Ask HN: Recommendations for ML training 2h a week
8 comments
Just a thought: is it practical to match the company time with some of your own time? e.g. two hours of company time plus two hours of your own time. The degree to which that sounds like a good idea might be a useful way of gaging your own interest in a particular method or course of study...material that is engaging enough to pursue outside of work is likely to be a good choice for you.
A second thought is to frame the project in calendar time. Two hours a week is a lot of hours over a year.
A third thought is to frame the project in workday time. A half hour a day, every work day is a steady discipline.
Finally, I'm biased toward beginning at the beginning because I often find that skipping the early basics based on what I know means I miss the author telling me which basics the author thinks are important (and since they're the expert, perhaps I ought to value their judgment a bit more than my own in that regard).
Good luck.
A second thought is to frame the project in calendar time. Two hours a week is a lot of hours over a year.
A third thought is to frame the project in workday time. A half hour a day, every work day is a steady discipline.
Finally, I'm biased toward beginning at the beginning because I often find that skipping the early basics based on what I know means I miss the author telling me which basics the author thinks are important (and since they're the expert, perhaps I ought to value their judgment a bit more than my own in that regard).
Good luck.
I am taking Machine Learning on coursera by Andrew Ag. Initially I was intimidated by the idea of ML as I had no prior programming experience. I started learning data science stuff less than 6 months ago. I started to feel motivated and confident about ML by taking this course. I highly recommend it.
https://www.coursera.org/learn/machine-learning
https://www.coursera.org/learn/machine-learning
Thanks. I'll look that up. However, how much time does that takes you (weekly)?
It's self-paced, so it's kinda up to you. There are "deadlines" to help keep you on track, but they're optional. To pass, you just have to pass all the graded assignments by the end date. But if you're getting close to the end and are behind, you can always shift your enrollment to the following session - but you keep all your progress and everything. It's pretty cool in that regard. Very low pressure.
That's cool. Can you tell me how much time you spend on it per week in average? Just so I get a general idea...
It's been a while since I took it, but I think I spent an hour or two a week watching the videos and reading notes and whatever, and then maybe another 3-5 hours on the programming assignment. It was probably less than that on the earlier programming assignments, and more on the later ones as things got more complicated later on. And I might have spent more time on the videos on certain sections, because of re-watching sections that weren't intuitively clear to me right away. In particular, some of the math'ier stuff where he explained the stuff about using partial derivatives to calculate the error gradients for neural networks... that stuff I had to put more work into since my math background isn't real strong (I never took multi-variable calculus).
All of that said, you can get through the class and learn and understand the material at the level he teaches it, even without completely understanding partial derivatives (a point he makes in the lecture). But having a strong calculus background certainly wouldn't hurt.
All of that said, you can get through the class and learn and understand the material at the level he teaches it, even without completely understanding partial derivatives (a point he makes in the lecture). But having a strong calculus background certainly wouldn't hurt.
I spend about min an hour a day. The general timeline is about 4 hrs per lectures, reading 1 hr reading (mostly optional) and 3 hrs per assignment. I take longer time on assignment since I am new. Sometimes I have to watch lectures over and over again to understand. In general I spend about 10 hrs/week.
I don't think there is any value in paying for the certification, since its an introductory course.
I don't think there is any value in paying for the certification, since its an introductory course.
If it isn't too introductory, I'd say take the Andrew Ng course on Machine Learning on Coursera. If that's too introductory, try the Geoffrey Hinton course on Neural Networks on Coursera, or the Google / Tensorflow course on Deep Learning on, er, I think it's EdX.
Just researching on myself is not out of the question, but something a bit more structured would be better. Still, 2h a week might be a bit too little to spend following a proper AI class online.
Also, I'm not new to the subject, but I haven't really done much in it in over 10 years. Therefore, a normal introductory course might be somewhat useless in some points but required in other points. As such, something that I can choose the pace would be the most appropriate.
Any recommendations would be very much appreciated...