Li Hongyi machine learning introduction -2022.07.11

Machine learning introduction summary ：
Machine learning is to make machines have the ability to learn .
Understand machines from two perspectives ：
1. personification ： The language can be recognized through the program , Or identify categories .
2. Pragmatic ： Find the right function , Identify data by function
The process of machine learning ：
1. Set a series of functions
2. Training materials ： The machine judges the quality of the function according to the quality of the training data
(1)inputs
(2)outputs
（1,2 Training process ）
3. Choose the best function
Enter information to start training , Output training results （ Testing process ）

The task of machine learning revolves around two categories ：
1.regression: Enter the value , The output is used to make future predictions based on existing data
2.classification:
(1)Birary Classification: Such as yes or no Simple classification of
(2)Multi-Classification: Classify different materials correctly , Not limited to two categories . Such as ： Classification of news articles , It can be divided into politics , economic , Literature, etc .

Machine learning classification ：
1.supervised Learning: Through a large number of label To learn
2.Semi_supervised Learing: Information has label and unlabel To learn
3.Transfer Learning: Through a large number of different label Materials to learn , To judge the effect of irrelevant material learning
4.Unsupervised Learning: Information unlabel, have only outputs No, Inputs Let the machine learn under such circumstances , To see the effect of learning
5.Reinforcement Learning: Don't tell the machine the correct answer , Only score the learned behavior of the machine , Let the machine learn from evaluation
6.Structured Learning: Let the machine output structural things . For example, will ‘ machine learning ’ Translated into English .

summary ： Machine learning is mainly through determining the behavior of learning , The problem to be solved , Use different methods to determine how the machine will solve the problem .