Reading research paper can be a big hassle for us. But if we follow some simple steps this can be very efficient. The simple steps stated below in this blog.
How To Read Research Paper
There are some method for reading a research or technical paper. But first of all we have know what is research paper.
What is a research paper?
According to Wikipedia, a paper that holds some results, review or some new invention based findings. So, by definition, It is a systematic or documentation way to share our new finding on various fields or review of some existing systems or improvement of existing system.
Who are going to read these research or technical Papers?
Researchers must read papers for some reasons. Some of the most common reasons are listed below.
- to review them for any conference or journal
- keep current knowledge in their field
- for literature survey of a new field
Here we are going to introduce the “Three Pass” method for reading research paper. The “Three Pass” method is way to read papers. This method is consists of 3 phases introduced as a pass. It is developed by some researchers based on their experience. So, Let’s start our journey of this blog.
The First Pass
What is it?
- it is a bird eye view of the paper
- this pass tells us whether we should read further or not
Duration: It is going to be 10–15 minutes read.
Steps:
- we need to carefully read the Title, Abstract, Introduction
- must need to only read the sections and sub-sections heading but by ignoring every else under the sections and sub-sections.
- read the conclusion
- take a glance over the reference section and mentally ticking off the ones we have already read. If we don’t have any previously read paper in the reference section that is also fine.
After “the first pass” we should answer some question. That is,
- Category: What type of paper is this? (measurement, existing system analysis, description of research prototype).
- Context: Which theoretical bases being used to analyze the problem?
- Correctness: Do the appeared assumptions are valid or not?
Don’t panic if you don’t find the correctness. It will come after some experience in this field.
- Contribution: What are the paper’s main contribution?
- Clarity: Is this paper well written?
Tips: If you are a researcher, it is going to very beneficial for you. Say, you submitted a paper of some journal or conference. Then this two things needed to ensured that,
-title, sections and sub-section are properly written
- a concise and comprehensive means a good sounded abstract
If a reviewer can’t understand your “The First Pass” of the paper, this will be ended up being a rejected one.
Using these information you need to decide whether to read further or not.
The second Pass
Duration: 1–2 hours
Steps:
- Look carefully at the figures, diagrams and any kind of illustrations in the paper. Carefully read the description of these. You can find the description behind or after the illustrations.
Try to answer these question. This will lead you to decide whether it is shoddily written or not.
- Are the axes properly labeled or not?
- Results shown in tabular format or using some plotting techniques. Plotting is more preferable for understanding the statistically significant conclusion.
- Remember to mark relevant unread references for further reading. (This is a good way to learn more about the background of this paper)
After the second pass we should be able to answer these questions.
- If we have the grasp of the content of this paper or not?
- Can we summarize the main trust of the paper with supporting evidence?
Tips: Sometimes we won’t understand the paper even after the second pass. This is very common to all of us. This may occur because,
- the subject matter is new to us
- unfamiliar terminology, acronyms
- Or, simply you are tired. Take some rest and read it again.
The Third Pass
Who need this pass?
- Reviewer
- Who wants to improve this study
- Who wants to know what are the papers achievement actually
Duration: 5-infinite hours. (Just make you life simple, take your time)
Steps:
Only one step needed in this pass. That is,
- Virtually re-implement the paper’s work
Example: VGG16 paper implementation, RESNET paper implementation.
After doing this one can have knowledge about these things,
- the paper’s innovation (if have any) or workings
- hidden failings and assumptions
- during this pass one should also can get jot down some ideas for future work of this paper
- identify strong and weak points
- point out the implicit assumptions, missing citations to the relevant work with experimental or analytical techniques
Conclusion: This is a simple method for reading any types of paper. The first pass is just reading the Title, Abstract, Introduction. And after this pass one should decide whether to read further or not. The second pass is mainly digging deep with help to illustrations of the paper. And grasp the work of the paper and can summarize the paper’s work. Finally the third pass, where we have re-implement the whole work virtually. Again this pass in not for all types of readers. Finally that’s how we can read a paper as simple as fast.
References: