3
$\begingroup$

I'm an artificial intelligence enthusiastic and I want to learn about it.

I want to ask you what do you think about the Udacity nanodegree Deep Learning Nanodegree Foundation.

I don't know if it is a good idea to pay for that course or maybe, there are better free resources.

I want to understand what artificial intelligence is, and also learn about machine learning, deep learning, and convolutional networks. I'm interested in image and speech recognition and also in artificial life.

$\endgroup$
  • 1
    $\begingroup$ Ideally,you won't receive any recommendations concerning reading resources;here. $\endgroup$ – quintumnia Jan 23 '17 at 8:22
10
$\begingroup$

It doesn't seem expensive at $399* (although the asterix needs to be taken into consideration.)

If you're interested in this subject, this may be a decent course, however it is certainly not an accredited institution, thus any "degree" you get from this course will be meaningless in an academic sense. On the other hand, a certificate shows that you've committed resources toward training.

My high level take on this is that the "Foundation" is looking to capitalize on the recent publicity for Deep Learning, per the AlphaGo milestones.

One thing I can tell you for certain is that this field requires advanced mathematics from a research perspective, and people who work in that area spend years training to gain the requisite skills.

The requirements for this class seem to be restricted to "Python knowledge" with no mention of mathematics, which raises some alarm bells, although it's worth mentioning that once solutions/applications arise, they can be applied to varying degrees without having to understand the mathematics. [See mindcrime's comment]

$\endgroup$
  • 2
    $\begingroup$ Thanks. Maybe there is another course that you know and you can recommend me. I did Andrew Ng's Coursera class on Machine Learning and you are right it requires advanced mathematics. $\endgroup$ – VansFannel Jan 20 '17 at 7:28
  • 2
    $\begingroup$ Depends on what you mean by "working in this field". There's no question that doing original research requires linear algebra, multi-variable calculus, etc. And there's no doubt that the more you understand the math and the theory. But at the same time, it seems clear to me that there's a distinction between "applied machine learning" and "ML research" and that you can do a lot of the former, and create a lot of value, without knowing a lot of math. $\endgroup$ – mindcrime Mar 17 '17 at 20:10
2
$\begingroup$

I studied Deep Learning as a part of my master's Degree. My suggestion would be to simply take the free online courses. There's so many amazing resources. There's really no need for you to pay for the course. If you are dedicated to learn the subject then you will. There's also a ton of books that have already been published going over the underlining workings of each of the deep learning techniques.

I don't think it is worth the money. And as Duke Zhou stated, if their is no mention of a deep understanding of mathematics as a prerequisite that is a red flag. Seems like a money grab to me.

$\endgroup$
  • $\begingroup$ I've found the research papers to be enormously helpful as well, and most can be found for free after a suitable period. $\endgroup$ – DukeZhou Mar 22 '17 at 13:57
2
$\begingroup$

I'm doing my master in computer science and i really liked the course from caltech ( complementary book is "learning from data") and the famous course from Andrew Ng which you can get free on cousera. On the programming side, i really like "Python machine learning" from Sebastian Raschka.

$\endgroup$
2
$\begingroup$

I know that I'm late in answering, and also notice that you've enrolled in the course already. But I wanted to suggest some courses available online for free for those who are in still in search.

As you've already done Andrew Ng's Machine Learning course, you can do CS231n: Convolutional Neural Networks for Visual Recognition course from Stanford taught by Dr. Fei Fei Li, Andrew Karpathy and Justin Johnson. It provides a very clear and thorough introduction to deep learning as well as on CNNs and how they are used in image recognition.

For NLP and speech processing/recognition, you can either do CS224n: Natural Language Processing with Deep Learning course from Stanford taught by Richard Socher and Chris Manning or Deep Learning for Natural Language Processing course from Oxford taught by Phil Blunson and many researchers from Deepmind. Both the courses teach NLP with a focus on the recent deep learning approaches within this domain.

All the above three courses have their class videos/notes, and assignments put up available online with decent communities of learners around them.

$\endgroup$
1
$\begingroup$

I'm a student in the Deep Learning Foundation Nanodegree. It's going great! The beginning was harder due to the math and we had to understand what is going under the hood. Only after this you use frameworks like Tensorflow.

My math is at high-school level, and the math that you must know or learn isn't much, they said that you must know basic algebra, calculus etc.

You must definitely know how to program and if you know how to program you should be able keep up with the course if you enjoy what you are doing.

The only real complaint I have is that they claimed it would cost 3-5 hours a week, but it's more like 9-15 hours.

Udacity isn't capitalizing on the A.I hype, it's trying to capitalize on the future and learning of technology.

$\endgroup$
  • $\begingroup$ The course enrollment is currently (almost) over I think. If you complete the foundation course you can enroll in the real nano degrees where you can get a job with. $\endgroup$ – Alexander Mar 16 '17 at 11:38
  • $\begingroup$ Thanks. I have enrolled it. You are right, I need a lot hours to study. $\endgroup$ – VansFannel Mar 16 '17 at 15:41

Not the answer you're looking for? Browse other questions tagged or ask your own question.