统,并对其排名。参赛方需遵守以下2018免费送彩金游戏训练方式的说明。参赛方可以使用基本的自然语言处理工具,例如中文分词和命名实体识别。

大概十年前我读研期间做得是统计机器翻译,那个时候能接触到的中英句对最多到过2、3百万,用得最多的工具是知名的开源统计机器翻译工具Moses,也在这里写了不少相关的文章。后来工作先后从事过机器翻译、广告文本挖掘相关的工作,与机器翻译渐行渐远。这一两年,我花了很多时间在专利数据挖掘上,深知专利数据翻译的重要性,也了解到机器翻译对于专利翻译有天然的吸引力。加之这几年来深度学习如火如荼,神经网络机器翻译横空出世,Google, 微软,Facebook等公司2018免费送彩金游戏机器翻译的PR一浪高过一浪,大有“取代”人翻译的感觉,这些都都给了我很大的触动,但是一直没有机会走进神经网络机器翻译。刚好这个时候自己又在家里重新组了一台1080TI深度学习主机,加上AI Challenger提供的机器翻译数据机会,我把这次参赛的目标定为:

  • 了解目前神经网络机器翻译NMT的发展趋势
  • 学习并调研相关的NMT开源工具
  • 将NMT应用在中英日三语之间的专利翻译产品上

相对于统计机器翻译,神经网络机器翻译的开源工具更加丰富,这也和最近几年深度学习开源平台遍地开花有关,每个深度学习平台基本上都附有一两个典型的神经网络机器翻译工具和例子。不过需要说明的是,以下这些2018免费送彩金游戏NMT工具的记录大多数是去年9月到12月期间的调研,很多神经网络机器翻译工具还在不断的迭代和演进中,下面的一些描述可能都有了变化。

虽然之前也或多或少的碰到过一些NMT工具,但是这一次我的神经网络机器翻译开源工具之旅是从OpenNMT开启的,这个开源NMT工具由哈佛NLP组推出,诞生于2016年年末,不过主版本基于Torch, 默认语言是Lua,对于喜爱Python的我来说还不算太方便。所以首先尝试了OpenNMT的Pytorch版本: OpenNMT-py,用AI Challenger官方平台提供中英翻译句对中的500万句对迅速跑了一个OpenNMT-py的默认模型:

Step 2: Train the model
python train.py -data data/demo -save_model demo-model
The main train command is quite simple. Minimally it takes a data file and a save file. This will run the default model, which consists of a 2-layer LSTM with 500 hidden units on both the encoder/decoder.

然后走了一遍AI Challenger的比赛流程,第一次提交记录如下:

2017.09.26 第一次提交:训练数据500万, opennmt-py, default,线下验证集结果:0.2325,线上提交测试集结果:0.22670

走完了比赛流程,接下来我要认真的审视这次英中机器翻译比赛了,在第二轮训练模型开始前,我首先对数据做了标准化的预处理:

  1. 数据shuf之后选择了8000句对作为开发集,8000句对作为测试集,剩下的980多万句对作为训练集;
  2. 英文数据按照统计机器翻译工具Moses 的预处理流程进行了tokenize和truecase;中文数据直接用Jieba中文分词工具进行分词;

这一次我将目光瞄准了Google的NMT系统:GNMT, Google的Research Blog是一个好地方: Building Your Own Neural Machine Translation System in TensorFlow,我从这篇文章入手,然后学习使用Tensorflow的NMT开源工具: Tensorflow-NMT,第一次使用subword bpe处理数据,训练了一个4层的gnmt英中模型,记录如下:

2017.10.05 第二次提交:训练集988万句对, tf-nmt, gnmt-4-layer,bpe16000, 线下验证集结果0.2739,线上提交测试集结果:0.26830

这次的结果不错,BLEU值较第一次提交有4个点的提升,我继续尝试使用bpe处理,一周后,做了第三次提交:

2017.10.12 第三次提交:训练集988万句对,tf-nmt, gnmt-4-layer,bpe32000, 线下验证集结果0.2759,线上提交测试集结果:0.27180

依然有一些提高,不过幅度不大。这一次,为了调研各种NMT开源工具,我又把目光锁定到OpenNMT,事实上,到目前为止,接触到的几个神经网络机器翻译开源工具中,和统计机器翻译开源工具Moses最像的就是OpenNMT,有自己独立的官网,文档相当详细,论坛活跃度很高,并且有不同的分支版本,包括主版本 OpenNMT-lua, Pytorch版本 OpenNMT-py, TensorFlow版本 OpenNMT-tf 。所以为了这次实验我在深度学习主机中安装了Torch和OpenNMT-lua版本,接下来半个月做了两次OpenNMT训练英中神经网络翻译模型的尝试,不过在验证集的结果和上面的差不多或者略低,没有实质性提高,所以我放弃了这两次提交。

也在这个阶段,从不同途径了解到Google新推的Transformer模型很牛,依然从Google Research Blog入手:Transformer: A Novel Neural Network Architecture for Language Understanding ,学习这篇神文:《Attention Is All You Need》 和尝试相关的Transformer开源工具 TensorFlow-Tensor2Tensor。一图胜千言,谷歌AI博客上给得这个图片让人无比期待,不过实际操作中还是踩了很多坑:

还是和之前学习使用开源工具的方法类似,我第一次的目标主要是走通tensor2tensor,所以跑了一个 wmt32k base_single 的英中transformer模型,不过结果一般,记录如下:

2017.11.03 第六次实验:t2t transformer wmt32k base_single, 线下验证集BLEU: 0.2605,未提交

之后我又换为wmt32k big_single的设置,再次训练英中transformer模型,这一次,终于在线下验证集的BLEU值上,达到了之前GNMT最好的结果,所以我做了第四次线上提交,不过测试集A的结果还略低一些,记录如下:

2017.11.06 第七次实验:t2t transformer wmt32k big_single,线下验证集结果 0.2759, 线上测试集得分:0.26950

不过这些结果和博客以及论文里宣称的结果相差很大,我开始去检查差异点,包括tensor2tensor的issue以及论文,其实论文里2018免费送彩金游戏实验的部分交代的很清楚:

On the WMT 2014 English-to-German translation task, the big transformer model (Transformer (big) in Table 2) outperforms the best previously reported models (including ensembles) by more than 2.0 BLEU, establishing a new state-of-the-art BLEU score of 28.4. The configuration of this model is listed in the bottom line of Table 3. Training took 3.5 days on 8 P100 GPUs. Even our base model surpasses all previously published models and ensembles, at a fraction of the training cost of any of the competitive models.

On the WMT 2014 English-to-French translation task, our big model achieves a BLEU score of 41.0, outperforming all of the previously published single models, at less than 1/4 the training cost of the previous state-of-the-art model. The Transformer (big) model trained for English-to-French used dropout rate Pdrop = 0.1, instead of 0.3.

For the base models, we used a single model obtained by averaging the last 5 checkpoints, which were written at 10-minute intervals. For the big models, we averaged the last 20 checkpoints. We used beam search with a beam size of 4 and length penalty α = 0.6 . These hyperparameters were chosen after experimentation on the development set. We set the maximum output length during inference to input length + 50, but terminate early when possible.

总结起来有2个地方可以改进:第一,是对checkpoints进行average, 这个效果立竿见影:

2017.11.07 第八次实验:t2t transformer wmt32k big_single average model, 线下验证集得分 0.2810 , 提交测试集得分:0.27330

第二,要有高性能的深度学习服务器。谷歌实验中最好的结果是在8块 P100 GPU的机器上训练了3.5天,对我的单机1080TI深度学习主机来说,一方面训练时对参数做了取舍,另一方面用时间换空间,尝试增加训练步数,直接将训练步数增加到100万次,结果还是不错的:

2017.11.15 第九次实验:t2t transformer wmt32k big_single 1000k 10beam,线下验证集得分0.2911,线上提交测试集得分0.28560

然后继续average checkpoints:
2017.11.16 第十次提交: t2t transformer wmt32k big_single 1000k average 10beam, 线下验证集得分0.2930,线上提交测试集得分0.28780

这两个方法确实能有效提高BLEU值,所以我继续沿用这个策略,按着训练时间推算了一下,估计这台机器在12月初比赛正式结束前大概可以训练一个250万次的模型,当然,这个给自己预留了最后提交比赛结果的时间。不过在11月27日,我在英中机器翻译比赛测试集A结束提交前提交了一个训练了140万次,并做了模型average的提交,算是这个赛道Test A关闭前的最后一次提交:

2017.11.27 第十一次提交 t2t transformer wmt32k big_single 1400k.beam10.a0.9.average, 验证集 0.2938 测试集 0.28950

12月1日凌晨测试集B正式放出,这个是最终排名的重要依据,只有2次提交机会,并且结果不会实时更新,只有等到12月3号之后才会放出最终排名。我的英中2500k Transformer模型大概在12月2号训练完毕,我做了Test B的第一次提交:

2017.12.2 average b10 a0.9: 0.2972(验证集)

之后,我逐一检查了保留的20个checkpoint在验证集上的得分,最终选择了高于平均值的11个checkpoint的average又做了第二次提交,虽然验证集只高了0.0001, 但是在这样的比赛中,“蚊子肉也是肉啊”:

2017.12.3 average select 11 b10 a0.9: 0.2973(验证集)

这就是我在英中机器文本翻译比赛中的整个历程,在Test A的最终排名大概在二十几名,但是最后一次模型的结果应该还能提高,所以预期是前20,剩下的就是等待TEST B的最终排名结果了。做到这个份上,其实我还挺满意的,不过故事如果真的到此就结束了,那算不上奇遇,有意思的事情才刚开始。

AI Challenger 2017有两个赛道和机器翻译有关,一个是英中机器文本翻译比赛(最高奖金30万),另外一个是英中机器同声传译比赛(最高奖金40万),一开始报名的时候,直观上觉得后者比较复杂,一方面奖金部分说明了问题,另外赛题描述部分也让人觉得涉及到语音处理,比较复杂:

简介
随着最近深度学习在语音、自然语言处理里面的应用,语音识别的错误率在不断降低,机器翻译的效果也在不断提高。语音处理和机器翻译的进步也推动机器同声传译的进步。如果竞赛任务同时考虑语音识别、机器翻译和语音合成这些项目,参赛队伍遇到的难度会很大。所以本次评测重点也在语音识别后的文本处理和机器翻译任务。翻译语言方向为英文到中文。

语音识别后处理模块:语音识别后的文本与书面语有很多不同。识别后文本具有(1)包含有识别错误;(2)识别结果没有标点符号;(3)源端为比较长的句子,例如对40~50s的语音标注后的文本,没有断句;(4)口语化文本,夹杂语气词等特点。由于本次比赛没有提供错误和正确对照的文本用于训练纠错模块。本次比赛提供的测试集合的源端文本是人工对语音标注后的文本,不包含识别错误。针对其它的特点,参赛队伍可以这几个方面考虑优化,但不限于以下几个方面:

1. 针对无标点的情况,参赛方可以利用提供的英文单语数据训练自动标点模块。用自动标点模块对测试集合文本进行添加标点。自动标点也属于序列标注任务,选手可以使用统计模型或是神经网络的模型进行建模。

2. 针对断句:源端文本都是比较长的文本,不利于机器翻译,参赛者可以设定断句策略。例如,参赛者可以依据标点来进行断句,将每个小的分句送入机器翻译系统。

3. 针对口语化:参赛队伍可以制定一些去除口语词的规则来处理测试集合。

机器翻译模块:将识别后处理的文本翻译成目标语言。参赛队伍需要根据评测方提供的数据训练机器翻译系统,可以自由的选择机器翻译技术。例如,基于规则的翻译技术、基于实例的翻译技术、统计机器翻译及神经网络机器翻译等。参赛队伍可以使用系统融合技术,但是系统融合系统不参与排名。

数据说明
机器翻译训练集。我们提供了1000万左右英中对照的句子对作为训练集合。训练数据领域为口语领域。所有双语句对经过人工检查,数据集从规模、相关度、质量上都有保障。一个英中对照的句子对,包含一句英文和一句中文文本,中文句子由英文句子人工翻译而成。

自动标点训练数据。选手可以利用提供的1000万文本训练自动标点系统。

验证集和测试集。我们会分别选取多个英语演讲的题材的音频,总时长在3~6小时之间,然后按照内容切分成30s~50s不等长度的音频数据,人工标注出音频对应的英文文本。人工标注的文本不翻译识别错误、无标点、含有语气词等。人工标注的好的英文文本会由专业译员翻译成中文文本,就形成了英中对照的句子对。抽取的英中对照的句子对会被分割为验证集和测试集。验证集和测试集最终是以标准的XML格式提供给选手。

我在一开始的时候考虑到这个比赛同样提供上千万句对的语料,所以当时顺手报名了这个同声传译比赛,但是直到最后一刻,我还没有仔细看过或者准备过这个任务。不过12月2号当我第一次完成英中机器翻译比赛的测试集B提交后,以完成作业的心态了解了一下这个英中机器同传比赛的题意以及数据集,发现这里提供的训练集和英中机器翻译比赛的数据是一致的,也就是说机器翻译模块可以复用之前训练的英中Transformer模型,而真正需要解决的,是标点符号自动标注模块以及断句模块。

感谢Google、Github和开源世界,在测试了几个自动标点标注模块后,我把目光锁定在 punctuator2(A bidirectional recurrent neural network model with attention mechanism for restoring missing punctuation in unsegmented text), 一个带attention机制的双向RNN无标点文本标点符号还原工具,通过它很快的构建了英文文本自动标点标注模块,并且用在了英中机器同声传译比赛的验证集和测试集上,验证集结果不算太差,所以对应英中机器翻译的模型,我也做了两次测试集B的提交,但是至于结果如何,我根本无法判断,因为在测试集A上,我没有提交过一次,所以无法判断测试集和验证集的正相关性。但是完成了 AI Challenger 的相关“作业“,我基本上心满意足了,至于结果如何,Who Care?

大约一个周之后测试集B上的结果揭晓,我在英中机器翻译文本比赛上进了前20,英中同声传译比赛上进了前10,不过前者的参数队伍有150多支,后者不足30支,特别是测试集B的提交队伍不到15支,有点诡异。原本以为这就结束了,不过到了12月中旬的某个周末,我微信突然收到了AI Challenger小助手的催收信息,大意是需要提交什么代码验证,问我为什么一直没有提交?我一脸错愕,她让我赶紧查看邮件,原来早在一个周之前的12月9号,AI Challenger发了一封邮件,主题是这样的:“AI Challenger 2017 TOP10 选手通知”

亲爱的AI Challenger,

恭喜你,过五关斩六将进入了TOP10,进入前十的机率是0.56%,每一位都是千里挑一的人才。非常不容易也非常优秀!

为了保证竞赛公平公正性,您还需要在12月10日中午12点前按如下格式提交您的代码至大赛核验邮箱aichallenger@chuangxin.com

邮件格式:
主题:AI ChallengerTOP10代码提交-队伍名称-赛道
正文:
队伍名称
全体队员信息:姓名-AI Challenger昵称-电话-邮箱-所在机构-专业&年级

附件:(文件名称)
1- 代码

非常感谢您的合作。

原来测试集B上的前10名同学需要提交代码复核,我原来以为只有前5名需要去北京现场答辩的同学要做这个,没想到前10名都需要做,赶紧和AI Challenger小助手沟通了一下,因为自己几乎都是通过开源工具完成的比赛,就简单的提交了一份说明文档过去了。正是在参加AI Challenger比赛的同一时期,我们的专利机器翻译产品也马不停蹄的开展了,出于对两个赛道前几名队伍BLEU值的仰望,我准备去北京旁听一下现场答辩,所以当天还和AI Challenger小助手沟通了一下现场观摩的问题,小助手说,前十名可以直接来,所以我觉得进入前十名还是不错的。

没想到第二天一早又收到Challenger小助手的微信留言,大意是:你不用自己买票来观摩比赛了,因为前面有几支队伍因种种原因放弃现场答辩,你自动递补为第5名,需要来北京参加12月21日的现场决赛答辩和颁奖礼,我们给你买机票和定酒店。吃不吃惊?意不意外?我当时的第一反应这真是2017年本人遇到最奇特的一件事情。。。然后很快收到了一封决赛邀请函:

亲爱的AI Challenger,

恭喜你,过五关斩六将走到了决赛,进入决赛的机率是0.28%,每一位都是千里挑一的人才。非常不容易也非常优秀!

“AI Challenger 全球AI挑战赛”面向人工智能领域科研人才,致力于打造大型、全面的科研数据集与世界级竞赛平台。由创新工场、搜狗、今日头条联合创建,旨在从科研角度出发,满足学术界对高质量数据集的需求,推进人工智能在科研与商业领域的结合,促进世界范围内人工智能研发人员共同探索前沿领域的技术突破及应用创新。

2017年是AI Challenger的诞生年,我们公布了百万量级的计算机视觉数据集、千万量级的机器翻译数据集,并主办多条细分赛道的AI竞赛。本次英中机器同传竞赛主要任务为集中优化语音识别后处理和机器翻译模块,旨在解决机器同声传译中的技术问题。

......

恭喜所有的入围选手!所有的入围者将在12月21日到中国北京进行现场答辩,本次大赛将以最终榜单排名结合答辩表现,加权计算总成绩,决出最终的大奖。

在答辩之前,我们需要Top5团队于12月18日下午17点前提交包括:
1-答辩PPT、
2-队员情况(个人姓名、个人高清半身照片、个人学校-年级-专业/公司-部门-职务、是否有指导老师-如有,请附上老师150字内简介)
3-团队出席名单(涉及报销事宜)
4-代码(供审查,如有作弊情况将按大赛规则处理)
5-150字内个人简介-选手手册素材(建议为三段话,第一段话是背景介绍,包括你的学校、实验室、师从老师等信息;第二段话可以介绍你的技术优势,包括Paper、竞赛履历、实习履历、项目经历;第三段话支持自由发挥,个人主页、你的爱好,让我们发现一个独一无二的你)
......

虽然去北京参加现场决赛也只是陪太子读书,不过最终还是决定去参加现场答辩,当然这里还有一关需要验证,前10名只需要提交代码或者代码描述即可,前5名参加决赛的同学还要复现整个流程,我很快被小助手拉入一个小群,里面有来自搜狗的工程师同学,他们给我提供了一台深度学习机器,让我复现整个过程以及最终核验比赛结果。当然,留给我的时间比较紧张,12月21号要去北京参加现场答辩,当时已经是12月18号了,所以Challenger小助手特地给我将时间留到了最后一刻。准备PPT和复现整个流程同时进行(复现并不是等于重新训练一遍,譬如机器翻译模型可以直接上传之前训练好的),终于赶在最后时刻完工。不过我自己答辩现场的感觉匆匆忙忙,效果也一般,但是学习了一圈其他获奖队伍的思路,很有收获:Transformer是主流获奖模型,但是很多功夫在细节,包括数据预处理阶段的筛选,数据 & 模型后处理的比拼,当然,牛逼的深度学习机器也是不可或缺的。

附上当时现场答辩PPT上写得几点思考,抛砖引玉,欢迎大家一起探讨机器翻译特别是神经网络机器翻译的现状和未来:

  • NMT开源工具的生态问题,这个过程中我们尝试了OpenNMT, OpenNMT-py, OpenNMT-tf, Tensorflow-nmt, Tensor2Tensor等工具, 总体感觉OpenNMT的生态最完备,很像SMT时代的Moses
  • NMT的工程化和产品化问题,从学术产品到工程产品,还有很多细节要打磨
  • 面向垂直领域的机器翻译:专利机器翻译是一个多领域的机器翻译问题
  • 由衷感谢这些从idea到开源工具都无私奉献的研究者和从业者们,我们只是站在了你们的肩膀上

当然,参加完AI Challenger比赛之后我们并没有停止对于神经网络机器翻译应用的探索,也有了一些新的体会。这半年来我们一直在打磨AIpatent机器翻译引擎,目标是面向中英专利翻译、中日专利翻译、日英专利翻译提供专业的专利翻译引擎,欢迎有这方面需求的同学试用我们的引擎,目前还在不断迭代中。

注:原创文章,转载请注明出处及保留链接“我爱自然语言处理”:

本文链接地址:AI Challenger 2017 奇遇记 /?p=10218

Coursera上数学类相关课程(公开课)汇总推荐

数学课程是基础,Coursera上有很多数学公开课,这里做个汇总,注意由于Coursera上有一批很有特色的统计学相关的数学课程,我们将在下一期里单独汇总。

1 斯坦福大学 Introduction to Mathematical Thinking(数学思维导论)

http://coursegraph.com/coursera-mathematical-thinking

引用老版课程一个同学的评价,供参考:

这门课是高中数学到大学数学的一个过度。高中数学一般重计算不太注重证明,这门课讲了基本的逻辑,数学语言(两个 quantifier,there exists, for all)和证明的几个基本方法,比如证明充要条件要从两个方向证、证伪只需要举个反例,原命题不好证的时候可以证等价的逆否命题以及很常用的数学归纳法。课程讲了数论里一些基本定理,然后通过让你证一些看起来显然而不需要证明的证明题来训练你证明的技能和逻辑思考的能力,看起来显然的命题也是要证明才能说服人的,课程最后简略的讲了下数学分析里面实数的引入,但这部分讲的不完整。Keith Devlin 是个 old school 的讲师,上课只用纸和笔,也是属于比较热情的讲师,他每周都会录几个答疑的视频。这门比较适合大一的新生上,开得也比较频繁。

课程简介:

Learn how to think the way mathematicians do – a powerful cognitive process developed over thousands of years. Mathematical thinking is not the same as doing mathematics – at least not as mathematics is typically presented in our school system. School math typically focuses on learning procedures to solve highly stereotyped problems. Professional mathematicians think a certain way to solve real problems, problems that can arise from the everyday world, or from science, or from within mathematics itself. The key to success in school math is to learn to think inside-the-box. In contrast, a key feature of mathematical thinking is thinking outside-the-box – a valuable ability in today’s world. This course helps to develop that crucial way of thinking.

2 加州大学尔湾分校 初级微积分系列课程

1)Pre-Calculus: Functions(初级微积分:函数)

http://coursegraph.com/coursera-pre-calculus

This course covers mathematical topics in college algebra, with an emphasis on functions. The course is designed to help prepare students to enroll for a first semester course in single variable calculus. Upon completing this course, you will be able to: 1. Solve linear and quadratic equations 2. Solve some classes of rational and radical equations 3. Graph polynomial, rational, piece-wise, exponential and logarithmic functions 4. Find integer roots of polynomial equations 5. Solve exponential and logarithm equations 6. Understand the inverse relations between exponential and logarithm equations 7. Compute values of exponential and logarithm expressions using basic properties

2)Pre-Calculus: Trigonometry(初级微积分:三角)

http://coursegraph.com/coursera-trigonometry

This course covers mathematical topics in trigonometry. Trigonometry is the study of triangle angles and lengths, but trigonometric functions have far reaching applications beyond simple studies of triangles. This course is designed to help prepare students to enroll for a first semester course in single variable calculus. Upon completing this course, you will be able to: 1. Evaluate trigonometric functions using the unit circle and right triangle approaches 2. Solve trigonometric equations 3. Verify trigonometric identities 4. Prove and use basic trigonometric identities. 5. Manipulate trigonometric expressions using standard identities 6. Solve right triangles 7. Apply the Law of Sines and the Law of Cosines

3 宾夕法尼亚大学的 单变量微积分系列课程

1)Calculus: Single Variable Part 1 - Functions(单变量微积分1:函数)
http://coursegraph.com/coursera-single-variable-calculus

Calculus is one of the grandest achievements of human thought, explaining everything from planetary orbits to the optimal size of a city to the periodicity of a heartbeat. This brisk course covers the core ideas of single-variable Calculus with emphases on conceptual understanding and applications. The course is ideal for students beginning in the engineering, physical, and social sciences. Distinguishing features of the course include: 1) the introduction and use of Taylor series and approximations from the beginning; 2) a novel synthesis of discrete and continuous forms of Calculus; 3) an emphasis on the conceptual over the computational; and 4) a clear, dynamic, unified approach. In this first part--part one of five--you will extend your understanding of Taylor series, review limits, learn the *why* behind l'Hopital's rule, and, most importantly, learn a new language for describing growth and decay of functions: the BIG O.

2)Calculus: Single Variable Part 2 - Differentiation(单变量微积分2:微分)

http://coursegraph.com/coursera-differentiation-calculus

Calculus is one of the grandest achievements of human thought, explaining everything from planetary orbits to the optimal size of a city to the periodicity of a heartbeat. This brisk course covers the core ideas of single-variable Calculus with emphases on conceptual understanding and applications. The course is ideal for students beginning in the engineering, physical, and social sciences. Distinguishing features of the course include: 1) the introduction and use of Taylor series and approximations from the beginning; 2) a novel synthesis of discrete and continuous forms of Calculus; 3) an emphasis on the conceptual over the computational; and 4) a clear, dynamic, unified approach. In this second part--part two of five--we cover derivatives, differentiation rules, linearization, higher derivatives, optimization, differentials, and differentiation operators.

3)Calculus: Single Variable Part 3 - Integration(单变量微积分3:积分)

http://coursegraph.com/coursera-integration-calculus

Calculus is one of the grandest achievements of human thought, explaining everything from planetary orbits to the optimal size of a city to the periodicity of a heartbeat. This brisk course covers the core ideas of single-variable Calculus with emphases on conceptual understanding and applications. The course is ideal for students beginning in the engineering, physical, and social sciences. Distinguishing features of the course include: 1) the introduction and use of Taylor series and approximations from the beginning; 2) a novel synthesis of discrete and continuous forms of Calculus; 3) an emphasis on the conceptual over the computational; and 4) a clear, dynamic, unified approach. In this third part--part three of five--we cover integrating differential equations, techniques of integration, the fundamental theorem of integral calculus, and difficult integrals.

4) Calculus: Single Variable Part 4 - Applications(单变量微积分4:应用)

http://coursegraph.com/coursera-applications-calculus

Calculus is one of the grandest achievements of human thought, explaining everything from planetary orbits to the optimal size of a city to the periodicity of a heartbeat. This brisk course covers the core ideas of single-variable Calculus with emphases on conceptual understanding and applications. The course is ideal for students beginning in the engineering, physical, and social sciences. Distinguishing features of the course include: 1) the introduction and use of Taylor series and approximations from the beginning; 2) a novel synthesis of discrete and continuous forms of Calculus; 3) an emphasis on the conceptual over the computational; and 4) a clear, dynamic, unified approach. In this fourth part--part four of five--we cover computing areas and volumes, other geometric applications, physical applications, and averages and mass. We also introduce probability.

4 杜克大学 Data Science Math Skills(数据科学中的数学技巧)

http://coursegraph.com/coursera-datasciencemathskills

这门课程主要介绍数据科学中涉及的相关数学概念,让学生了解基本的数学概念,掌握基本的数学语言,内容涵盖集合论、求和的Sigma符号、数学上的笛卡尔(x,y)平面、指数、对数和自然对数函数,概率论以及叶斯定理等:

Data science courses contain math—no avoiding that! This course is designed to teach learners the basic math you will need in order to be successful in almost any data science math course and was created for learners who have basic math skills but may not have taken algebra or pre-calculus. Data Science Math Skills introduces the core math that data science is built upon, with no extra complexity, introducing unfamiliar ideas and math symbols one-at-a-time. Learners who complete this course will master the vocabulary, notation, concepts, and algebra rules that all data scientists must know before moving on to more advanced material.

5 加州大学圣迭戈分校 Introduction to Discrete Mathematics for Computer Science Specialization(面向计算机科学的离散数学专项课程)

http://coursegraph.com/coursera-specializations-discrete-mathematics

面向计算机科学的离散数学专项课程(Introduction to Discrete Mathematics for Computer Science Specialization),这个系列包含5门子课程,涵盖证明、组合数学与概率、图论,数论和密码学,配送问题项目等,感兴趣的同学可以关注: Build a Foundation for Your Career in IT-Master the math powering our lives and prepare for your software engineer or security analyst career

Discrete Math is needed to see mathematical structures in the object you work with, and understand their properties. This ability is important for software engineers, data scientists, security and financial analysts (it is not a coincidence that math puzzles are often used for interviews). We cover the basic notions and results (combinatorics, graphs, probability, number theory) that are universally needed. To deliver techniques and ideas in discrete mathematics to the learner we extensively use interactive puzzles specially created for this specialization. To bring the learners experience closer to IT-applications we incorporate programming examples, problems and projects in our courses.

1) What is a Proof(什么是证明)

http://coursegraph.com/coursera-what-is-a-proof

There is a perceived barrier to mathematics: proofs. In this course we will try to convince you that this barrier is more frightening than prohibitive: most proofs are easy to understand if explained correctly, and often they are even fun. We provide an accompanied excursion in the “proof zoo” showing you examples of techniques of different kind applied to different topics. We use some puzzles as examples, not because they are “practical”, but because discussing them we learn important reasoning and problem solving techniques that are useful. We hope you enjoy playing with the puzzles and inventing/understandings the proofs. As prerequisites we assume only basic math (e.g., we expect you to know what is a square or how to add fractions), basic programming in python (functions, loops, recursion), common sense and curiosity. Our intended audience are all people that work or plan to work in IT, starting from motivated high school students.

2)Combinatorics and Probability(组合和概率)

Counting is one of the basic mathematically related tasks we encounter on a day to day basis. The main question here is the following. If we need to count something, can we do anything better than just counting all objects one by one? Do we need to create a list of all phone numbers to ensure that there are enough phone numbers for everyone? Is there a way to tell that our algorithm will run in a reasonable time before implementing and actually running it? All these questions are addressed by a mathematical field called Combinatorics. In this course we discuss most standard combinatorial settings that can help to answer questions of this type. We will especially concentrate on developing the ability to distinguish these settings in real life and algorithmic problems. This will help the learner to actually implement new knowledge. Apart from that we will discuss recursive technique for counting that is important for algorithmic implementations. One of the main `consumers’ of Combinatorics is Probability Theory. This area is connected with numerous sides of life, on one hand being an important concept in everyday life and on the other hand being an indispensable tool in such modern and important fields as Statistics and Machine Learning. In this course we will concentrate on providing the working knowledge of basics of probability and a good intuition in this area. The practice shows that such an intuition is not easy to develop. In the end of the course we will create a program that successfully plays a tricky and very counterintuitive dice game. As prerequisites we assume only basic math (e.g., we expect you to know what is a square or how to add fractions), basic programming in python (functions, loops, recursion), common sense and curiosity. Our intended audience are all people that work or plan to work in IT, starting from motivated high school students.

3)Introduction to Graph Theory(图论导论)

We invite you to a fascinating journey into Graph Theory — an area which connects the elegance of painting and the rigor of mathematics; is simple, but not unsophisticated. Graph Theory gives us, both an easy way to pictorially represent many major mathematical results, and insights into the deep theories behind them. In this course, among other intriguing applications, we will see how GPS systems find shortest routes, how engineers design integrated circuits, how biologists assemble genomes, why a political map can always be colored using a few colors. We will study Ramsey Theory which proves that in a large system, complete disorder is impossible! By the end of the course, we will implement an algorithm which finds an optimal assignment of students to schools. This algorithm, developed by David Gale and Lloyd S. Shapley, was later recognized by the conferral of Nobel Prize in Economics. As prerequisites we assume only basic math (e.g., we expect you to know what is a square or how to add fractions), basic programming in python (functions, loops, recursion), common sense and curiosity. Our intended audience are all people that work or plan to work in IT, starting from motivated high school students.

4) Number Theory and Cryptography(数论和密码学)

We all learn numbers from the childhood. Some of us like to count, others hate it, but any person uses numbers everyday to buy things, pay for services, estimated time and necessary resources. People have been wondering about numbers’ properties for thousands of years. And for thousands of years it was more or less just a game that was only interesting for pure mathematicians. Famous 20th century mathematician G.H. Hardy once said “The Theory of Numbers has always been regarded as one of the most obviously useless branches of Pure Mathematics”. Just 30 years after his death, an algorithm for encryption of secret messages was developed using achievements of number theory. It was called RSA after the names of its authors, and its implementation is probably the most frequently used computer program in the word nowadays. Without it, nobody would be able to make secure payments over the internet, or even log in securely to e-mail and other personal services. In this short course, we will make the whole journey from the foundation to RSA in 4 weeks. By the end, you will be able to apply the basics of the number theory to encrypt and decrypt messages, and to break the code if one applies RSA carelessly. You will even pass a cryptographic quest! As prerequisites we assume only basic math (e.g., we expect you to know what is a square or how to add fractions), basic programming in python (functions, loops, recursion), common sense and curiosity. Our intended audience are all people that work or plan to work in IT, starting from motivated high school students.

5)Solving Delivery Problem(解决旅行商问题)

http://coursegraph.com/coursera-delivery-problem

We’ll implement together an efficient program for a problem needed by delivery companies all over the world millions times per day — the travelling salesman problem. The goal in this problem is to visit all the given places as quickly as possible. How to find an optimal solution to this problem quickly? We still don’t have provably efficient algorithms for this difficult computational problem and this is the essence of the P versus NP problem, the most important open question in Computer Science. Still, we’ll implement several efficient solutions for real world instances of the travelling salesman problem. While designing these solutions, we will rely heavily on the material learned in the courses of the specialization: proof techniques, combinatorics, probability, graph theory. We’ll see several examples of using discrete mathematics ideas to get more and more efficient solutions.

6 伦敦帝国理工学院 Mathematics for Machine Learning Specialization(面向机器学习的数学专项课程系列)

http://coursegraph.com/coursera-specializations-mathematics-machine-learning

伦敦帝国理工学院的面向机器学习的数学专项课程系列(Mathematics for Machine Learning Specialization),该系列包含3门子课程,涵盖线性代数,多变量微积分,以及主成分分析(PCA),这个专项系列课程的目标是弥补数学与机器学习以及数据科学鸿沟,感兴趣的同学可以关注:Mathematics for Machine Learning。Learn about the prerequisite mathematics for applications in data science and machine learning

For a lot of higher level courses in Machine Learning and Data Science, you find you need to freshen up on the basics in maths - stuff you may have studied before in school or university, but which was taught in another context, or not very intuitively, such that you struggle to relate it to how it’s used in Computer Science. This specialisation aims to bridge that gap, getting you up to speed in the underlying maths, building an intuitive understanding, and relating it to Machine Learning and Data Science. In the first course on Linear Algebra we look at what linear algebra is and how it relates to data. Then we look through what vectors and matrices are and how to work with them. The second course, Multivariate Calculus, builds on this to look at how to optimise fitting functions to get good fits to data. It starts from introductory calculus and then uses the matrices and vectors from the first course to look at data fitting. The third course, Dimensionality Reduction with Principal Components Analysis, uses the maths from the first two courses to do simple optimisation for the situation where you don’t have an understanding of how the data variables relate to each other. At the end of this specialisation you will have gained the prerequisite mathematical knowledge to continue your journey and take more advanced courses in machine learning.

1) Mathematics for Machine Learning: Linear Algebra(面向机器学习的数学:线性代数)

http://coursegraph.com/coursera-linear-algebra-machine-learning

In this course on Linear Algebra we look at what linear algebra is and how it relates to vectors and matrices. Then we look through what vectors and matrices are and how to work with them, including the knotty problem of eigenvalues and eigenvectors, and how to use these to solve problems. Finally we look at how to use these to do fun things with datasets - like how to rotate images of faces and how to extract eigenvectors to look at how the Pagerank algorithm works. Since we're aiming at data-driven applications, we'll be implementing some of these ideas in code, not just on pencil and paper. Towards the end of the course, you'll write code blocks and encounter Jupyter notebooks in Python, but don't worry, these will be quite short, focussed on the concepts, and will guide you through if you’ve not coded before. At the end of this course you will have an intuitive understanding of vectors and matrices that will help you bridge the gap into linear algebra problems, and how to apply these concepts to machine learning.

2)Mathematics for Machine Learning: Multivariate Calculus(面向机器学习的数学:多变量微积分)

http://coursegraph.com/coursera-multivariate-calculus-machine-learning

This course offers a brief introduction to the multivariate calculus required to build many common machine learning techniques. We start at the very beginning with a refresher on the “rise over run” formulation of a slope, before converting this to the formal definition of the gradient of a function. We then start to build up a set of tools for making calculus easier and faster. Next, we learn how to calculate vectors that point up hill on multidimensional surfaces and even put this into action using an interactive game. We take a look at how we can use calculus to build approximations to functions, as well as helping us to quantify how accurate we should expect those approximations to be. We also spend some time talking about where calculus comes up in the training of neural networks, before finally showing you how it is applied in linear regression models. This course is intended to offer an intuitive understanding of calculus, as well as the language necessary to look concepts up yourselves when you get stuck. Hopefully, without going into too much detail, you’ll still come away with the confidence to dive into some more focused machine learning courses in future.

3)Mathematics for Machine Learning: PCA(面向机器学习的数学:主成分分析)

http://coursegraph.com/coursera-pca-machine-learning

This course introduces the mathematical foundations to derive Principal Component Analysis (PCA), a fundamental dimensionality reduction technique. We'll cover some basic statistics of data sets, such as mean values and variances, we'll compute distances and angles between vectors using inner products and derive orthogonal projections of data onto lower-dimensional subspaces. Using all these tools, we'll then derive PCA as a method that minimizes the average squared reconstruction error between data points and their reconstruction. At the end of this course, you'll be familiar with important mathematical concepts and you can implement PCA all by yourself. If you’re struggling, you’ll find a set of jupyter notebooks that will allow you to explore properties of the techniques and walk you through what you need to do to get on track. If you are already an expert, this course may refresh some of your knowledge.

注:本文首发“课程图谱博客”:http://blog.coursegraph.com

同步发布到这里, 本文链接地址:http://blog.coursegraph.com/coursera上数学类相关课程数学公开课汇总推荐 http://blog.coursegraph.com/?p=804