At least from the 1st few comments, it seems that people want to point out the ploy to get you using MS' products/tools rather than give you a wide, general education.
So... I don't mind this. Learning via a toolset is an OK approach, in a lot of cases. I personally like it. Realistically, a big part of basic competency is familiarity with a functional toolset anyway. Some people will expand their toolsets later.
Also, if MS succeed in providing good learning tools as a way of creating demand for their services, this is cool too. Ideally AWS and others will continue to do the same, and there'll be an incentive to create and maintain free courses.
If you want a more general/abstract course, these exist too. This just isn't it.
Based on a quick glance I'd say the course wants to get you to a point where you connect to data, do a little cleanup coding, use it to build a model & then turn it all into a web service. It probably has a little bit of stats & ML "theory," but not a lot. This is useful if its useful to you. If you're mostly interested in theory, this just isn't for you. Try the big coursera course, maybe.
Speaking as a CS educator, I agree, with a caveat:
In my free time, I took the big Coursera course, and it is (was?) also tied into a proprietary toolset, GraphLab. They have "you can also do this in pandas" instructions on lab assignments, but the instruction and tools are aligned to GraphLab. I imagine the vast majority of students will not even consider using pandas.
Graphlab Create has since been open-sourced (as Turi Create) by its new owner, Apple. I didn't have a problem with the setup back then.
Stressing over toolsets is like stressing over your first programming language. It's counterproductive. You will do a lot better by finding a source of instruction you trust and a toolset you can work with, and then actually making something.
Quality instruction vastly outweighs the tradeoffs with learning a new language/api/etc.
From having skimmed the Microsoft course's offering, it looks like it would prepare a good student to understand and apply the algorithms to make models, but not to implement the algorithms themselves (which the Coursera course does do).
Which is fine! I could tell you how to implement elliptic curve cryptography and prove most of the mathematical underpinnings, but I have no intention of ever writing an implementation myself.
fast.ai’s course looks great, and they even have a numerical linear algebra course.
For a theoretical background, I particularly like the Shai-Shalev Schwartz book, though it’s not about deep learning. I’m told the Goodfellow book is good only in the sense that a better book simply doesn’t exist.
If you are looking for profound theoretical insights in Deep Learning, you will be disappointed, but amazed with classical ML. If you are looking for practical capabilities, you'd be amazed by Deep Learning but frustrated with classical ML. Deep Learning is fairly trivial math-wise (unless you look carefully into non-linear optimization it uses, which is as brutal as any other advanced part of math).
Agreed. I have used some of the Azure ML services a bit and its really easy to understand concepts and visualize models via a designer, like a workflow engine. Pretty neat. I havent used it for anything professional yet though.
So I just signed up, and I see that this is implemented as a track, meaning there are several edX courses required to complete it. Among those are Python for Data Science, Math for Artificial Intelligence, legal stuff, NLP etc.
I like how comprehensive this is. The problem (as I've stated before) with Ng's course is he hand waves a lot of the required math away which concerns me. Specifically vectorization I remember constantly being surprised what was possible mathematically instead of 'getting it'.
That said, is there an AWS version of this track? Maybe Math/Spark/Scala/ML oriented?
I do like that it seems comprehensive, so I think I'll invest time into doing it.
consider a mathematician, who already knows python, is he forced to first complete the python, math etc tracks, or can he simply follow the course without needing to prove he meets the requirements?
You can skip straight to the assessments without watching any videos or doing any labs, if that's what you mean. But you need to complete the assessments if you want to get a certificate at the end.
Their strategy is to get you to use the their tools rather than teaching you.
Well you have to use some tool, and to teach a course everyone has to use a common tool, so what would you suggest? Your own favourite editor on your own favourite Linux distro I suppose? What if other students favourites are different? The class will spend its whole time getting their environments set up..
FWIW I’ve taken some of MS EdX courses and they use Python, Jupyter, R, RStudio... loads of open source stuff. See https://notebooks.azure.com for the typical environment they use
The ones I've taken did teach but were both extremely basic and much more focussed on teaching the tool than the application domain. And, yeah, they're pretty promotional.
Then again, that's very similar to many of the expensive Microsoft-certified trainings I've seen, so honestly it's still an improvement.
does anyone know if the courses are actually free (as in free beer)? Universities have been making training courses "available to the public" for ages...
there the section "How much will this cost" reveals:
"You must purchase Verified Certificates from edX.org as proof of the successful completion of courses in the Microsoft Professional Program. The cost for Verified Certificates varies by course. The prices are published in the course information on the edX site. You may always audit the course for free on edX.org, but "audit mode" does not provide MPP credit. Pricing is subject to change. Additional charges may apply."
this is clickbait advertising...
EDIT 2:
Oh its not that bad:
"Can I take the courses in the track and not sign up for the Verified Certificates of completion from edX.org?
You are welcome to audit any online course for free. However, the only way to receive completion credit toward an MPP learning path is to obtain a Verified Certificate from edX.org."
I read some of the other commenters who claimed that the tutorials were more like teaching how to use Microsoft tools.
At this point I started visualizing "tutorials" on how to open and close their application, describing the GUI layout, the workbenches, ... which is not hard for me to imagine about microsoft. I simply did not feel like checking for myself, since I would risk wasting my time. Time better spent reading actual articles...
This looks really cool, and kudos to them making it free. However, this seems very much applied, and I'd really like a course that dives more into the math and theory of AI, neural networks, deep learning, etc. etc. Anyone have any recommendations for any video lecture series, MOOCs, or anything like that?
However, if you know the very basics of matrices (multiplication, transpose) and calculus (derivatives of basic functions, and partial derivates, and chain rule) I'd highly recommend first trying basic applied ML before diving deep into the math.
It'll help you see where the math you're learning is actually used, as you learn them.
Try deeplearning.ai first, then try this "math for ML" course.
The importance of math in understanding AI is kind of overblown. Binary Neural Networks for example are in theory worse, but the savings in computational time makes up for quite a bit.
That said, a wide foundation including a deep understanding of linear algebra is more useful than just covering the specifics relevant to AI.
That makes sense. But for me for instance, who's trying to learn about AI and have a math background, I like courses that explain the math behind everything because it makes it easier to understand the concepts. It's also a good way to check that I understand what's going on. And finally, I find it easier to skim over a text has equations, they easier to parse for me since I'm a slow prose reader. So I can go through a text by skimming through the math, and when I stop understanding, I try to backtrack what's going on and only move on once I understand what was blocking me.
EDIT: And this applies to everything. Ex: I've been playing the guitar for over a decade but have been struggling with music theory. I've recently tried to apply the little algebra I know to it to try and find structure. I've found some really cool articles, for instance [0], and it's helped quite a bit believe it or not !! For French speakers that like music, math, and algorithms, I highly encourage a presentation by Moreno Andreatta [1] where he closes the presentation by performing a "rotation around the Do of the Beatles' Hey Jude".
Recent report from my mother, who's a physicist with a strong math background, no statistics: Take the coursera intro course.
She said that the first hour or two were a little frustrating for the kinds of reasons you mention. After that, she could skip the "long" explanations & examples by going to the math directly. You'll progress 5X faster than normal after that.
She "got it" by creating ML "proofs" for math problems without doing math. She didn't like all the image recognition examples. Said it's a confusing place to start.
You are probably stuck with learning from a source that assumes you don't know math, because most people (including me) don't. This course is probably not for you if you want theory, as it's focused on "using" ML algorithms, not writing them. That said, this might be useful to you if you already have stuff that you want to do, and need a way of doing the "sticks and duct tape."
It looks like the courses use cognitive toolkit aka CNTK for deep learning. This is sort of the equivalent of watching a class from Google where they use tensorflow or fb where they use pytorch.
There are several courses in that "class" (over the course of a year) that require access to the Azure cloud. The free Azure trial runs only 30 days, though.
is it just me or these are really poorly designed courses? I took the RL course and it first started describing papers in RL with little context (under applications of RL). The course then jumped to bandits and I was hoping it will describe the foundations but it assumed students know about the regret minimization framework. Researchers are traditionally not the best educators -- I usually pick tutorials in conferences that are done by professors as opposed to researchers despite their content.
You jumped straight in at the 8th course in the programme and are complaining that it assumes you know something?
So many people on this thread looking for any nitpick they can, we get it guys, you just hate Microsoft. This isn't for you, then, no-one's pointing a gun at your head and saying, take this free course.
I went through all of them and I did not see any regret minimization background. If you think there is, please point me to it and clear my misunderstanding.
Enrolled, but I doubt I will learn that much more than just using their cognitive services which seems cool and all, but that is not really doing AI yourself. Is it?
Cognitive Services may or may not usher it in, but the inevitable democratization of AI means the definition of "doing AI" will eventually be the same as "doing Excel". i.e., it is just a means to an end, and the underlying level of automation is irrelevant.
Following the fiasco with Tay, I think it's better if 4chan were to teach this course.
EDIT: the more I think about this, the more I like the idea of a course about an adversarial approach to AI. They could name it something like "Abusing public machine learning services for fun and profit". And given the previous successful work on Katawa Shoujo, I really believe that 4chan could pull this off.
So... I don't mind this. Learning via a toolset is an OK approach, in a lot of cases. I personally like it. Realistically, a big part of basic competency is familiarity with a functional toolset anyway. Some people will expand their toolsets later.
Also, if MS succeed in providing good learning tools as a way of creating demand for their services, this is cool too. Ideally AWS and others will continue to do the same, and there'll be an incentive to create and maintain free courses.
If you want a more general/abstract course, these exist too. This just isn't it.
Based on a quick glance I'd say the course wants to get you to a point where you connect to data, do a little cleanup coding, use it to build a model & then turn it all into a web service. It probably has a little bit of stats & ML "theory," but not a lot. This is useful if its useful to you. If you're mostly interested in theory, this just isn't for you. Try the big coursera course, maybe.