Public account [Machine Learning and AI Generated Creation] Backstage reply : 168 . You can participate in the free delivery of [Hands-On Machine Learning] book activity. The deadline for the activity is 10 pm on September 14, 2023 .
Official account backend reply: 168 (it is recommended to select and copy)
You can also purchase it yourself by clicking on the following JD.com exclusive link:
▲
Click on the cover to get a 50% discount on the book, limited to 50 people
Recently, the team of Professor Yu Yong from the ACM class of Shanghai Jiao Tong University released a heavyweight new book - "Hands-On Machine Learning" . This is a rare blessing for technical people, because it explains machine learning clearly at once.
How tough is the ACM class at Shanghai Jiao Tong University? Just look at its distinguished alumni.
In the field of scientific research, there are Yang Diyi who teaches at Stanford and Chen Tianqi who teaches at Carnegie Mellon University. He is also the author of XGBoost and the initiator of TVM. In the corporate world, there are Lin Chenxi, co-founder of Yitu Technology, and Dai Wenyuan, founder of 4Paradigm. Li Mu, the author of MXNet, and his book "Learning Deep Learning by Hand" is a must-read classic for many technical people.
Therefore, the ACM class is known as the "Whampoa Military Academy" for Chinese AI talents in the industry. Behind the great achievements is a great vision. Professor Yu Yong, the founder of the ACM class, has been working hard to cultivate Chinese Turing Award winners. The organization that awards the Turing Award is ACM (Association for Computing Machinery), which shows how much Professor Yu Yong had high expectations when he named the ACM class.
ACM class instructor: Professor Yu Yong
The content of "Hands-On Machine Learning" comes from the front-line teaching practice of ACM classes . Zhang Weinan, one of the authors of this book, systematically organized the theoretical knowledge and practice codes into a first draft based on his own experience in teaching machine learning courses and student feedback.
Subsequently, under the careful review and arrangement of the main creative team, this masterpiece in the field of machine learning was finally delivered. The book's lesson pack is now available for pre-sale. Scan the QR code and add the assistant WeChat to purchase the book pack at a discounted price of 99 yuan !
02
What do you need to learn to understand machine learning?
"Hands-on Machine Learning" is positioned to guide beginners to get started. In terms of content setting, it starts from explaining basic theoretical algorithms and gradually moves into the discussion of supervised learning models and unsupervised learning models, striving to allow learners to systematically master machine learning. main knowledge.
The main creative team of this book has three authors. In addition to the chief instructor, Professor Yu Yong , there is also Associate Professor Zhang Weinan , who compiled the teaching results into a book . He has made great achievements in the fields of reinforcement learning, data mining, and knowledge graphs. The author Zhao Hanye also has in-depth research on reinforcement learning and machine learning.
Such a strong technical team has set an authoritative benchmark for this book in the industry.
The contents of the four main parts of the book are explained below.
Machine learning basics
In the basic part, it mainly helps learners grasp the core concepts and principles, and explains the two most basic algorithms: KNN (K nearest neighbor algorithm) and linear regression. Based on these two algorithms, the basic ideas and experimental principles of machine learning are discussed.
Once you have a solid grasp of this part of the content, you will have the ability to practice problem solving in most machine learning scenarios.
Parametric model
This part mainly discusses parametric models for supervised learning tasks, including logistic regression, bilinear models, neural networks and multi-layer perceptrons, convolutional neural networks, recurrent neural networks, etc.
The common feature of these methods is mainly to find the gradient of the model parameters based on the loss function of the data, and then update the model.
non-parametric model
This part focuses on non-parametric models of supervised learning, including support vector machines, decision trees, ensemble learning and gradient boosting decision trees.
The reason why non-parametric models are included as a separate part is to allow learners to better understand the differences, advantages and disadvantages from parametric models in terms of principles and codes.
Unsupervised model
This section discusses unsupervised learning methods for processing unlabeled data, including K-means clustering, principal component analysis, probabilistic graphical models, EM algorithms, autoencoders, etc.
Unsupervised learning is discussed from different tasks and different technical perspectives, so that learners can fully understand the difference from supervised learning.
Now that you know what you want to learn, let’s talk about how to learn it.
03
Use your brain but also your hands
The biggest highlight of "Hands-On Machine Learning" is that it provides a fool-like experience environment for hands-on practice . The creative team has refined the practical results of the ACM class and combined theory with code so that learners can get started smoothly.
The book includes conceptual definitions, theoretical analysis, algorithmic processes and runnable codes of machine learning. Learners can flexibly choose the content they want to read based on their own learning conditions.
However, after all, it is not a zero-based introduction. Before starting, learners must have two basic abilities. One is mathematical concepts and mathematical statistics knowledge , including matrix operations, probability distribution and numerical analysis methods, etc.; the other is basic Python programming ability . Able to understand code and debug and run it.
The rich technical cases in this book cover basic algorithms, parametric and non-parametric models of supervised learning, and unsupervised models. Learners can practice the knowledge points mentioned in the book by themselves. The main creative team carefully selected the code examples, striving to be simple in function and easy to modify.
All sample codes are classified by chapters and generated in ipynb format. Learners can download them from https://github.com/boyu-ai/Hands-on-ML .
The codes provided in the book are all implemented based on Python 3 and the PyTorch framework, and can be run with any online or local tool that supports the ipynb format.
The Python tool libraries used in the examples are briefly described. Each sample code contains variables that can be set by learners. Learners can modify and run the observation results in real time.
The following is an example of running the KNN algorithm in the Visual Studio Code environment.
code area
Run output
Isn’t it still full of ideas? There are more surprising rich media learning materials, let’s take a look.
Online teaching videos: The main creative team thoughtfully recorded video courses. How to watch: When you see the icon "Scan the QR code to watch the video course" in the book, pick up your phone and scan the QR code immediately.
Supporting PPT courseware: It is convenient for learners to review the main knowledge, and it is also helpful for teachers to prepare lessons.
After-class exercises: See if you have learned well. Don’t be a half-empty teacher. You must be able to do all the questions.
Learning community: When three people travel together, there must be a teacher. The book provides everyone with a way to learn in groups and communicate with peers anytime and anywhere.
The content of "Hands-On Machine Learning" has been tested by ACM classes before it was written . The principle explanations, algorithm descriptions, and code cases in the book are all eliminated through the interaction of teaching and learning. It can be said that this book is The crystallization of thinking and practice.
When advanced teaching results leave the campus, spread and popularize machine learning knowledge in the industry, and promote the development of the industry, they also attract praise from the big guys.
This book starts with the basic concepts of machine learning, combines it with the sklearn machine learning algorithm library, and leads readers into the world of machine learning with a large number of examples and codes.
——Zhou Zhihua, Director of the Department of Computer Science and Technology, Nanjing University
And Dean of the School of Artificial Intelligence
This book is derived from the machine learning course notes of the ACM class of Shanghai Jiao Tong University. It focuses on helping readers integrate, deeply understand the principles of machine learning and improve their practical capabilities. It is a rare good book for cultivating a new generation of practical artificial intelligence talents.
——Zhu Jun, Bosch Professor of Artificial Intelligence, Department of Computer Science and Technology, Tsinghua University, IEEE Fellow
This book is a comprehensive and systematic machine learning textbook, which mainly introduces the core concepts and representative methods of machine learning. This book not only covers classic machine learning theories such as neural networks and ensemble learning, but is also equipped with codes that can be run online to help readers enhance their understanding of machine learning technology through hands-on practice. Whether you are a beginner or an industry professional, you can get valuable help and guidance from this book.
——Li Wenxin, professor at the School of Computer Science at Peking University, famous teacher in Beijing, and director of the Computer Experimental Teaching Center at Peking University
The form of this book is novel and rich. Paper books provide a systematic introduction to the knowledge system of machine learning, theoretical interpretation video courses provide a more thorough explanation of difficult knowledge points, codes support online running and modification, exercises help readers test and consolidate learning effects, and teaching courseware helps college teachers prepare lessons. This book can not only provide systematic learning resources for industry professionals to learn machine learning on their own, but also help college teachers and students complete the teaching and learning of machine learning through hands-on learning.
——Huang Xuanjing, Professor, School of Computer Science and Technology, Institute of Modern Linguistics, Fudan University
Intelligent upgrades have become a new ladder for contemporary enterprises to climb after digital transformation. To this end, a large number of artificial intelligence talents with professional knowledge are needed. They must not only have an in-depth understanding of the theories and algorithms of artificial intelligence, but also have practical capabilities to cope with challenges in practical application scenarios. This book is written to cultivate this kind of practical artificial intelligence talents. It has detailed content, rich examples, and clear code. It is highly recommended to all readers who want to understand artificial intelligence in depth and carry out hands-on practice.
——Liu Tieyan, Distinguished Chief Scientist of Microsoft, Head of Asia Region of Microsoft Research Scientific Intelligence Center, Vice President of Microsoft Research Asia
The ACM class at Shanghai Jiao Tong University focuses on cultivating students' solid professional foundation and hands-on practical ability. This book written by the team of teacher Yu Yong embodies the unique concept of the Shanghai Jiao Tong University ACM class in cultivating talents in the basic module of artificial intelligence. It is recommended to all readers.
——Tianqi Chen, Assistant Professor, Department of Machine Learning and Department of Computer Science, Carnegie Mellon University
04
To play with machine learning, you need to make good use of non-explicit programming
Fundamentally speaking, machine learning is a discipline that studies algorithms , and the role of these algorithms is to use empirical data to improve the performance indicators of a certain task through non-explicit programming.
So the difficulty here is that even if we understand the algorithm, we cannot go back from the results, which is difficult to restore. Explicit programming is much more intuitive and provides answers directly to the problem. For example, if you want to sort a set of data, implement a quick sort algorithm.
For learners, if they want to apply machine learning well in their own business fields, they must change their thinking, "step back", and switch from explicit programming to writing machine learning algorithm programs . This is what non-explicit programming means. In different tasks, a model to solve the problem is trained based on the data of the task itself.
1-1 Traditional explicit programming and non-explicit programming of machine learning
"Hands-On Machine Learning" is to help learners systematically understand machine learning, understand the principles of algorithms, and learn to use non-explicit programming to solve their own problems. The book provides out-of-the-box sample code for each algorithm, so learners can easily expand it and use it in actual work.
▲ Click on the cover to get 50% off the book purchase
Now, analyze the tasks and data at hand, quickly use machine learning methods, and become a super individual in the intelligent era!
—END—