Literature reading is an important skill for postgraduates , Therefore, how to efficiently read literature is a problem worthy of thinking . This article is from David R. Cheriton Written How to read a Paper, Deepen your understanding of literature reading .

Abstract

Scholars spend a lot of time reading literature , But I don't know how to read the literature , It leads to a waste of time and energy . This article outlines an efficient and practical “three pass” Method , It also describes how to conduct literature survey through this method .

Introduce

Most scholars read articles for the following purposes ：

• Course needs or meeting needs
• Know the latest progress in your research field , To obtain a new idea
• Literature survey on new research fields , Build awareness of the unknown

For graduate students , How to read literature needs to correct your reading style by constantly trying and making mistakes , This process often takes a lot of time , Sometimes it even makes people feel at a loss .
And use “three pass” A good way of reading will keep you from getting caught up in the details of the article until you know the general meaning of the article , Save a lot of time ; Allows you to assess the time it takes to review a series of articles ; Let you adjust the depth of your understanding of the article according to your needs and time .

“three pass” Method

** Get an article , Remember not to read from beginning to end , It's a waste of time , And can't grasp the key .** Should use “three pass” Methods , Each step needs to achieve a specific goal .

The first pass

Browse the full text quickly , Decide whether this article is worth further reading . This process takes almost 5-10 minute , And it includes the following steps ：

1. Read... Carefully title,abstract and introduction part
2. Read only article section as well as subsection, Don't worry about what's under the chapter
3. Look at the mathematical formula in the article , Try to understand the underlying mathematical foundation
4. Reading articles conclusions part
5. Browse references , Pay special attention to the references you have read

After the above steps , You need to answer the following questions ：

1.== What kind of article is this ？ It's a measurement paper ？ Analysis of existing systems ？ Description of the research prototype ？==
2. Which papers does this paper relate to ？ What mathematical basis is used to analyze the problem in this paper ？
3. Whether the assumptions in the paper are reasonable and effective ？
4. What is the main contribution of the paper ？
5. Is this paper well written ？

The first pass after , You may not want to read the article in depth . Maybe the paper is not what you are interested in , Or you don't have enough expertise in this field , So that you can't understand the paper well , Or the author did not make reasonable assumptions . For not in your field of study , But papers with some relevance ,The first pass That's enough .

The second pass

The second pass Need to read the paper more carefully , But ignore things like proof . This will help you in the process of reading , Jot down some key points or make comments in the blank space of the page . What comments should I make ？ come from Uni Augsburg Of Dominik Grusemann It is recommended to write down Terms you don't understand during reading , Or the question you want to ask the author . These comments are helpful when you really start writing a literature review .

1. Read the illustrations carefully 、 Chart . Whether the shaft is correctly marked ？ Whether the result is displayed in the form of error bar ？ Whether the conclusion has statistical significance ？ Common mistakes like this will distinguish between really good work and carelessness 、 Bad work .
2. Don't forget to mark some relevant unread relevant literature , In order to further understand the background of the article .

For an experienced scholar ,The second pass It should cost about 1 Hours of time . Through this reading , We should grasp the main content of the article , Summarize the main points of the article .
But sometimes , At the end of the process the second pass after , I still don't understand the article well . There are many reasons , It may be that the research topic of the paper is new to you , There are too many terms and acronyms ; Maybe the author used a proof you can't understand 、 Experimental techniques ; It may also be that this paper is not well written , There are too many unconfirmed assertions and a lot of cutting-edge literature ; Or maybe it's late at night , I'm a little tired . When you can't understand the article well , You can choose the following methods ：

1. Put the paper aside , I hope you can succeed without understanding these materials
2. Put the paper aside first , After reading more background knowledge , Go back and read the paper
3. If you can't understand it, you can't understand it , Go straight ahead the third pass

The third pass

The third pass The key is Repetition paper , Is to make the same assumptions as the author , The process of realizing the paper step by step . In the process of reproduction , It is easier to find the innovation of the paper , At the same time, you can also find the hidden errors and assumptions .
The third pass You need to pay great attention to details , You need to find the assumptions in the article and question them .** Besides , You need to think about how you can express a particular idea .** Reproducing the paper can make us have a deep understanding of the proof and demonstration skills in the paper , It also helps to add this knowledge to your tool base . And in the process of reproduction , You should write down your ideas for future work .
For beginners , Reproducing code is a difficult process , It takes time to stick to , To ponder . stay the third pass Last , You should reconstruct the framework of the full text in your mind , And know the advantages and disadvantages of this article . especially , You should be able to accurately point out the assumptions implied in the article , Missing references to related work , And potential problems in the process of experiment and Analysis .

Literature survey

For an unfamiliar field , Literature survey needs to read a lot of literature . So what kind of literature is worth reading , How to read literature ? It can be divided into the following steps ：

1. Using academic search engines , And the most common keyword search in this field 3 To 5 A highly cited article in recent years . Reading articles the first pass, Understand relevant work .
2. Look for references in the article and the names of repeated authors , Because these are the key papers and researchers in this field . Download these key papers from the Internet , And visit the websites of key researchers , Look where they've published recently . This will help you identify the top journals or conferences in the field , Because these best researchers usually publish papers in top journals or conferences .
3. Visit the websites of top journals or conferences , See their progress . Take a general look at , Learn about recent high-quality work . Found so far 、 The selected articles constitute the first version of the literature survey .
4. For these articles the second pass and the third pass.
5. If you read these articles , It is found that they all cite key articles and have not been searched before ,find it！！！, Iterate as necessary .