Research Papers for Beginners


  • I am going through the Deep Learning Specialization. I have finished the first two courses and will finish the rest as well.
  • I am proficient in coding.
  • I am familiar w/ the basics of calculus, linear algebra and probability having done electrical engineering courses in college and computer science courses in grad school.
  • I am taking courses in the Math for ML and Data Science Specialization to refresh my math concepts. I have finished the first two courses (though they seemed way too basic).

As I am doing these courses, I wanted to get in the habit of reading 1-2 research papers every week. However, I do not want to start w/ papers that are way too specialized or deal w/ a narrow field or are experimental or have advanced math or a poorly written for the general audience.

I prefer well-written seminal/foundational/mainstream/popular papers (even if they are old) that deal w/ the general concepts and the most popular/well-established algorithms. I prefer papers with some math and that have links to any code/notebooks.

PLEASE can you recommend me a reading list. A list of 10 papers should suffice. I will REALLY appreciate it!

The number of research papers that meet your criteria is extremely small. No one writes research papers on basic concepts.

1 Like

I think you may have misunderstood my point - sorry if I haven’t explained myself well.

By “basic”, I mean the initial/foundational papers that say came up w/ the concept of Gradient Descent or Neural Networks. As opposed to say some advanced and specialized tuning methods of a minor hyper-parameter. For example, if I had to learn about MapReduce I would read the popular/foundational white paper that came out of the Google team authored by Jeff Dean.

In any case, for a beginner in ML/AI who is interested in Deep Learning/Neural Nets which papers would you recommend they read first after they have done some basic coursework? There must be some selection criteria, no?

1 Like

The FAQ for the Deep Learning Specialization has a list of reading materials. You can find it here:

Hi @Nandan1,

I don’t recall we have such a list of papers as you requested, but I do think you can filter papers with the idea of how you found the paper for MapReduce.

I googled “most cited machine learning papers”, and found this , for example. Scrolling through the list therein, technique-oriented and DLS-covered papers include Adam, BatchNorm, Dropout, Resnet, and so on.

They are well cited, covered by DLS lectures, and popular so that you can easily find other explanations online.

You can decide whether they are right for your level, taking into account other explanations available to you.

If you need more papers, you may google for more similar lists with different keywords (many people like to share their lists on their Github besides in articles), simply scan through the DLS’s tables of content and sort a list of skills in which you are interested and find their papers directly, or use Google Scholar.

In fact, Andrew might have cited some papers in those videos that you are interested in.

If you are interested in a particular subfield, add “literature review” to your keywords.

If you want to help future learners who may have the same request as yours, I encourage you to share your list to the DLS Resources category.

Enjoy your research process!


1 Like

Thank you Raymond for such a detailed and thoughtful answer!

Hey nandan, im sorry for out of topic but can you share about your leaning experiences in mathematic especially calculus? because i dont know how to understand it xD