I'm taking this class now. I've finished the first two weeks of the first course and am about to begin the first set of programming assignments.
Some initial impressions:
- I really like Andrews teaching style, which is why I took the course. If you are familiar with his machine learning coursera class and enjoy it, you will enjoy this as well. It really feels like a seamless continuation of the ML course and the concepts taught there are helpful. You may want to go through that course first to learn the basics, but if your math is solid you can jump right in to this.
- the course teaches python, numpy, and tensorflow. Some folks had trouble with Octave in the ML course, so many will appreciate the stack being taught here.
- there is lots of foundational mathematics. Some like that (I do) and some don't. If you are not interested in core calculus or linear algebra details and just want to learn applied deep learning through code, you may enjoy the fast.ai courses more (which to me felt a bit cargo culty)
- it's still early in the specialization for me so take the above with a grain of salt!
Do you know of any courses that will bring one up to speed on the math component?
I really love this format of learning, and I want to take this course as it's something I'm interested in and I like Andrew Ng, but the Week 2 content was a complete non-starter for me. I've been writing software professionally for a decade now, but because I have no mathematical background I'm very far from understanding even the first step of this course (which really is Week 2, Week 1 is just a formality).
Audit edx's Calculus sequence, taught by MIT. The courses are: Calculus 1A, 1B, 1C. You can watch MIT OCW's Linear Algebra course with Strang and/or enroll in: "Linear Algebra - Foundations to Frontiers" on edx. Use Khan academy for supplemental Calculus & Linear Algebra review. You can get Stewart's Calculus text and read through it/attempt the problem sets. Once you have a solid Calculus/Linear Algebra review you can take a look at: "Statistics 110: Probability" which is found free here: https://projects.iq.harvard.edu/stat110/home.
Get familiar with linear algebra. What I did was go back to school ( community college for cal2, linear algebra, differential equations and multi-variable calculus. ) I'm comfortable with the math now. It's a commitment for about a year but really worth your time.
You may just want to try Andrews original ML course. There is some introductory calculus and linear algebra early on that is solidified through the 11 weeks, and it is all in context. I didn't come into that course cold, but it had been 18 years since I took any calculus and didn't remember much...but everything I needed was taught there. The good thing about the ML course is that if a concept is taught, you know exactly what the context is and can do some further research on the side through youtube, Kahn academy, or other resources.
If you had high school algebra, then with some rigor you could get through the ML course and walk away with a great foundation.
... and click "Enroll", you can only proceed by supplying payment info. However, if you scroll down to that page to the box titled "Course 1", at the bottom of that box is a link "You can choose to take this course only. Learn More".
Click on THAT to go to the individual course page. Then, click Enroll, and in the first box that pops up, you'll see the link "Or audit this course" in the lower left.
This allowed me to sign up for all five without supplying payment info.
The basic problem is that Coursera wasn't successful in attaching some meaningful value to their certificates as credentials. And, if the credential isn't meaningful, why on earth would I want to pay a VC-funded company for a PDF that has zero value to me? Taking the course may be worthwhile but a certificate adds essentially nothing to that.
So now they've effectively eliminated just about the only thing that distinguishes them from some YouTube videos and a textbook.
The basic problem is that Coursera wasn't successful in attaching some meaningful value to their certificates as credentials.
The EdX solution to this is that their courses are all endorsed or run by brick-and-mortar universities with considerable investments in their brand that they won't want to tarnish by attaching it to any random certificate.
I'm not sure to what degree EdX has really "solved" this. My impression is there's quite of range of quality and rigor on EdX as well. And, more centrally, most people still don't see EdX certificates as general substitutes for more conventional educational degrees.
I wish I could figure out what the actual price is. I'm logged in and all I see is "Enroll free!". The only pricing disclosed is $49/month. Courses 1-3 are 9 weeks, so does that mean it costs about $100? (No lengths on courses 4 and 5 yet.)
To get the course material, you go to each course link and click on "Enroll". Then look for the "Audit" link at the bottom left of the modal dialog that comes up.
It says it is free to audit the videos, but I can only find a button that lets me enroll for free for 7 days before paying $49 a month. Am I missing something?
EDIT: nevermind when you click enroll then there is a small text in the left bottom of the pop up thats says audit the course.
Is there any past discussions whether these courses will actually get you a job doing deep learning? I am skeptical these courses would get you a job in deep learning.
I'm pretty excited about this... at work, we are using machine learning to train models that allow us to automate the process of grading student essays. The product is called LightSide. I hope to apply my learning from the class to improving our models and our scoring accuracy. Woowhoo, thanks Dr. Ng!
I would be furious if my essay was graded by a computer. Is the model really good enough to account for all the variances in human language? Is the whole thing graded automatically or just the grammar?
I don't know about the specific project SoMisanthrope is talking about, but these types of tools are often used in conjunction with human graders. e.g. Instead of having 2 human graders, you automate the grading and have 1 human grader, and if the grades differ by some amount, only then do you bring in a secondary grader.
Good point DamVigilante. We trained the model using hundreds of human-scored essays. They were all double or triple scored, to validate IRR. I think that's why the model is performing so well. But, there is always room for improvement! :-)
I hear you WoodenChair. For this particular task, the stakes are very low (really, zero). The model agrees with our human graders, quite well. In fact, statistically speaking, the machine scoring is actually tightly correlated to the highest inter-rater reliability that we could achieve, among our human graders. You would be surprised.
