So far it's pretty easy. I took this mostly as a guided tutorial in Python. I'd already read through the Python site and had a false sense of security about the language. I find that I can see the lectures and do the homework in the same day without too much difficulty. That's not to say I'm getting perfect scores as I'm not. On the other hand, it doesn't disturb me deeply to guess wrong on the Collatz Conjecture ;-)

I think this class will turn out to be useful for the purpose I expect of it.

asked 08 Mar '12, 13:51

robrambusch's gravatar image

robrambusch ♦
24.2k12101350


I think this is partially because they have chosen 7 week length for their classes, while Coursera and MITx classes are much longer (even #daa, because it's just split in half). And I think that they chose this kind of course length by looking at retention rates from AI-class. It's hard for people with real life work and families to put aside so much time for many weeks and keep going. And since their classes are shorter, they have to give up something (more involved math), if they want to still cover the whole subject. I would love to see a harder follow-up, maybe they could split it up in 2 parts, like #daa - like, CS373 PRC - I (7 weeks) and then CS374 PRC - II (7 weeks) with more math, harder programming assignments, and a PRC - I as a recommended prerequisite. There is now a question by Andy @ PRC forum, - homework-feedback, where they mention that they want to have more challenging content.

Every single person who took AI-class had the opportunity to learn the same, and more, from the AIMA book. As for free - hey, you can find it in torrents, if you are very poor, but determined. How many did ? I bet not that many, compared with the amount of people who watched the whole course, and not many even compared to how many got the certificate. But even the people who did not finish, they got an insight to a field, as @rseiter said - it was a great popularization of AI and CS in general. I think this is a lot about educating masses about what is out there, what are the opportunities, what are they missing, where can they go and what can they learn.

And that is exactly what CS101 is doing - it is popularizing Computer Science basics to a population that has had no exposure to it before, but are curious. It's easy ? Of course it is! We here are all AI-class alumni (I assume), most of us here are professional programmers or systems engineers, many have CS degrees. Obviously a course that is targeted to people who don't even know what a "string" or "bit" is, is very easy for a person who has learned "around 10 programming languages" in his life. Because we are not a target audience for this class. I am taking it mostly to see how things are being taught, in what order CS concepts are introduced etc and I find it very interesting. I was taught BASIC at school and university, and it wasn't very exciting.

link

answered 10 Mar '12, 02:42

Gundega's gravatar image

Gundega
7.5k11346

I think the Udacity courses are a bit too easy. This applies to CS373 specially, if you meet the prerequisites of course, since unlike CS101 it's not introductory but supposedly advanced. Yet it stops short of introducing "too much" algebra, the programming assignments are small, etc. It gets you started in robotics, but if you wanted to make something a thousandth as complex as what Thrun has made, you'd need to complement it with lots of reading and lots of math.

Currently Coursera and MITx seem to be closer to offering the same material as in brick and mortar.

I think it's too bad, because the whole point of the revolution is that it's higher education. I don't want it to be harder only to have less people complete it; it's not knowledge or even teaching what's in short demand or expensive, but rather, as Thrun knows perfectly well, the kind of education that enables people for the job market. People who are in the end really interested in learning robotics are more than willing and able to learn the prerequisites beforehand. And needless to say, people in underdeveloped countries don't need the material to be dumbed down, on the contrary they need the empowerment that free higher education will give them.

link

answered 09 Mar '12, 12:43

XavierP's gravatar image

XavierP
1.9k1521

edited 09 Mar '12, 12:45

@XavierP I agree with your assessment, but would phrase it as "a bit easier than I would like". One thing I liked about the AI class was that between the optional textbook and Pacman programming assignments (which were discussed on the forums) it was possible to adjust the experience myself and retain the forum community because others were doing the same. I am looking at the Udacity classes more as popularizers than as college/graduate level education and I think they fulfill that role well (I am enjoying CS373 and am a bit glad it's not taking more time given all the other classes coming). The Coursera classes look like they will be structured more like "real college classes" and come closer to meeting your (our) desires.
(09 Mar '12, 12:58) rseiter ♦ rseiter's gravatar image

Udacity may be one of the the finest learning places in the Internet, but people have been able to learn from the Internet since it began to exist, and I don't think Udacity's goal is simply to teach. Sebastian has stated several times that his goal is to revolutionize higher education. If their courses aren't comparable to brick and mortar college courses, there's no revolution. The Khan Academy likewise, is very nice to learn but has nothing to do with certifying this knowledge.

You can spend your whole life learning from the Internet and books, but if you don't have any kind of certificate or degree, employers won't want to bother to find out whether you really know anything.

If Udacity builds a reputation of being as demanding as brick-and-mortar universities, employers will be interested in people who took their courses, regardless of how "official" their certificates are. But if their level falls short compared to colleges, people can use the knowledge they acquired in Udacity in their daily life and work, but Udacity won't impact the education market, and it will remain a place where it costs $20,000 to get a degree, and raising.

The point is that Stanford could (can now) teach thousands of times as many students as it does, without lowering the bar a millimeter.

There's also the "weedout" tactic, but the usual criticism against it is that it's bad to be a poor teacher in order to have less students succeed; not that it's bad that the material is advanced, even if the teacher is good.

There will always be a need for higher education, where the state of every art is taught. Of course there's also need for all the other levels of education. But if higher education remains untouched by online popularization, they'll continue charging $20,000 or more for a degree. And we know it's possible to bring online popularization to higher education, without lowering its level.

link

answered 11 Mar '12, 13:45

XavierP's gravatar image

XavierP
1.9k1521

The enthousiast student can always do more. Where is the problem? It's too easy, you finish the homework fast? Fine, you get more time to dedicate to experiment with the material and do complementary readings. Or, do you just fear many students may get the same score as you and you believe you should emerge from the hord? You know, it all depends on the goals of UDacity in providing these courses. If Sebastian was enthousiast about teaching the world, if he become to multiply difficulties to retain only a bunch of 10 top performer students, he ends up doing what he was already doing at Stanford. Teach a small class of selected students.

link

answered 11 Mar '12, 12:56

AchilleTalon's gravatar image

AchilleTalon
1.7k15

I agree with you to some degree, but think you underestimate the benefit to the enthusiast student of a structured environment and associated community. The enthusiast student has always been free to seek out books, internet resources, other people, etc. to learn on their own. I consider myself a fairly enthusiastic student (and like self teaching), but I do see value in online classes. I would argue that the teaching, structure, and community is most valuable when a student is pushing their limits. This is why I would like to see the availability of more difficult classes. Audience sizes would probably decrease for more difficult classes. This does not mean they would be only 10 people though. If the more difficult classes attract an audience 1/10 that of the existing classes we are still talking about thousands of people in a class (I think the number would be between .1 and .01 of the current audiences, but that's just a guess). You said it best with this: "You know, it all depends on the goals of UDacity in providing these courses."
(11 Mar '12, 14:30) rseiter ♦ rseiter's gravatar image
1
I do not underestimate the enthusiast student, I consider myself in that category. And the playground is just there for them to show up and lead this somewhere for their own benefit and the benefit of others. However, I believe this should be done in the respect of the pace of the courses themselves. You know, once you have learn the particle filter algorithm and the Kalman filter algorithm, why would you want to introduce relativistic quantum mechanics into it just because you didn't find it hard enough? Currently, I can see a parallel community which is creating itself within the course and I believe they will do something by their own on the side of the course and it will be just fine that way too.
(11 Mar '12, 16:48) AchilleTalon AchilleTalon's gravatar image
1
Maybe my statement wasn't clear: "think you underestimate the benefit to the enthusiast student" I wasn't saying you underestimate the student, I was saying I think you underestimate the benefit that student can get from the class setting. (maybe I am the one underestimating the enthusiast student by thinking they need the class to provide structure for the more in depth material ;-) I think the parallel community you describe is a great way to handle the difficulty problem. I would like to see more support for this in some of the courses (AI had the textbook and Stanford pacman assignments, Algorithms has textbooks and optional lectures/theory problems). The Coursera courses seem better than the Udacity courses in this regard (probably from their genesis in existing university courses), but it looks like the CS373 students are managing just fine creating an enhanced environment on their own. I think we agree that a course shouldn't just be hard for no reason (e.g. a weedout course). The difficulty I have in mind is more references to more detailed explanations or derivations. Thanks for your thoughtful analysis and comments.
(11 Mar '12, 17:25) rseiter ♦ rseiter's gravatar image
Yes, we agree, I misunderstood your statement.
(11 Mar '12, 18:44) AchilleTalon AchilleTalon's gravatar image
Your answer
toggle preview

Follow this Question via Email

Once you sign in you will be able to subscribe for any updates here

Q&A Editor Basics

  • to upload an image into your question or answer hit
  • to create bulleted or numbered lists hit or
  • to add a title or header hit
  • to section your text hit
  • to make a link clickable, surround it with <a> and </a> (for example, <a>www.google.com</a>)
  • basic HTML tags are also supported (for those who know a bit of HTML)
  • To insert an EQUATION you can use LaTeX. (backslash \ has to be escaped, so in your LaTeX code you have to replace \ with \\). You can see more examples and info here