I am considering taking Udacity's machine learning nanodegree [0] with zero machine learning background. It seems interesting. Any thoughts?
[0] https://www.udacity.com/course/machine-learning-engineer-nan...
Understanding Machine Learning: From Theory to Algorithms (2014)
21–30 of 58 posts
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#22Top comment in another thread whines that nobody understands what the machinery of ML algorithms are really doing.
Top comment here whines that math is hard.
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#23Earlier quoted context omitted.
Andrew Ng's Coursera ML course is supposed to be pretty accessible. I've also heard good things about Machine Learning for Hackers ( http://www.amazon.com/Machine-Learning-Hackers-Drew-Conway/d... ). Ultimately, ML is a mathematical discipline. You can ask for a gentle approach that gets you to the foot of the mountain, but "if you want to learn about nature, to appreciate nature, it is necessary to understand the la…
I'm currently taking Andrew Ng's Coursera course and I'd agree it's quite accessible. In fact, if you have a solid understanding of calculus and linear algebra, you might find it a bit slow at times.
This was my problem; it was incredibly boring (at least for the first few classes) and the video-based nature of it meant that I had to skip around and try to pick up what was going on later, which seldom worked. I hear it's more self-paced now, so that might be a better option these days.
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#24I feel like the barrier to machine learning for me, as I've seen in many tutorials and books and is an immediate discouragement, is the massive amount of math thrown in your face. Many of us didn't just graduate, need glasses and fall asleep at 8pm on the couch when the kids are in bed... Math is this distant fragment of memory buried under years of everything not Math. It feels like machine learning is only taught b…
I have seen many people make this observation. I'm just curious about your problem. Is your problem
1 : Mathematics in ML - proofs etc.
2 : Understanding the intuition behind ML algorithms without the requirement of higher order math?
Personally, I feel that 2 can be tackled quite easily. The core issue is that most people who teach want to stay on a "higher dimension". ;)
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#25Just a curiosity: One of the authors also proposed Pegasos SVM [1] which is a nice online approximation to SVM and that can be written in 15 lines of code or so.
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#26I feel like the barrier to machine learning for me, as I've seen in many tutorials and books and is an immediate discouragement, is the massive amount of math thrown in your face. Many of us didn't just graduate, need glasses and fall asleep at 8pm on the couch when the kids are in bed... Math is this distant fragment of memory buried under years of everything not Math. It feels like machine learning is only taught b…
I interview a lot of developers for ML positions at our company. The first red flag is always a lack of math. Candidates who come in with API-level competence ie. can implement an ML algo using this,that or the other API, without any understanding of some basic math behind it, always fare poorly. Atleast in ML, not having an understanding of math is pretty much like claiming expertise in riding a bicycle by watching 100s of bicycle videos on youtube without ever riding a bicycle yourself. This isn't the case with say, front-end web dev, where nobody really cares whether what's under the hood is php or jquery or elegantly handcrafted elm as long as the webpage looks good and functions as advertised. You can get a lot of mileage by hiring someone who can "make webpages" without knowing what tools they use to make that webpage. Once you get traction etc. you'll probably rewrite all of the cruft :) With ML, if you hire somebody who can "create decision tree in python mllib" without knowing the first thing about what entropy is & how to compute entropy ( a real candidate, unfortunately ), you are simply inflicting a lot of pain on yourself and your customers. Suppose such a person is deciding whether or not to give you a loan, and he decides to construct a decision tree. He'll happily throw in your zip code and credit card number into the mix, not realizing that those two features have super high entropy but the tree will have serious overfitting issues ie. the tree will simply not generalize to unseen data. He won't realize these things because he won't know what entropy is in the first place, since he only thinks of decision trees in terms of some black box that comes out of some ML api.
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#27I feel like the barrier to machine learning for me, as I've seen in many tutorials and books and is an immediate discouragement, is the massive amount of math thrown in your face. Many of us didn't just graduate, need glasses and fall asleep at 8pm on the couch when the kids are in bed... Math is this distant fragment of memory buried under years of everything not Math. It feels like machine learning is only taught b…
I interview a lot of developers for ML positions at our company. The first red flag is always a lack of math. Candidates who come in with API-level competence ie. can implement an ML algo using this,that or the other API, without any understanding of some basic math behind it, always fare poorly. Atleast in ML, not having an understanding of math is pretty much like claiming expertise in riding a bicycle by watching…
I get that math is an important aspect, but being able to realise that throwing in those variables would create entropy sounds like more of a common-sense thing, no?
I guess I should still brush up on math though, it seems.
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#28I feel like the barrier to machine learning for me, as I've seen in many tutorials and books and is an immediate discouragement, is the massive amount of math thrown in your face. Many of us didn't just graduate, need glasses and fall asleep at 8pm on the couch when the kids are in bed... Math is this distant fragment of memory buried under years of everything not Math. It feels like machine learning is only taught b…
As others have said, you need to at least get the gist of what's going on mathematically to be able to make sensible decisions about model selection, structure, etc.
That said, I find a lot of the introductions to the theory behind ML techniques to be very poorly written. It's often worth giving a new student a conceptual simplification before introducing a rigorous definition.
Without linear algebra and basic probability/calculus though, forget it. Luckily there's great sources to brush up on it.
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#29I feel like the barrier to machine learning for me, as I've seen in many tutorials and books and is an immediate discouragement, is the massive amount of math thrown in your face. Many of us didn't just graduate, need glasses and fall asleep at 8pm on the couch when the kids are in bed... Math is this distant fragment of memory buried under years of everything not Math. It feels like machine learning is only taught b…
I interview a lot of developers for ML positions at our company. The first red flag is always a lack of math. Candidates who come in with API-level competence ie. can implement an ML algo using this,that or the other API, without any understanding of some basic math behind it, always fare poorly. Atleast in ML, not having an understanding of math is pretty much like claiming expertise in riding a bicycle by watching…
A lack of mathematical intuition is a serious problem for many people from engineering to biology to economics. It certainly plagued me throughout my engineering bachelors studies and is something I continually work to get better at.
In my opinion, physics students learn the best framework for thinking and get a very good mathematical intuition. For example, here's a problem from an introductory QM book that really threw me for a loop when I was studying:
A needle of length L is dropped at random onto a sheet of paper ruled with parallel lines a distance L apart. What is the probability that the needle will cross a line?
Re: Understanding Machine Learning: From Theory to Algorithms (2014)
#30Earlier quoted context omitted.
I interview a lot of developers for ML positions at our company. The first red flag is always a lack of math. Candidates who come in with API-level competence ie. can implement an ML algo using this,that or the other API, without any understanding of some basic math behind it, always fare poorly. Atleast in ML, not having an understanding of math is pretty much like claiming expertise in riding a bicycle by watching…
A lack of mathematical intuition is a serious problem for many people from engineering to biology to economics. It certainly plagued me throughout my engineering bachelors studies and is something I continually work to get better at. In my opinion, physics students learn the best framework for thinking and get a very good mathematical intuition. For example, here's a problem from an introductory QM book that really t…
> A needle of length L is dropped at random onto a sheet of paper ruled with parallel lines a distance L apart. What is the probability that the needle will cross a line?
Thickness of line is needed right? Otherwise P approaches 100% as thickness approaches 0?