How to Read Academic Papers without Freaking Out

James Lee
Dec 29, 2017 · 6 min read

Research is one of the key contributions of any educational institution. Months and even years worth of knowledge and experiments are all condensed into the few pages of an academic paper.

There is an abundance of academic papers to be read right now, whether it’s for some novel method in Artificial Intelligence or a whitepaper of an ICO (not necessarily an academic paper but presented in the same way), reading academic papers can be quite the headache, especially for the uninitiated.

“We have enough papers. Stop publishing, and start transforming people’s lives with technology!” — Andrew Ng from his speech “AI is the new electricity” at the AI Frontiers Conference

To implement papers, we first have to able to read them efficiently. For the most part, people aren’t always taught how to best approach academic papers. As a result of this a lot of time and effort goes to waste, and reading academic papers has sometimes become a chore, even for those who have spent many years doing it already.

So before you dive head first into exploring the world of emergent research, I’ll share a few tips and tricks that I find useful when reading papers.

Before You Begin

Reading an academic paper is an entirely different process than compared to reading a blog post or news articles. It is often better to not read the sections in the order that they are presented, but you would also have to go through it multiple times. If you’re just starting out, it’s likely that you will have to spend a lot of time on a single paper (I took a few hours when I started), but you’ll get faster over time for sure. :)

Most research papers will be divided into the following sections: Abstract, Introduction, Methods, Results, and Conclusions/Interpretations/Discussion, References. It really depends on what the subject area is and which journal it’s published too. Some papers come with supplementary material or appendices that complement the existing sections.

It’s All About Reading Efficiently

If you intend to become an avid reader, you will have to learn to be able to extract the maximum amount of information with the least effort and time expended.

To do that, there are a few sections of a paper you should head to first:

  1. The Abstract — If provided, this section of a paper gives a general overview of the paper’s content. It’s good to start here first, to sort of “seed” the ideas and concepts of the paper into your head.
  2. The Introduction — Like the abstract, this segment gives you the general overview and setting of the paper. Sometimes you’ll find brief descriptions of key concepts or phrases, which are always a big plus to go through.
  3. The Conclusion /Discussion — This the section of the paper where the hypothesis is either accepted or rejected. More often than not I find that reading the conclusion can help you decide if you want to spend your time reading the paper or not if you are looking for a specific method of doing something.

If you read these 3 sections first, you will find that you’ll be able to have a better sense of what’s going on in the paper. Especially in the technically heavier parts like the Methods section. You’ll also be able to decide if this paper is something that is truly relevant to what you are looking for.

Reading the introduction and conclusion is the fastest way to determine the problem statement and the approach taken to the problem by the authors.

Read the Methods Section Last

Read the Methods section only after you’ve determined that the problem statement, methods taken to solve it and the findings are relevant. Else you might be lost in all the technical jargon that they often include.

The Methods section is more often than not the most technical part of a paper. It describes the procedures that the authors take to solve the problem. Be prepared to go through this section multiple times, especially if it’s a math heavy piece. Oh and don’t be afraid of losing track of all the mathematical symbols, it’s perfectly normal.

You Will Not Understand Everything in One Read Through

It’s true. Even after reading hundreds of papers myself, I still wouldn’t be able to understand everything in one go.

When you encounter something you don’t understand, there are 2 things you can do:

  1. Plow right through the unknown word or term. If it’s something important, you’ll most likely be able to derive it’s meaning through the context in which is used in. Make a list of each word/term that you don’t understand. You can always come back to it later. This method prioritizes speed.
  2. Pause and look-up the meaning of the word. I only do this when I want to study a paper in-depth. This is usually after curating my papers and deciding on a paper that will give me key insights on a specific area. I don’t really recommend this, cause it’ll disjoint your attention on the paper and place it elsewhere, but it works for some people. This method prioritizes understanding.

Don’t Ignore the References

If you’re new to the subject area, it’s also a good idea to read papers from the references. As you read more and more papers from a particular subject area, you’ll start to realize that a few common titles would appear more than once. They might be a little dated, but these are often defining papers that introduce the best way to do something.

Finding and reading these papers often boost your understanding of the subject as a whole much more than compared to randomly picking papers to read.

Lookout for Keywords

There are a few phrases that you should always be on the lookout for. To my knowledge there are 2 that are important — “key contribution” and “significant”

The phrase “key contribution” often indicates a finding that the paper discovers that captures the entire essence of the paper. If present, finding these phrases alone is enough to give you a brief overview of the paper’s work.

“Significant” on the other hand indicates something of statistical importance. It’s often a key result from the paper.

Be Curious

As an academic, it is important that you be inquisitive and curious about the things authors declare in their papers. Do not merely take the words you read as truth and accept them. Pick up the habit of questioning them, ask why this method was used and not another, put yourself in the author’s perspective and figure out their train of thought.

These are all habits that will naturally groom you into an efficient reader of academic papers, as well as a solid scientist. 👍

Prioritize Speed, Efficiency and Value

Above all else, prioritize the speed at which you read. Lingering on a single paper only cuts you off from reading other papers that will yield more insights.

Prioritize efficiency. Don’t waste your time and effort on papers that are irrelevant. There’s way too many out there for anyone to ever hope to read all. You have to learn to be objective about your topics.

And lastly, prioritize the value you can derive from any paper. Your goal is to learn, and if you can’t derive any value from it you should move on to something else that will. After all, the world needs scientists like you to make the world a better place. 👊


Find this useful? Feel free to smash that clap and check out my other works. 😄

James Lee is an AI Research Fellow at Nurture.AI. A recent graduate from Monash University in Computer Science, he writes about on Artificial Intelligence and Deep Learning. Find him on Twitter at @jamsawamsa.

AI Saturdays

Making rigorous AI education accessible and free, in 50+ cities globally. Sign up at https://nurture.ai/ai-saturdays

James Lee

Written by

James Lee

Future Tech. Ai, Blockchain and game design enthusiast. AI Research Fellow at Nurture.Ai & moderator of the FB group Awesome AI Papers

AI Saturdays

Making rigorous AI education accessible and free, in 50+ cities globally. Sign up at https://nurture.ai/ai-saturdays