A method of directly advancing from AI Xiaobai to an artificial intelligence engineer with one year of experience

In 2017, artificial intelligence gave us too many surprises and changes. From this year, international giants have begun to make great strides in strategic shifts - from mobile priority to AI priority: Microsoft in March, Facebook in April, Google in May. Apple in June...and even in the Baidu meeting some time ago, the title of Director Li: The unmanned car ticket has come, will the volume be far behind? Directly pick up the climax of the day!

It can be seen that artificial intelligence has entered the eve of the global outbreak. Artificial intelligence technologies such as personalized information push, face recognition, and voice control have “invaded” the details of daily life. How hot is artificial intelligence? An AI company was born every 10.9 hours. Moreover, it is well known that a large number of domestic first-line Internet companies have been laying out machine learning teams to optimize smart products. Under such a background, it is not difficult to imagine that future machine learning technology will be a new threshold and field for technicians.

So the question is, how can I transform/learn AI technology as a technician? How to join this wave called the "fourth" technological revolution?

The "Artificial Intelligence Engineer" course project produced by CSDN's educational platform and CSDN Academy will take you step by step and learn from the shallower and deeper, master the knowledge and technology application practice of machine learning!

This is a 120-day study plan, the goal is: to allow you to directly upgrade from AI Xiaobai to an artificial intelligence engineer with one year experience!

"Artificial Intelligence Engineer" is divided into three stages, from machine learning to deep learning to project actual combat, step by step, and deeper.

Knowledge point overview:

This stage mainly explains the principles of machine learning, including common algorithms, model evaluation and selection, feature engineering and other essential knowledge of machine learning, and takes you to fully grasp the basic ideas and processes of machine learning. Finally, a commodity recommendation system will be implemented, combining various feature engineering techniques and machine learning algorithms to improve the ability to use algorithms, data cleaning and feature processing, and lay a solid foundation for industrial combat.

Join the first phase of the course and challenge the following practical projects:

Project 1: Housing price forecast case; data set exploration

Familiar with the classic algorithms, models and tasks of the machine learning field, learn to build and configure the machine learning environment, and learn to solve a practical problem with linear regression.

Project 2: House Price Forecast Case II

Commodity classification was realized by logistic regression, neural network and SVM; the performance of SVM (different regular parameters and kernel functions) under different models and different parameters was compared, and the characteristics of each model were realized.

Project 3: E-commerce commodity classification case

Learning Boosting integration ideas and tree-based integration algorithms, implementing e-commerce commodity classification projects through XGBoost, learning complex model parameter tuning

Item 4: Face image feature extraction: PCA, ICA, NFM. E-commerce user clustering case.

Learn to use dimensionality reduction techniques to reduce dimensionality of high-dimensional features, and learn to use non-supervised learning algorithms to accomplish related tasks through two practical cases.

Item 5: Product recommendation case

Learn common data preprocessing methods and feature coding methods; general principles of learning feature engineering; combine various feature engineering techniques and machine learning algorithms to implement recommendation systems

Project 6: The graduation project implements an actual product recommendation system.

Knowledge point overview:

Comprehensive understanding and mastered supervised learning, unsupervised learning, intensive learning and deep learning in the field of machine learning, and personally challenge cutting-edge applications.

Join the second phase of the course and challenge the following practical projects:

Item 1: Mnist Handwritten Number Identification

Familiar with the common terminology in the field of neural networks, install and configure the deep learning framework Tensorflow, learn to solve a practical problem with Tensorflow.

Project 2: Handwritten digit recognition (Mnist dataset) with CNN; verification code recognition

Commodity classification was realized by logistic regression, neural network and SVM; the performance of SVM (different regular parameters and kernel functions) under different models and different parameters was compared, and the characteristics of each model were realized.

Item 3: 20 categories/11530 image data sets: image detection tasks

Learn the image classification task and the current main model algorithm of the detection task, and train the CNN model under the Tensorflow framework through two practical case studies.

Project 4: 330,000 image datasets: Image semantic segmentation tasks.

Learn the mainstream image segmentation model and train and tune the CNN model under the Tensorflow framework through practical case study.

Project 5: CNN+RNN implements writing poetry robot

Learn the principles and applications of cyclic neural networks, and train and tune the CNN+RNN model under the Tensorflow framework through practical case studies.

Knowledge point overview:

At this stage, a large number of real data sets will be provided, and a complete practical project design will be carried out based on the interests and actual conditions. In the actual combat process of the project, the students are grouped, and the team members discuss the training model on their own, and constantly adjust and optimize, and finally show the project results. The teacher will review and provide guidance based on the specific project situation.

List of actual combat projects (optional):

Item 1: Natural Language Processing: Text Classification. According to the company's registration, investment and business scope and other relevant information, the company is classified and provides a reference for the company's valuation.

Item 2: Ad Click-through Rate Forecast (CTR) predicts the user's click-through rate for viewing a given page, improving the accuracy of ad delivery.

Project 3: Vehicle Detection and Model Identification - Use the deep learning method to detect the vehicle from the picture and identify its model number.

Project 4: Viewing a Talking Robot - Analyze image content using computer vision and deep learning methods and automatically generate textual descriptions of the images.

To learn any programming technique, you need a highly professional instructor, systematic, scientific curriculum, practical exercises and high-quality, efficient study and guidance, because it will not only help you save a lot of time, but also get twice the result with half the effort; Can guarantee that it will not be abandoned halfway.

CSDN Academy produced "Artificial Intelligence Engineer Through Train", dedicated to excellent learning, professional lecturers and enthusiastic teaching assistants, three characteristics to ensure learning results. 4 months, harvest yourself differently:

Exclusive course system, ingeniously polished, planned learning every day

Lecturer live Q & A, multiple assistants are available for counseling

The class teacher can check the progress of the students at any time to ensure the learning effect.

There are customs clearance exams at each stage, and the examination results ensure the learning effect and progress.

Congratulations on the certificate of completion! Become an AI engineer! Also get a job recommendation!