-1
$\begingroup$

I have done Andrew Ng's ML and DL courses, and some projects and implemented some important ML algorithms from scratch. Now reading the deep learning book. <=(Edited)

I want to start from the beginning (in terms of reading research papers), i.e, deep feedforward networks, regularization techniques,{then maybe conv nets and others}etc, etc and some tips on how to tackle the difficulty in understanding it. Thank You.

$\endgroup$
4
  • 3
    $\begingroup$ Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. $\endgroup$
    – Community Bot
    Oct 21, 2022 at 18:21
  • $\begingroup$ For your question, I strongly suggest you start from this book: deeplearningbook.org instead of papers. $\endgroup$ Oct 21, 2022 at 20:00
  • 1
    $\begingroup$ The best research paper for you to read will depend on your specific interests and background knowledge in deep learning. However, a good place to start might be the CS230 Deep Learning, which covers a wide range of topics in deep learning and provides links to further reading on each topic.[cs230.stanford.edu/] $\endgroup$
    – Faizy
    Oct 21, 2022 at 22:11
  • 1
    $\begingroup$ @Faizy cs230.stanford.edu $\endgroup$ Oct 22, 2022 at 20:31

4 Answers 4

1
$\begingroup$

At your stage, I don't think jumping straight into reading research papers would be efficient. Generally, reading textbooks/review-articles, or simply watch a couple introductory youtube courses would do a better job at getting you up to speed with the background knowledge. Of course, you can always find a project that interests you and try to incorporate some elements of ML into it, which allows you to naturally learn ML at the same time.

Some standard introductory textbooks/courses are:

which should cover the topics you mentioned.

If you want to focus on a specific topic (e.g. ConvNets, transformers, recurrent networks, etc.), it's generally helpful to find a recent review article on this topic and read through it. This is just to understand the current state of the field, and you can then read specific papers that interests you with this contextual knowledge in mind. Note these fields are moving so fast that certain seminal papers are no longer hugely relevant (e.g. many network architectures and training methods proposed in the classic AlexNet paper are outdated.)

$\endgroup$
2
  • $\begingroup$ I have done Andrew Ng's ML and DL courses, and some projects and implemented some important ML algorithms from scratch. Now actually reading the deep learning book. $\endgroup$
    – vivian.ai
    Oct 23, 2022 at 3:17
  • $\begingroup$ Also, wiki is free XD. $\endgroup$ Apr 19 at 11:09
1
$\begingroup$

There is no "beginning" with research papers. Papers are published as they are ready, in no particular order with respect to complexity or topic. I think you just have to jump in.

Pick papers that match your interests. Look things up as you read to understand. You might need to brush up on Math.

Here is a list of resources where you can find research papers to start. I am sure you can find others as you learn what interests you.

$\endgroup$
0
$\begingroup$

The good news is that there are many freely available educational resources online and at your local library. Here are some that I used to get me started:

  1. Kaggle: Intro to Deep Learning
  2. Kaggle: Computer Vision
  3. Pattern Recognition & Machine Learning (Chapter 5)
  4. Machine Learning (Chapter 4)

Good luck and happy learning!

P.S. I know these are not research papers, but I would encourage you to start with these anyway.

$\endgroup$
1
  • $\begingroup$ Thanks, some of the links were helpful, but I want an answer specific to my question. $\endgroup$
    – vivian.ai
    Oct 21, 2022 at 15:28
0
$\begingroup$

Not sure what to recommend, since you say "from the beginning" in the text but "intermediate" in the title...

Anyway, for the "then maybe conv nets" part, there is a tutorial from 2021 that relates convolutional networks with the matched filter, a well grounded technique in signal processing. I find this a great idea; depending on your background, it may be interesting for you as well.

https://arxiv.org/abs/2108.11663

Is this the kind of papers you are looking for?

$\endgroup$
1
  • $\begingroup$ In terms of reading research papers, I meant the beginning. $\endgroup$
    – vivian.ai
    Oct 23, 2022 at 3:11

Not the answer you're looking for? Browse other questions tagged .