I am a computer science student with about one semester left before graduation. Currently, I have some free time — about a month before the final semester begins.

I am interested in artificial intelligence (AI) and have taken classes / projects in machine learning. I am currently using this time to read / revise machine learning theory from Christopher M. Bishop's Pattern Recognition and Machine Learning [PDF, 18 MB]. I am doing this purely because I am interested in the mathematics and statistics behind machine learning.

I feel like going so in-depth into machine learning theory is only good for academia (research). I do not see how such theory will be applied in the future should I take a job on developing AI systems which is more end to end machine learning.

  • 6
    If you are asking if reading ML theory is good for working in industry, it seems to me you are on the wrong site.
    – Nobody
    Dec 3 '20 at 6:53
  • Well, if you take an implementation job it may not be too useful. But for an industry R&D job a thorough understanding of the theoretical concepts is beneficial, if not required. (As you may have noticed in the book, Chris Bishop also works in industry, so clearly what we writes about is relevant there.)
    – cheersmate
    Dec 3 '20 at 7:09
  • 1
    I would think it would be much more helpful to work on kaggle projects then for such "implementation" jobs ?
    – calveeen
    Dec 3 '20 at 7:15
  • @calveeen Don't worry, remaining will take less time.
    – hanugm
    Dec 3 '20 at 8:46
  • 2
    It's also worth pointing out that just because you understand the high-level theory doesn't mean you actually understand how to use the industry-standard machine learning packages to make working code.
    – nick012000
    Dec 3 '20 at 12:02

The more depth, the better, IMO.

You can never learn "too much" ML theory, just like one cannot learn "too much" stats, probability, linear algebra.

The theory helps one understand the bigger picture, and so the deeper and braoder, the better.

Not the answer you're looking for? Browse other questions tagged or ask your own question.