Python now plays an irreplaceable role in the field of artificial intelligence and data analysis. Many machine learning frameworks support Python API, which facilitates data analysis, storage, acquisition and computation. Therefore, Python has become the first language of machine learning in the field of artificial intelligence.
I am writing this article because many people want to know how to get started/switch to Python machine learning and hop on the 21st century train of artificial intelligence. In addition, this problem will be mentioned every once in a while, so I want to write an article to once and for all.
Purpose of the article:
- Point out some misunderstandings in learning
- Provide an objective and feasible study list
- Give suggestions for further learning
The target audience is:
- Zero-based readers interested in artificial intelligence
- Friends who have the foundation and want to combine machine learning/data analysis with their job
- Student friends
- Has worked, has other programming foundation wants to turn artificial intelligence friend
Some mistakes in learning
1. Don’t try to master all the relevant math before starting machine learning
It takes a long time for ordinary people to complete all these knowledge before starting machine learning, which is easy to give up halfway. And that knowledge is a tool, not a goal. It’s not our goal to be mathematicians. It is suggested that the machine learning process should be more purposeful and less time-consuming.
2. Do not collect more information & determine the timeliness of the information
There are a lot of machine learning materials, often hundreds of gigabytes of material can be downloaded to watch, many friends have “collecting addiction”, in fact, is just put there. In the introductory period, it is recommended to “small and fine” selection of materials, find suitable for you, understand the start of action.
3. Practice more and get to know more industry giants and communicate more
Machine learning some algorithm selection is no exaggeration in practice to explore experience and skills, not action is certainly not. In addition, with some industry more exchanges and learning, have the opportunity to know must get in touch. These people have a lot of resources and contacts on them or around them. Try your best to find some of them, whether it’s questions on the way to study or future job opportunities.
Machine learning course schedule
| Python based | Mathematical basis |
|---|---|
| Function-class-object-oriented containers, file processing modules, standard library data structures | Probability theory statistics linear algebra calculus |
| Python Data Science | Supervised learning |
| NumPy SciPy Pandas Matplotlib Scikit-Learn | Decision tree linear regression logistic regression naive Bayesian support vector machine ensemble learning EM algorithm |
| Unsupervised learning | Semi-supervised learning |
| K-means algorithm DBSCAN clustering principal component analysis | Collaborative filtering for label propagation |
| Deep learning | Deep learning framework |
| BP neural network Convolutional neural network Cyclic neural network recursive neural network deep neural network | TensorFlow MXNet Caffe2 PaddlePaddle Keras PyTorch |
| Natural language processing | The project of actual combat |
| Tf-idf Word2Vec FastText | Spam filtering License plate number recognition Face recognition financial intelligent decision making system natural language emotion analysis recruitment network anti-fraud system…… |
This is a more systematic learning outline, involving a lot of knowledge is very wide, during the learning methods and skills in this with this text is certainly not over, also do not understand.
In view of this, I have prepared several open classes, which will be supplemented and improved by video explanation, animation demonstration, application scenes, internship recommendation and other aspects.
Content of open class (start at 20:00) :
- Learning Python machine Learning in 3 months (11.20)
- Naive Bayes Spam Filtering (11.21)
- Financial Intelligent Decision System (Time Series)(11.28)
- Mechanism of Natural Language Emotion Analysis (12.05)
- Face Recognition (Neural Network \OpenCV) (12.12)
If you don’t catch up, there will be video playback, temporarily set these 5 themes, continue to update later, all free, all free, all free.
You can add wechat: MIDU25 or long press to identify the following QR code to consult the details of the class, marked: class
My name is what, you can call me Pierre (peer), focusing on the Python machine learning research, Chinese Academy of Sciences in 2013, Dr After graduation go to France to work for more than 2 years of the national academy of sciences, considering the AI for the motherland development contributes an own strength (albeit small), returned to her home in 16 years, a top AI company currently works in Beijing.
Friends in Beijing can meet offline, I also have some resources and channels to recommend you to study or internship or employment.
In addition, if you come to the class, each of you will get a free copy:
- Python 3.5 Zero-basics tutorial;
- Introduction to Applied Mathematics & Machine Learning 0
In-depth research, practice
Congratulations to you! If you have completed the schedule above, you are already quite capable of machine learning.
The next is to contact the actual combat as soon as possible, can be a variety of forms, such as practice, work, scientific research, into the laboratory and so on.
For most of you who are already working, going back to school for a degree isn’t practical, so try applying machine learning to your own work.
In the end, whatever direction you choose, the most important thing is the ability to think independently and the courage to take the first step.
\
\