Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon. Entire thread

Stanford's Free AI Course

Name: Anonymous 2011-10-15 4:55

Is anybody else in the class, what are your thoughts so far? Who wants to have a thread where we can share class notes or answer questions?

Name: Anonymous 2011-10-15 5:36

Name: Anonymous 2011-10-15 5:36

Name: Anonymous 2011-10-15 5:39

Name: Anonymous 2011-10-15 5:40

Name: Anonymous 2011-10-15 6:15

wut...

Name: Anonymous 2011-10-16 19:25

bump

Name: Anonymous 2011-10-16 19:31

The loaded coin question is a mindfuck. Really, wtf. Doesn't make any sense.

Name: Anonymous 2011-10-16 21:51

>>8
What doesn't make sense about it?

Name: Anonymous 2011-10-16 22:32

>11 Nice ``DUBZ'', bro.

Name: Anonymous 2011-10-16 22:33

>>10 yainorite?

Name: Anonymous 2011-10-17 4:48

>>9
What are the sensors and the actuators? Are we observing the coin flips from distance and updating the model based on the flips? Or are we actually flipping and the probability depends on physics as well?

Name: Anonymous 2011-10-17 11:43

>>12
The former latter.

Name: Anonymous 2011-10-17 12:45

>>12
>What are the sensors and the actuators?
That's not even part of the problem. They could be anything, for instance, depending on whether we were talking about a robot doing the flipping AND making the predictions, or whether we were talking a human doing the flipping and inputting the results into a program. If that detail was important, they would have told you in the video.

>Are we observing the coin flips from distance and updating the model based on the flips?
Again, not part of the problem, but what do you mean by "model"? If you mean the algorithm, then no; the dataset that the algorithm learns from, then yes.

>Or are we actually flipping and the probability depends on physics as well?
Not part of the problem but at least somewhat related to whether this is a fully or partially observable problem. Now, you can philosophize all you want about whether some omniscient being could predict the outcome of a coin flip by calculating all the variables in the universe, but you can't and neither can your computer, so we're clearly dealing with a partially observable problem.

Try not to think too much about the problems other than what is being asked or you're going to lose sight of what you should be learning. They will get the bigger questions later, I assure you.

Name: Anonymous 2011-10-17 12:59

>>14
Imho the partial observability yields from the fact that we need to memorize past flips. In other words, just by looking at a coin flip does not give us the optimal results.

Name: Anonymous 2011-10-18 5:35

The course consists completely of set theory bullshit, but mathematics is a pseudoscience, so AI must be pseudoscience too.


- you know you say things are "much of a muchness"--did you ever see such a thing as a drawing of a muchness?'
- Really, now you ask me,' said Alice, very much confused, `I don't think--
- `Then you shouldn't talk,' said the Hatter.


Same for "Infinite Set" - math people cant draw, but constantly talk about it.

Name: Anonymous 2011-10-18 11:36

I'm still lost with A* functions.

Name: Anonymous 2011-10-18 11:42

Now that the answers have been presented, can anyone explain to me why the loaded coin question is discrete?

'A function is called continous if there is an infinite number of actions the agent can take, or an infinite number of things an agent can sense'. While the actions the agent can take are certianly discrete, the coin can be tossed an infinite number of times - and thus there is an infinite amount of data for it to sense.

Name: Anonymous 2011-10-18 11:45

>>16
Sure. And since CS uses mathematics, then CS is a psuedoscience too.

Name: Anonymous 2011-10-18 12:09

>>19
Don't see him.

Name: Anonymous 2011-10-18 12:11

>>19
Yes. Turing Machine even has infinite tape.

>>20
just add "math infinity turing" into wordfilter. it'll remove 95% of my posts.

Name: Anonymous 2011-10-18 12:18

In truth, I was expecting a much more demanding course; and perhaps in a couple of weeks things will be more up to speed. As of now, the material covered is fairly fundamental and far from engaging. The former is understandable given that it's just started and the latter's certainly on Thrun and his inability to communicate effectively. I wonder when Norvig will finally intervene.

I'm following the Machine Learning class as well and the content is also very basic, but Ng is an outstandingly engaging lecturer. Unlike Thrun, who breathes heavily on the camera and can not seem to spell properly nor formulate coherent quiz questions (curiously, the Homework Assignment was very clear).

Yudkowsky does a much better job at explaining Bayes' Theorem…

Name: Anonymous 2011-10-18 12:24

>>18
Well that would happen if the outcome was to keep flipping the coin. In one flip though, there is a finite result, either heads or tails. The coin won't flip eternally, it will eventually land on one position.

Name: Anonymous 2011-10-18 12:26

>>22
Is the Machine Learning one still open enrollment?

Name: Anonymous 2011-10-18 12:32

>>24
Probably, but not on the Advanced track I'm guessing.

Name: 25-san 2011-10-18 12:35

>>24

http://www.ml-class.org/
Hm, it seems the Advanced track is still open for enrolment...

Name: Anonymous 2011-10-18 13:50

Are there any good AI book that doesn't rely on useless mathematics? I mean why they require infinitesimals, if in real world qualia is always fixnum?

Name: Anonymous 2011-10-18 13:58

Modern mathematics has no contact with reality.

In the past physics provided problems for mathematicians to work on, but to-day we have plenty of problems in mathematics itself.

None of the really new mathematics of this century has found any practical application whatever.

It should be studied as an art form.

Name: Anonymous 2011-10-18 14:02

I tend to quickly forget math unless I see it applied to real world applications because it helps me visualize the math in a way that pure memorization can never achieve.

I was surprised that she disagreed. She told me that "back-in-the-day", non-applied mathamaticians looked down upon applied mathamaticians as the working class. She told me that she was about halfway inbetween the 2 opinions, although the sections in the book she is choosing to skip tell a different story. She refered to math as "mental msturb...ion" (I would hope this word is censored in the forum). It's not hard to see her point. I get that warm fuzzy feeling too, every time I struggle for a half hour on a problem and finally figure it out. But that said, I have to admit I would have zero interest in math if math could not be applied to solve real world problems. If 2 dollars + 2 dollars didn't equal 4 dollars, I wouldn't even have an interest to know simple addition.

Name: Anonymous 2011-10-18 14:21

Most math has to be applied. How do you get the pytagorean Theorem without a physical triangle?

Name: Anonymous 2011-10-18 14:24

Three of my friends, over the years, got A's. Unlike me, they were very diciplined students: no partying until the homework is finished. But a few weeks ago, I gave all of them the same challenge: create a word problem that requires Calc I material to solve. You do not have to solve the problem. Just come up with one. I only got answers like "well, its the slope of the curve...". I'd respond with "why might I want to know that?" But not one of the A students could come up with a real-life application

Name: Anonymous 2011-10-18 14:29

>>30
ahahahahaha

Name: Anonymous 2011-10-18 14:30

>>31
Derivation is very useful for creating beautiful danmaku.

Name: Anonymous 2011-10-18 14:30

triples

Name: Anonymous 2011-10-18 14:32

Those calling maths useless are putting up some arbitrary standard of their own, by which the designate "usefulness" to some class of activity, and "uselessness" to some other class, rarely bothering about defining what either term means.

Why should they be regarded as the hallowed possessors of objective judgment?

For example, someone with the individual quirk who gets pleased by seeing an elegant proof for some assertion, is perfectly justified in saying that maths is useful in generating personal pleasure.

And what's wrong with that?

Why is mental "masturbation" worse than fiscal masturbation?

Name: Anonymous 2011-10-18 15:03

>>27
Enjoy being a disposable ENTERPRIZE Java. That is, if you achieve your full potential, which you probably won't.

Name: Anonymous 2011-10-18 15:21

>>22
I wonder when Norvig will finally intervene.
Norvig has given most of the talks so far. Are you sure you're not confusing him with Thurn? Thurn is the one with the German accent.
Ng is an outstandingly engaging lecturer.
Ng's assburger's is too much for me.

Name: Anonymous 2011-10-18 16:06

>>37
Ng
U MENA Ŋ

Name: Anonymous 2011-10-18 18:03

>>36
Java is:
1. Based of Set Theory (classes = sets)
2. Ugly mathematics-like infix notation

I'll rather use Lisp, cuz it looks nothing like math.

Name: Anonymous 2011-10-18 19:29

>>39
Set Theory (classes = sets)
Whoa, just who are you trying to troll? Bertrand Russell doesn't hang out here you know. You might check LtU Zombie forums though.

Newer Posts
Don't change these.
Name: Email:
Entire Thread Thread List