Researchers usually need to spend a lot of time and energy on paper reading. However, few people will teach us how to read a paper, and the wrong reading method will often waste a lot of our energy. This paper puts forward a practical and efficient reading method of scientific research papers-three times reading method. ...
1. Nobody wants to read newspapers when there is nothing to do. When a researcher decides to read a paper, he usually has the following reasons that he can't refuse: ...
2. The key point of the three-time reading method is that you should read a paper (at most) three times instead of reading it from beginning to end (multiple rounds, nonlinear). ...
2. 1. The first pass is a quick browse, so as to have a macro understanding of the paper and determine whether further reading is needed. ...
2.2. In the second reading, you need to read the paper more carefully, but you should ignore the details such as proof.
2.3. The third time
If you want to fully understand a paper, you need to read it three times. The focus of the third reading is to try to reproduce this paper, that is, to deduce the work in the paper again based on the same assumptions as the author. By comparing the differences between you and the author, you can not only easily get the innovation of the paper, but also easily find the hidden assumptions and shortcomings.
The third time needs to pay attention to details. You need to find out every hypothesis hidden in every sentence and question it. Besides, you need to think about how to come up with an idea yourself. This process of contrast between reality and reality can help you deeply understand the methods of proof and presentation in the paper and internalize these methods into your own knowledge. In this process, you should write down your thoughts about your future work.
This reading may take 4 ~ 5 hours for beginners, and it will take about 1 hour after proficiency. After this round of reading, you should be able to reconstruct the structure of this paper and find out its advantages and disadvantages. You should be able to find out hidden assumptions, missing references to related work, and problems in experiments and analysis.