We do this as part of our university practical course as well (2 week challenges in teams of 4, sometimes participating also in external competitions). While I agree that this good for learning, I would, however, say it is completary to lectures. It makes no sense if people just learn to overfit some hyperparams to beat a score. So rating should at least be done wisely. Otherwise I think people need lectures to understand the real limits of tech and actually find novel uses.
With many good lecture contents online I think it also would be good to also switch to more live competitions at university to do inverted classroom style learning. I think it is the mix that matters (along with the quality)
Every time I read a slogan like "learn X in Y weeks" or "learn through games instead of lectures", I have to think of a Peter Norvig piece: Teach Yourself Programming in Ten Years.
Honestly instead of taking paid courses that promise to teach incredibly difficult subjects in weeks, just take one of the longer courses like Andrew Ng's that start with the fundamentals and are free. Bootcamp style education is awful.
I'm really sad our culture is encouraging this. I know this is an unpopular opinion and may be seen as gatekeeping by some, but I believe that these "learn X in Y weeks" for deeply technical topics (it's totally fine for X framework) do more harm than good. People who rely on these usually have no real understanding, and just slap something on their resume. As someone involved in hiring, I actively treat anything like this on a resume as a big red flag. If you say you know ML but list some short-term course, I know that you're exaggerating and likely don't have a real understanding. This also tells me that you are likely exaggerating in other parts of your resume and profile too [0]. The same goes for projects. If your "projects" consist of simply calling X library with new data and doing nothing novel, that's a red flag more than anything else. I see this a ton: Just fork some Github project and make a few tiny changes and put it on your resume.
On the other hand, I respect people who are willing to spend the time and effort to truly learn something from scratch. Start with a basic math and ML course, then go on more difficult ones, read papers, implement papers and projects, and so on. The same goes for other topics such as distributed systems, compilers and programming languages, etc. There are no "learn X in Y weeks" shortcuts. I understand that not everyone may be able to do this and many will give up along the way, and I don't think that's a bad thing. It's a great filter.
The course is "Introduction to Reinforcement Learning". We make no claims that you'll be writing papers come the end of 4 weeks! :)
Really interesting reading that Norvig's piece. I agree with almost all of it and think what we're doing at Delta Academy lines up with most of it!
Particularly:
> The key is deliberative practice: not just doing it again and again, but challenging yourself with a task that is just beyond your current ability, trying it, analyzing your performance while and after doing it, and correcting any mistakes. Then repeat. And repeat again.
This is exactly what we do - providing weekly challenges that stretch your ability (that also happen to be fun), discussing the approaches taken by teams & giving expert feedback on code.
And his recipe for programming success:
> Get interested in programming, and do some because it is fun. Make sure that it keeps being enough fun so that you will be willing to put in your ten years/10,000 hours.
> Program. The best kind of learning is learning by doing.
> Talk with other programmers; read other programs. This is more important than any book or training course.
> Work on projects with other programmers.
Again - team projects which are fun sound right up his street.
There are so many high quality, free resources online for learning RL [1,2,...]. Four weeks of primarily self-guided M-F study isn't nearly enough time to obtain anything more than a cursory understanding of the topic. Kaggle, various gym environment baselines, and workshop competitions exist for those who want to compete.
Everytime I read something like "learn ML" I think about learning some descendant of the ML programming language like OCaml, F# or Standard ML instead of machine learning.
It's more a cohort-based online class that's punctuated by competitions (once per week). The competitions serve to motivate you to learn in a fun finale each week, rather than all being about winning. :)
The price tag includes 12 tutorials with exercises, 4 competitions (incl live discussion of solutions with the cohort) and expert code review from instructors on all the exercises.
Kaggle has big competitions over several months. They're designed to find top machine learning talent & innovative solutions to the problems they set. Typically the winners aren't just learning ML, they're seasoned pros doing it as a side project.
Our competitions are designed to teach. Each is a progression in difficulty over the previous one. Also there are a set of tutorials preceding each competition which get you up to speed on what you'll learn in the competition.
Plus we're organised into a cohort so you're not competing with the whole world - rather you're competing with your peers who are also learning. You work in a pair on the competition. Then we discuss the solutions teams came up with & what an 'ideal' solution would look like (if one exists - sometimes it doesn't!).
With many good lecture contents online I think it also would be good to also switch to more live competitions at university to do inverted classroom style learning. I think it is the mix that matters (along with the quality)