Doctor of MCU Computer, Chief Scientist of Amazon, "Hands-on Deep Learning" at Station B (Learn AI with Li Mu's personal space _ beep _ Bilibili) course instructor Mu Shen's advice on advanced learning for machine learners

Knowledge accumulation

You can passively obtain information through the circle of friends, but more importantly, actively obtain information and build a knowledge base

Accumulate knowledge by reading papers

• Read every sentence of a classic paper
• Combined code to understand the details of the paper, study website: The latest in Machine Learning | Papers With Code
• See open review: About | OpenReview reviews
• Organize the papers I have read, and introduce the method of organizing: How do I quickly read and organize literature_bilibili

Personal practice

Participate in the competition, but pay attention to the difference between competition and research, industrial application

The difference between machine learning in competition and industry applications_Mao Feilong's blog-CSDN blog

• Competition: To adjust the best parameters, complex methods such as model fusion will be used, just to get the best evaluation index results
• Research: new model, good parameter tuning, pursuit of innovation
• Industrial: abstracting the problem into a machine learning problem, data collection, the focus is on solving the problem, the model is usually not the most important

Do more research on open source code, communicate with developers, and actively contribute to the open source community

If you don't know how to contribute, you can start by organizing the documentation