Machine Learning ten years ago. Ten years ago, Ten years ago, Ten years ago, ten years ago, ten years ago, ten years ago, ten years ago, ten years ago, ten years ago, ten years ago, ten years ago, ten years ago, ten years ago, ten years ago. \

Back in early April, Ng introduced the book: “The point of this book is not to teach you machine learning algorithms, but to teach you how to use them for your own benefit.” Like some technical AI classes give you a hammer, but this book teaches you how to use the hammer. If you aspire to be a technology leader in artificial intelligence and want to learn how to set the direction for your team, this book will help.

After reading this book, you will have learned:

Identify the most promising directions for ai projects;

Diagnosing errors in machine learning systems;

Build ML in complex Settings such as mismatched training/test sets;

Set up an ML project to compare with human performance;

Learn when and how to apply end-to-end learning, transfer learning, and multitasking learning.

Screenshots:

Full contents: \

1 Why does machine learning need strategy? \

How can you use this book to help your team

3. Knowledge and symbol description

4 scale drives machine learning

5 Definition of development set and test set

Development and test sets should follow the same distribution

How big should the development and test sets be? ?

8. Use single value evaluation indicators for optimization

9 optimization indicators and satisfaction indicators

Accelerate iteration through development sets and metrics

11 When to modify the development set, test set, and metrics

Summary: Build a development set and a test set

Quickly build and iterate on your first system

Error analysis: Evaluate ideas against development set samples

Evaluate multiple ideas in parallel during error analysis

16 Clean mislabeled development set and test set samples

Divide the large development set into two subsets, focusing on one

How big is this setting?

Summary: Basic error analysis

Bias and variance: Two major sources of error

21 Bias and variance examples

22 compared with the optimal error rate

23 Deal with bias and variance

Tradeoff between bias and variance

Reduce techniques that can avoid deviations

26 Error Analysis of training set \

Techniques for reducing variance

Diagnostic bias and variance: the learning curve

29 Draw the training error curve

Reading the learning curve: High bias

Reading the learning curve: Other cases

32 Draw a learning curve

33 Why is it compared to human performance

How to define human performance level

35 Exceeds human performance

When to train and test on different distributions

How do you decide whether to use all your data

38 How do I determine whether to add inconsistent data

39 Add weights to data

40 From training set generalization to development set

Identify bias, variance and data mismatch errors

42 Resolve the data mismatch problem

43 Synthetic data

44 Optimize validation tests

Optimize the general form of validation tests

46 Examples of reinforcement learning

The rise of end-to-end learning

More examples of end-to-end learning

The pros and cons of end-to-end learning

Pipeline component selection: data availability

Pipelining component selection: task simplicity

Page 4 Machine Learning Yearning-Draft Andrew Ng

52 Direct learning for richer output

53 Perform error analysis based on components

54 Error is attributed to a component

The general case of error attribution

56 Component error analysis compared to human level

57 Discovering a flawed machine learning pipeline

Build superhero teams – Get your teammates to read this book


Deeplearning. ai has published a translated version of the book.

Link: pan.baidu.com/s/1OmlO8fAZ…

Extraction code: ZR03

If the link is invalid, please reply to “Training Secrets” to obtain the download address.


Machine learning enthusiasts QQ group: 654173748\

Please follow and share our official account: \

Essential article:

Machine learning for Beginners

Introduction to Deep Learning – Python Deep Learning, annotated version of the original code in Chinese and ebook

Printable version of Machine learning and Deep learning course notes \


Ng personal notes on machine learning course online edition \

DeepLearning. Ai DeepLearning course notes online edition

Mathematical foundations of machine learning \

Zotero paper Management tool