Getting Started with ML and DL

Hey everyone!! My name is Kunal and I am from an electronics background. I want to learn ML and DL. I read many great reviews about fast.ai on twitter and was eager to know more.
I have a few questions that I’d like to ask:

  1. I have some knowledge of python and basic mathematics. Which course should I begin? Which textbook should I refer? There tons of courses out there but I want to stick with one.

  2. I was quite fascinated by the blog post on why everyone should blog. I want to host a similar a blog post that describes my learning journey, but along with it, I want to add an about me page that would serve as a resume, like integrate linkedin, github and twitter links. How do I it?

  3. If anyone is beginner and would like to study together, do let me know <3.

Since, this is my first post, I do not know the exact category where I should ask these questions. I would appreciate if someone replies to this and help me out :slight_smile:

2 Likes

For a blog, check out https://github.com/fastai/fastpages.

For courses to start with, there’s 2 I would recommend as a starting point. You could start with 1 or the other, or both. Here’s the courses and you can decide. Ultimately, you probably want to do both eventually so you are more well rounded.

Intro to Machine Learning - This is a gentler introduction. It mostly covers Random Forests and Decision Trees. While it’s an introductory course, it’s not the kind where you learn a bunch of trivial stuff that nobody actually uses. You will be learning things that I use in my job as a data scientist all the time You will learn why to have training vs validation sets, how these algorithms work, and more importantly code and apply this to real problems. A lot of the methods and techniques in the course are really good for tabular data (ie doing machine learning on data stored in SQL tables). There’s a lot more there, but I think that’s why you’d do that course. I don’t think there’s a book that goes with it.

Practical Deep Learning for Coders - This move a bit faster as it goes, but it is still a course you can do to start with if you want. You will be doing deep learning and learning more cutting-edge stuff. The biggest reason to start here in my opinion is if you just think Neural Networks are cool. I don’t think there’s a bad decision in which you go with. A new version of the course along with a new book and a new version of the library are being released in the next month, so it may make sense to work on the intro to machine learning for the next month. Then you can start fresh on this course with the new tools and best practices. Book can be found at github.com/fastai/fastbook currently, but you should definitely buy it at https://www.amazon.com/Deep-Learning-Coders-fastai-PyTorch/dp/1492045527/ref=sr_1_2?dchild=1&keywords=deep+learning+for+coders&qid=1593261408&sr=8-2v

1 Like

Thank you so much Ezno for solving my doubts. My target for next month is to complete the intro to ML :slight_smile:
Btw, how much math would I need to do the intro to ML course?

With basic high school math, you can learn the rest as you need it. He is good about telling you if/when you need to go learn something like that. But bottom line, machine learning isn’t complicated math typically - It’s really basic math + small tricks + doing that basic math a very large number of times.

1 Like

Yes. I will be graduating next month.

I will have to revise the basics, but as said, I can do it on the go.

Thank you @Ezno and @dejavucoder :slight_smile:

If I have any doubts during the course, which channel should I look for?

1 Like

Totally agree with @dejavucoder.

In the categories area you can see all of them (https://forums.fast.ai/categories). There is a Part 1 category (which is for practical deep learning for coders course). There is also an “Intro to Machine Learning” category for that course. Just use your best judgement on where to post it. If you post it somewhere that makes some sense, it’ll be fine.

1 Like

Amazing community at fast.ai. Excited to be here and learn new things. Once again, thanks!

1 Like