13 ms·

For people getting started with ML do you think it is more important to learn first principles and the "boring" math like this, or do you think it is important

For people getting started with ML do you think it is more important to learn first principles and the "boring" math like this, or do you think it is important to give the learner some quick wins and keep the excitement and interest levels up?

closed 10y ago
This feels like a pretty loaded question. It seems like you can have math with quick wins, keeping excitement and interest. When you say "boring" math, are you referring to the overall content or the way it's taught? Most of my experiences with "boring" math was because it felt taught poorly or I wasn't ready for it. ML is such a broad canopy that it probably includes many who aren't ready for the math, and will find it boring. It's the same with the distinction between appliers and "methodologists" in statistics. Breaking down "people getting started with ML" into what they want to do with it feels more tractable. Maybe it's an issue of courses signaling who they are geared for.
RockyMcNuts 10y ago
Do what feels good for you :) Ng is a fine place to start, you get some pretty quick wins, doing MNIST from first principles within a month or two. You just need to know or get comfortable with matrix multiplication. It strikes a reasonable balance between being rigorous and approachable for a committed student at an undergrad level. Principles of Statistical Learning is easier https://lagunita.stanford.edu/courses/HumanitiesandScience/StatLearning/Winter2015/about https://lagunita.stanford.edu/courses/HumanitiesandScience/S... LAFF linear algebra is just starting http://www.ulaff.net/ http://www.ulaff.net/ Hinton's Neural Networks is offered in the fall https://www.coursera.org/learn/neural-networks https://www.coursera.org/learn/neural-networks For my money, I wouldn't do something like Practical Machine Learning in R, because I think you'll learn more R than machine learning. I wouldn't do the Udacity TensorFlow course because I think it assumes a lot of stuff you would learn in Ng's class ... I think Ng is a fine place to start.

closed 10y ago

This feels like a pretty loaded question. It seems like you can have math with quick wins, keeping excitement and interest. When you say "boring" math, are you referring to the overall content or the way it's taught? Most of my experiences with "boring" math was because it felt taught poorly or I wasn't ready for it. ML is such a broad canopy that it probably includes many who aren't ready for the math, and will find it boring. It's the same with the distinction between appliers and "methodologists" in statistics. Breaking down "people getting started with ML" into what they want to do with it feels more tractable. Maybe it's an issue of courses signaling who they are geared for.

RockyMcNuts 10y ago

Do what feels good for you :) Ng is a fine place to start, you get some pretty quick wins, doing MNIST from first principles within a month or two. You just need to know or get comfortable with matrix multiplication. It strikes a reasonable balance between being rigorous and approachable for a committed student at an undergrad level. Principles of Statistical Learning is easier https://lagunita.stanford.edu/courses/HumanitiesandScience/StatLearning/Winter2015/about https://lagunita.stanford.edu/courses/HumanitiesandScience/S... LAFF linear algebra is just starting http://www.ulaff.net/ http://www.ulaff.net/ Hinton's Neural Networks is offered in the fall https://www.coursera.org/learn/neural-networks https://www.coursera.org/learn/neural-networks For my money, I wouldn't do something like Practical Machine Learning in R, because I think you'll learn more R than machine learning. I wouldn't do the Udacity TensorFlow course because I think it assumes a lot of stuff you would learn in Ng's class ... I think Ng is a fine place to start.