Wisdom is not a product of learning, but a lifelong attempt to acquire it.

Albert Einstein

Everyone who is seriously engaged in machine learning needs to learn to understand what is published in scientific articles. Such publications are made by scientists who are at the forefront of research in relevant fields. These are artificial intelligence (AI, Artificial Intelligence), machine learning (ML, Machine Learning), deep learning (DL, Deep Learning) and many other areas.

ITKarma picture

In order to stay up to date with the latest discoveries and expand your own knowledge, you need to have a scientific mindset and appropriate habits. AI, ML and DL technologies are developing at an incredible rate. Therefore, we need to keep up with progress, stock up with relevant knowledge. This knowledge can only be obtained in the course of work with scientific publications.

Here you will find guidance on how to work effectively with scientific articles. In particular, we will focus on the following topics:

  • A systematic approach to reading collections of publications to gain knowledge in your area of ​​interest.
  • Rules for reading scientific articles.
  • Useful online resources that can help you find publications and critical information.

Who is Andrew Eun?

This article is based on lectures delivered by Andrew Eun. In addition, I have added here some of my own recommendations and useful information found on the Internet.

To get started, I want to briefly talk about who Andrew Eun is ( Andrew Ng ).

Andrew Eun is an assistant professor of Stanford University . He is probably the most famous (and one with the largest audience) lecturer in the field of machine learning. Andrew is also a co-founder of Deeplearning.ai and Coursera .

How do people gain useful skills?

It’s completely natural for a person to adopt the skills and habits that those around him demonstrate. It is precisely because of this that doctoral students develop skills for the effective assimilation of information from scientific publications. This is, to a certain extent, a well-known fact. Andrew mentions this at the very beginning of his video lecture, to which I gave the link above.

But we are not students, although some of you may be a student. Therefore, we are faced with the question of how an ordinary person should acquire the skills necessary for reading and a deep understanding of scientific publications.

Methodology for selecting reading materials

To a talented person who wants to work in the field of machine learning, it is best to specialize in something. For example, possessing general knowledge from the field of computer vision (Computer Vision) is commendable. But someone who has specialized knowledge and experience in using the basic methods for solving the problem of assessing the position of an object in space (Pose Estimation, PE) will look much more attractive from the point of view of a potential employer who needs a specialist in this field.

Let us, using the example of the PE problem, analyze the methodology of working with scientific publications on the topic of interest to us.

▍1. Material Selection

In the first step of the work, we will formulate a selection of resources related to the subject of our interest. The concept of “resources” includes scientific publications, articles from Medium, materials from blogs, videos, GitHub repositories, and so on.

If you search on Google for the phrase pose estimation , we’ll have a set of links to the main resources related to the question that interests us.At this step of the work, our goal is to collect everything that may suit us. This is a video from YouTube, and documentation relating to the practical implementation of the machine vision mechanisms of interest to us, and, of course, scientific articles. Here, ideally, you do not need to be limited to any specific amount of resources that you consider important. The main thing is to make a final list of materials that can benefit you.

▍2. Analysis of materials and assessment of their level of understanding

Here we will analyze the resources that were previously recognized as potentially useful and related to the issue of interest to us. It is very important to consider that there is a method for assessing the level of understanding of materials included in the final list at the previous step of the work. Andrew Eun advises making a table in which you can indicate the level of understanding of each of the materials. This table may look something like the following.

ITKarma picture

The resource table used to assess their level of understanding

It is recommended that you try to read at least 10-20% of the contents of each document added to such a table. This will allow you to get acquainted with such a volume of material, which is enough to accurately check how much the material meets our needs.

Materials that are better than others correspond to the subject of our interest, you need to understand more deeply than with others. As a result, you will find some suitable resources that you fully understand.

Perhaps you are now wondering how many articles or other resources should be fully understood by you.

I have no answer to this question, but Andrew has an answer.

Namely, he says that an understanding of 5-20 materials indicates a basic level of orientation in the issue. Perhaps this level is enough to move to the practical implementation of various methods.

If it comes to understanding 50-100 materials, it means that you are very well versed in the issue.

After you analyze the resources and learn something about them, your table will look something like the one shown below.

ITKarma picture

Updated table to assess the level of understanding of resources

▍3. Brief Description of Materials

The third step of preliminary work with materials is based on my own experience. When I try to understand scientific publications, I make structured notes in which I briefly summarize, in my own words, the main results, findings, methods described in the papers.

Now, after we find the materials that are worth reading, we move on to reading them.

Read a scientific article

When they read something in order to understand, just reading the material is not enough. Andrew says that reading an entire article in one sitting may not be the best way to get an understanding of the material.

Be prepared for the fact that in order to properly master the article, you will have to read it at least three times.

▍4. First Reading

When you read the article for the first time, focus on the headline, annotations, and drawings.

▍5. Second Reading

At the second reading, pay particular attention to the introduction and conclusion. Also, take another look at the drawings and skim the rest of the article.

The introductory and final sections of the article contain clear and concise information about its contents; they summarize the main findings made by the authors of the article. These sections usually do not include any supporting information. Here is just what is really important. Thanks to this, the reader is prepared to perceive the rest of the article.

▍6.Third reading

At the third pass through the article, the main text is read while skipping complex mathematical calculations or descriptions of techniques new to the reader. At this stage, you can also skip obscure or new terms.

▍7. Next Reading Article

Anyone who is engaged in deep research in a certain field can read the article several more times. These additional reading sessions will be mainly aimed at the analysis of mathematical calculations, the development of techniques, to clarify the meaning of unfamiliar terms.

Anyone who usually reads scientific articles for informational purposes, and in order to quickly get acquainted with their proposed methods of work, may be faced with the fact that in-depth study of articles takes up a lot of his time. Especially when it comes to dozens of articles that need to be worked out.

Here is an example of what I call "deep research" here. I read this article using the methodology presented here, figured it out, and then, on it basis, wrote 4 own articles ( 1 , 2 , 3 , 4 ).

Questions to ask yourself when reading an article

Andrew gives us a list of questions to ask yourself when reading this article. These questions are generally aimed at revealing an understanding of what you are reading. I use the following questions as guidelines, so that I can’t deviate from my main goal - understanding the essence of the article.

These questions are:

  1. Describe what the author of the article is trying to achieve, or perhaps what he has already achieved.
  2. If you are faced with a new approach to solving a problem, with a new technique or technique, describe their key elements.
  3. What exactly in the article do you consider the most useful for yourself?
  4. What materials mentioned in the article would you like to read?

Additional resources that may help research

In my search for information, some resources helped a lot, the list of which I give below.

You may well make such a list yourself.


Long-term results are provided by constant stable study, not learning anything by “raids."

Andrew Eun

I am still a relatively new person in the fields of machine learning and computer vision. And here, to put it mildly, there is still a lot of things that I do not know. But, despite this, I am sure that if a person is consistent in his search for knowledge, no matter in what area, he will be rewarded with understanding and skills that will allow him to rise above a certain average level.

Using the methodology for working with scientific articles proposed by Andrew Eun, I plan to read at least four scientific articles a month. Read them until I understand. To be honest, it took me a week and a half to read and understand the above article about LeNet. But the more I read, the faster and better I will be able to do it. This applies not only to me.

Andrew says that he constantly carries with him a stack of articles that he planned to read. Andrew is a famous person in the field of machine learning.I think that someone who takes over his habits and methods of learning new things can be a big plus.

How do you read scientific articles?

ITKarma picture.