Neural networks and deep learning Chapter 2: machine learning overview reading questions

1. What is machine learning ？

informally , machine learning （Machine Learning,ML） That is, let the computer learn from the data automatically , Get some kind of knowledge （ Or law ）． As a subject , Machine learning usually refers to a kind of problem and the method to solve it , That is, how to get data from observation （ sample ） Looking for the law in the world , And use the learned rules （ Model ） Predict unknown or unobservable data ．

2. Why should each sample in the training set be independent and identically distributed ？

The question of Zhihu ： Why in machine learning , Let's assume that our data is independent and identically distributed ?

3. There are three basic elements of machine learning ？

Model , Learning rules , optimization algorithm

4. Why does the model over fit ？ What is the method to prevent the model from over fitting ？ It's not fitting ？

5. Why use SGD, and mini-batch gradient descent ？

6. Why use Ridge Regression instead of vanilla Linear regression of ？

7. What is the maximum a posteriori estimate ？

8. What is the connection and difference between deviation and variance ？

9. What is supervised learning , What is unsupervised learning , Illustrate with examples ？

10.L1 Why does regularization lead to sparse features ？

11.F1 When to use macro average for scores , When to use micro averaging ？

12. Why use cross validation ？

13. Talk about right no free lunch The understanding of the theorem

14. Talk about the understanding of the principle of Occam razor

15. What is? inductive bias（prior）？