• What is machine learning？ Machine learning is to find functions automatically.For example, speech recognition, image recognition, playing go, dialogue system… 1.regression prediction 2.binary ...
Machine learning is to find functions automatically.For example, speech recognition, image recognition, playing go, dialogue system…
Regression
Prediction.
Binary Classification
Output is yes or no.
Multiclass Classification
Let the machine do multiple choice questions.The machine selects the correct one in the options.
Generation
Produce structured complex things.Like images,sentences created like a human.How to tell the machine what kind of function to look for？
Supervised Learning
Labeled data.Need to give the machine some labeled data for training.
Function Loss
Judge whether the function is good or bad. The basic idea is as follows: First, the image is given to the function for output, and the output result is compared with the label. The proportion of the correct answer is related to the quality of the function.The smaller the LOSS, the closer to the function we need to find. Next, the machine will automatically select the function with the lowest LOSS value.
Reinforcement Learning
The result will allow the machine to evaluate previous decisions.And alpha go both use supervised learning and reinforcement learning.
Unsupervised Learning
Unlabeled data.What can machine learn from unlabeled data.
Function Search Range
Linear,Network Architecture,like RNN and CNN.
Function Search Method
Other Application
Explainable AI
Tell us why this result is output.
Deliberately modify the input data to confuse the machine, resulting in an error in its selection function.
Network Compression
Compress the data input and put it on the phone.
Anomaly Detection
Enter something completely different,Can you let the machine know that ‘i don’t know’?
Transfer Learning
Training data AND testing data.learn by analogy.BUT when testing data totally different with the training data, what tragety will machine choose.
Meta Learning
Let machine can learn by itself! learning algorithmn.Can you make the machine smarter？create learning algorithmn.
Life-long Learning
Machine life-long learning.Continuous learning.


展开全文
• 02What is Machine Learning, Deep Learning and Structured Learning
• What is Machine Learning Two broad classification Supervised Learning Regression Classification Unsupervised Learning What is Machine Learning?One definition of Machine Learning: “A computer program ...

What is Machine Learning
Classification

Unsupervised Learning

What is Machine Learning?

One definition of Machine Learning: “A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E.”

Supervised Learning

Regression

In a regression problem, we are trying to predict results within a continuous output, meaning that we are trying to map input variables to some continuous function.

Classification

In a classification problem, we are instead trying to predict results in a discrete output. In other words, we are trying to map input variables into discrete categories.

Unsupervised Learning

Unsupervised learning allows us to approach problems with little or no idea what our results should look like. We can derive structure from data where we don’t necessarily know the effect of the variables.

With unsupervised learning there is no feedback based on the prediction results.
展开全文
• What is machine learning? machine learning : experience—-&gt;skill The skill is something which can improve the performance of program Why use machine learning For example,how to recognize ...
What is machine learning?

machine learning : experience—->skill

The skill is something which can improve the performance of program

Why use machine learning

For example,how to recognize a tree？

One way is that,we may write down many rules of the tree(the rules can be so many),once we need to judge a object whether tree or not we can use these rules to determine.But in fact ,we didn’t recongize by this way.Just think how a three-year-old baby do this.They can recongnize the tree after seen a lot of tree,they don’t need to write down rules.Because they have seen too many tree so when they meet a new thing they can determine whether it is a really tree.

PC can recognize a tree like a three-years-old baby,they just need some data to deal with just like to see many trees like for training.And it would be more effective than writing rules.

What fileds can machine learning apply to?

For example,some interesting field

Navigating on Mars: we can  not make rules in advance for the reason that we haven’t reach Mars before,we need machine to collect data and use these data to maker decision when it arrive to Mars

Speech/visual recognition: the feature of voice and image is hard to catch, it also means makeing rules is difficlut so it’s better to use machine

High-frequency trading marketing: it’s  hard for human beings to make decision in only just seconds in trading market but marchine learning can do it.The program just need to be trained by the data earlier in the trading market and it would make decision for the later quickly.

Consumer-targeted marketing : make service for people in a large scale in the case the service is targeted,we can’t do it by artificial method,we need use machine learning

What problems is suitable for machine learning

There are three keys which would determine whether a problem fit to machine learning:

(1)Exists some underlying pattern to be bearning

(2)It’s hard to definite rule

(3)There should be a certain amount of data

What’s the aim of machine learning?

We have some input:x and some output:y==>(x,y)

There exists a underlying function f which make f(x)=y,but we don’t know the function f

So we should  find hypothesis g : g ≈$\approx$$\approx$ f,there could be many function g ,but we need to chose the best one

What is the difference between machine learning,data mining,artical intelligence and statistics?

Machine learning: use data to compute hypothesis g ≈$\approx$$\approx$targeted f

Data mining: use huge data to find property that is interesting, when the interesting property is hypothesis ,data mining  ==machine learning

Artificical Intelligence : compute something that shows intelligent behavior,machine learning can relaize AI

Statistics :use data to make inferenceabout unknown process,traditionaly statistics values useing math provable result ,but machine learning puts highlight on computing data.Statistics can be used to realize ML,provides tools for ML
展开全文
• 一、What is Machine Learning ? Two definitions of Machine Learning are offered. 1、Arthur Samuel described it as: “the field of study that gives computers the ability to learn withou...
开始Machine Learning !
一、What is Machine Learning ?
Two definitions of Machine Learning are offered.
1、Arthur Samuel described it as: “the field of study that gives computers the ability to learn without being explicitly programmed.” This is an older, informal definition.
2、Tom Mitchell provides a more modern definition: “A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E.”
Example:playing checkers
E = the experience of playing many games of checkers
T = the task of playing checkers.
P = the probability that the program will win the next game.( performance measure:性能指标）

In general, any machine learning problem can be assigned to one of two broad classifications:
Supervised learning and Unsupervised learning.

附：入门机器学习的知乎帖子：
https://zhuanlan.zhihu.com/p/29704017


展开全文
• we will try to define what it is and also try to give you a sense of when you want to use machine learning. Even among machine learning practitioners, there isn't a well accepted definition of what i....
• I’v been preparing to learn Machine Learning for a few days, but always being delayed by some stuffs like building my blog. After beautifying the blog for a long time, I pick up my course today. ...
• 新定义：A computer program is said to learn from experience E with respect to *(关于，至于)*some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, impro.....
• t learning anything and I wanted to know, if you had some further debugging mechanics to see better, what is happening inside.</p><p>该提问来源于开源项目：facebookresearch/votenet</p></div>
• 【转载】https://hips.seas.harvard.edu/blog/2013/02/25/what-is-representation-learning/ Posted in Machine Learning. By Roger Grosse – February 25, 2013 In my last post, ...
• one is to learn by yourself, and the other one is learning by lecturers if we think learning as an approach to eat things from the cook that is an interesting approach to see basically, if you learn ...
• Two definitions of Machine Learning are offered. Arthur Samuel described it as: “the field of study that gives computers the ability to learn without being explicitly programmed.” This is an older, ...
• learning : acquiring skill with experience accumulated from observations. machine learning : acquiring skill with experience accumulated/computed from data ML: an alternative ...
• What is Machine Learning? What is Machine Learning? Two definitions of Machine Learning are offered. Arthur Samuel described it as: “the field of study that gives computers the ability to learn ...
• Deep learning 什么是深度学习？ Deep v.s. Shallow 深度学习的大起大落 世人都道深度好，为啥深度好？ Structured Learning 链接：http://speech.ee.ntu.edu.tw/~tlkagk/courses/MLDS_2015_2/Lecture/Brief%...
• 摘要: 本文是吴恩达 (Andrew Ng)老师《机器学习》课程，第一章《绪论：初识机器学习》中第2课时《什么是机器学习？》的视频原文字幕。...What is machine learning? In this article we will try...
• 1.what is Machine Learning? Learning：通过观察获取技能(skill) observation—&gt; learning —&gt;skill Machine Learning：电脑通过数据获得技能(skill) data—&gt;ML—&gt;skill(improve ...
• * supervised learning:  * regression   * classification *unsupervised learning
• 1.2 [what is machine learning？] 1.人：observation --> learing --> skill 机器：data --> ML --> improved performance measure /skill 2.什么情况下适合使用机器学习： （1）some 'underlying ...
• Meta learning is Learn to learn based on tasks (machine learning is learn to do a given task based on dataset). It does mean that it can handle new tasks with existing status, but it is likely to ...
• 转自：https://www.quora.com/What-is-regularization-in-machine-learning Regularization is a technique used in an attempt to solve the overfitting[1] problem in statistical models.* First of all, I...
• https://www.growthengineering.co.uk/what-is-game-based-learning/ Gamification and game-based learning have been hot topics in the world of learning and development for some time now, but did you...
• Book: An Introduction to Statistical Learning ...http://www-bcf.usc.edu/~gareth/ISL/这是第二章，简要介绍统计学习中的一些基本概念2.1 What Is Statistical Learning? 假定我们观察到一个定量响应变量
• What is machine learning Informal definition:the field of study that gives computers the ability to learn without being explicitly programmed. Modern definition:A computer program is said to learn fr...
• 中文笔记见：...What is Machine Learning? Two definitions of Machine Learning are offered. Arthur Samuel described it as: "the ...
• https://www.datamachinist.com/reinforcement-learning/part-1-what-is-reinforcement-learning/

...