Return to Video

01_Definition of ML

  • 0:00 - 0:01
    Hi, Michael.
  • 0:01 - 0:02
    >> Hey, Charles, how's it going?
  • 0:02 - 0:04
    >> It's going quite well. How's it going with you?
  • 0:04 - 0:05
    >> Good. Good.
  • 0:05 - 0:08
    >> Good. Good. So today, I thought we would
  • 0:08 - 0:10
    talk a little bit about the philosophy of machine learning.
  • 0:10 - 0:12
    >> Ooh. I hate philosophy.
  • 0:12 - 0:15
    >> I don't like it much either, although I'm a doctor of philosophy.
  • 0:15 - 0:16
    >> Oh, that's very impressive.
  • 0:16 - 0:17
    >> Aren't you a doctor of philosophy, too?
  • 0:17 - 0:18
    >> I am. It's kind of impressive.
  • 0:18 - 0:20
    >> It is. It is kind of impressive. So
  • 0:20 - 0:21
    what we wanted to get across today was a
  • 0:21 - 0:23
    little bit about why the class is structured the
  • 0:23 - 0:27
    way it is. What the different parts are, and maybe
  • 0:27 - 0:29
    go a little bit of back and forth about what we
  • 0:29 - 0:31
    think you should be getting out of the course. That seem reasonable?
  • 0:31 - 0:31
    >> Sure.
  • 0:31 - 0:34
    >> Okay. Well so first off, by the way, before we get started, I wanted
  • 0:34 - 0:35
    to thank you for coming down to
  • 0:35 - 0:38
    Atlanta and joining me in these beautiful studios.
  • 0:38 - 0:41
    >> Well it's, it's very nice to be here. Thank you for inviting me.
  • 0:41 - 0:43
    >> Oh, no, no. Thank you for coming, Michael.
  • 0:43 - 0:46
    >> Thank you for asking me to do the course, this has been a lot of fun.
  • 0:46 - 0:47
    >> Oh, the whole point was to be able
  • 0:47 - 0:49
    to do the course with you, Michael We like each
  • 0:49 - 0:50
    other, and that's one of the things that we
  • 0:50 - 0:52
    want you to get, want to get across in this
  • 0:52 - 0:54
    class, because we like machine learning and lot of
  • 0:54 - 0:57
    stuff in common, but I'm not sure we completely
  • 0:57 - 1:00
    agree on the most important parts of machine learning
  • 1:00 - 1:01
    and why we do the things that we do.
  • 1:01 - 1:03
    >> Hm. Alright.
  • 1:03 - 1:04
    >> So I think people in the outside world Michael
  • 1:04 - 1:06
    would claim that you're more theoretical than I. But I am?
  • 1:06 - 1:08
    >> In theory.
  • 1:08 - 1:10
    >> In theory. And I'm more practical than you are.
  • 1:10 - 1:12
    >> Practically.
  • 1:12 - 1:15
    >> At least in practice. And hopefully some of that
  • 1:15 - 1:17
    tension will come out in the class. But I think in
  • 1:17 - 1:19
    order to see why that tension works that way,
  • 1:19 - 1:22
    you have to understand what machine learning is. So, Michael.
  • 1:22 - 1:23
    >> Right.
  • 1:23 - 1:23
    >> What's machine learning?
  • 1:23 - 1:25
    >> It's about proving theorems.
  • 1:25 - 1:27
    >> [LAUGH] No.
  • 1:27 - 1:27
    >> No.
  • 1:27 - 1:29
    >> I would not say it's about proving theorems,
  • 1:29 - 1:31
    although proving theorems is often important in machine learning.
  • 1:31 - 1:33
    >> I agree with that
  • 1:33 - 1:33
    >> Okay.
  • 1:33 - 1:34
    >> So we're on the same page.
  • 1:34 - 1:37
    >> We're partially on the same page. What is machine learning?
  • 1:37 - 1:39
    >> What is machine learning? Right.
  • 1:39 - 1:40
    >> Give me a definition.
  • 1:40 - 1:41
    >> So,
  • 1:41 - 1:46
    it is computational statistics. How's that for a definition?
  • 1:46 - 1:50
    >> That is a definition. It is wrong on so many levels. However, a lot of
  • 1:50 - 1:51
    people would agree with that statement. They would
  • 1:51 - 1:54
    say that machine learning is really just applied statistics.
  • 1:54 - 1:56
    >> Not applied statistics, computational statistics.
  • 1:56 - 1:59
    >> Computationally applied statistics.
  • 1:59 - 2:00
    >> Psh.
  • 2:00 - 2:04
    >> I don't like that definition. I think that it's a
  • 2:04 - 2:07
    bit too narrow. I think that machine learning is about this broader
  • 2:07 - 2:12
    notion of building artifacts, computational artifacts typically that
  • 2:12 - 2:14
    learn over time based on experience. And then in
  • 2:14 - 2:17
    particular, it's not just the act of building these
  • 2:17 - 2:19
    artifacts that matter, it's the math behind it, it's
  • 2:19 - 2:22
    the science behind it, it's the engineering behind
  • 2:22 - 2:24
    it, and it's the computing behind it. It's everything
  • 2:24 - 2:28
    that goes into building intelligent artifacts that almost by
  • 2:28 - 2:31
    necessity, have to learn over time. You buy that?
  • 2:31 - 2:31
    >> Yeah, so you,
  • 2:31 - 2:35
    you have data, and you do analysis of the data and try to glean things from
  • 2:35 - 2:38
    the data and you use various kinds of
  • 2:38 - 2:43
    computational structure to do that. So, computational statistics.
  • 2:43 - 2:46
    >> I don't think that's computational statistics.
Cím:
01_Definition of ML
Video Language:
English
Team:
Udacity
Projekt:
UD675: Machine Learning 1 - Supervised Learning
Duration:
02:47
Cogi-Admin edited Angol subtitles for 01_Definition of ML

English subtitles

Felülvizsgálatok