How I started with Machine Learning

It has been almost a year now when I first started learning Machine Learning. I am still a newbie…learning new things every day. This is a fairly new and exciting field with research happening really fast and not easy to catch up.

I will share how I started…and what I am doing right now.

Before joining my college, I talked to few seniors to recommend me some pre-requisite video courses as I have never done ML before in my B.E. and IITs are really competitive.

For the first semester, we are given introductory courses like Introduction to Machine Learning which are not that easy…so If you are all decided to jump to ML…then definitely go for this course in you curriculum. To start as pre-requisites, one can learn probability and linear algebra if you have not focused on these two topics before.

Courses that I started with are:

1. MIT 18.06 Linear Algebra ( https://youtu.be/J7DzL2_Na80 )
2. Statistics 110 by Harvard University ( https://youtu.be/KbB0FjPg0mw )
3. Essence of Linear Algebra – 3B1B
   (https://youtube.com/playlist?list=PLZHQObOWTQDPD3MizzM2xVFitgF8hE_ab)
4. Essence of calculus – 3B1B
   (https://youtube.com/playlist?list=PLZHQObOWTQDMsr9K-rj53DwVRMYO3t5Yr)

These are the starting videos that you can finish before joining to strengthen your concept… or learn any if you don’t already know. There will be other good videos as well on YouTube that you can search but…I followed these.

If you directly want to jump to Machine Learning then

1. Stanford CS229 – Machine Learning
   ( https://youtube.com/playlist?list=PLoROMvodv4rMiGQp3WXShtMGgzqpfVfbU)
2. Neural Networks
   (https://youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi)
3. CS231n – Computer Vision
   (https://youtube.com/playlist?list=PLC1qU-LWwrF64f4QKQT-Vg5Wr4qEE1Zxk)
4. MIT 6.S191 – Deep learning
   (https://youtu.be/-boCMDouF2g)
5. CS 224N – NLP with DL
   (https://youtube.com/playlist?list=PLoROMvodv4rOhcuXMZkNm7j3fVwBBY42z)
   (https://youtu.be/knTc-NQSjKA)
6. COL 774 Machine Learning
   (https://www.cse.iitd.ac.in/~parags/teaching/col774/)
7. Advance data structures and Algorithms
   ( https://youtube.com/playlist?list=PLXFMmlk03Dt7Q0xr1PIAriY5623cKiH7V)
   (https://youtube.com/playlist?list=PLUl4u3cNGP61Oq3tWYp6V_F-5jb5L2iHb)
   
   Some standard books that were used during my course work:
   
8. Information Retrieval – There is no online course that I followed as it was taught really good in the college itself
   Book – Introduction to information retrieval
   https://nlp.stanford.edu/IR-book/pdf/irbookonlinereading.pdf
   
9. Book – Deep Learning
   https://www.deeplearningbook.org/
   
10. Pattern Recognition and Machine Learning
    http://users.isr.ist.utl.pt/~wurmd/Livros/school/Bishop%20-%20Pattern%20Recognition%20And%20Machine%20Learning%20-%20Springer%20%202006.pdf
    
11. Speech and Language Processing
    (https://web.stanford.edu/~jurafsky/slp3/)
    

How to start with research and choose problem statement?? – A typical question
Even I am learning the answer to this one…
It has been a year now…and I still don’t have a finalized problem statement to write a paper on…maybe I am slow or this process is slow I don’t know yet. But surely its all self study, the more you read, the more you learn how other people find and solve problems.
There is no particular paper that one can start with my first paper was introduced to me in one of my coursework and from there I started reading more on my own as the course progressed and we were also asked to develop a project…which literally made me read 10 papers and finally I implemented one of them.

If someone really has an answer to this one, please comment…it will surely help many students like me.

Some things that I think can be beneficial to know are:

1. Many GitHub repos have have the link to good papers from particular fields that one can follow to read… just search for your topic on GitHub
2. One can also follow some profs or industry ppl in MSR, Google Brain, Facebook AI, or influencers as well on Twitter, they share good videos and papers.
3. There is an archive with e-prints of the papers – https://arxiv.org/
4. Suppose some paper is not available for free download. In that case, good universities generally have subscriptions to online journals and digital libraries where you can login with your institute id and get access to restricted papers.
5. Other than this there is Sci-Hub, which is a shadow library website that provides free access to millions of research papers and books, without regard to copyright …just be careful while using this…there are chances of getting your IP blocked as well.