The Big Beard Team is taking the most popular Machine Learning course on the deep reaches of the internets and you can follow along us. Coursera’s Machine Learning by Andrew Ng is one of the cornerstone courses for Coursera and one of the first MOOC (Massively Open Online Courses) in the Machine Learning category. Erin, Kyle, Brett, and myself have decided to sign up to attempt for our certification in this course. In the Big Data Beard fashion we want to record our journey throughout this course. Yes that includes the good and the bad.
the coursera machine learning Andrew Ng week 1
Kicking off our YouTube series on the Coursera Machine Learning course Erin and I covered week one. The course is broken out over 11 weeks which leaves no time for an easy week. The first week jumps right into so deep math from my perspective. Check out the Machine Learning course syllabus below:
Machine Learning Syllabus Week 1
- Introduction – Basics covering how to take the course and introducing yourself to other students in MOOC.
- Linear Regression with one Variable – Explanation of running examples for Machine Learning course. Predicting mortgage rates based on multiple parameters and detecting cancer in patients. Definitions are presented in this section like supervised/unsupervised learning, cost function, classification models, and etc.
- Linear Algebra Review – Review for some (not me!!) on the concepts of Linear Regression and how to use it for data modeling.
Learn more about Machine Learning – Andrew Ng Week 1 and tips for taking this course from Erin and myself in the video below.
Video – Machine Learning – Andrew Ng Week 1
Erin K. Banks: Hey everyone, Erin Banks and Thomas Henson here from the BigDataBeard.com group. And recently, we decided to sign up for this class on Coursera, and it’s the Introduction to Machine Learning with Andrew Ng. And it’s an 11-week course, and we figured we’d just talk about it every week. So, if you’re interested in taking the course or thinking about taking the course, subscribe to the channel, to the YouTube channel, and follow us every week to kind of get our updates. Again, this is just for week one. And this is, again, our first time talking about the course, and our thoughts on it, and our opinions on everything like that.
So, Thomas, what were your thoughts about week one for this course?
Thomas Henson: Well, I thought week one would be a little easier. We might have to change the name of the series to “Watch Me Fail the Machine Learning Course”.
Erin K. Banks: [Laughs]
Thomas Henson: So, I don’t know. I mean, it was pretty difficult. You know my background, I have got an undergrad and a grad degree, but they’re all on the business side. So, I was computer information systems, and even on the master’s, so I didn’t really go too deep into math. So, this is like my first kind of go around going really deep in math. And also, I think the last time I sat in a college class was, I don’t know, six, seven years ago.
Erin K. Banks: Yeah. Yeah. I had that same problem. So, my background definitely has math in it. It’s electrical engineering I have my degree, but it was, I don’t even know if I want to tell you when I graduated college. It was a long, long time ago.
Thomas Henson: [Laughs] Tell us.
Erin K. Banks: And I was like…
Thomas Henson: This is the truth area.
Erin K. Banks: [Laughs] So, 1997, I graduated from Northeastern University in Boston, Massachusetts, [Smacks forehead] ugh, with an electrical engineering degree. So, I haven’t done any of it. So that’s what I thought (A) it was a really…I thought the best part was having other people take the class because I have to just always start off…I always have to start off with a negative and then be prepared and do the positive, because my biggest frustration — I totally agree with you — I thought the practice exams, like, I was going to do great. But then when it came to the last test, which you can take three times and then if you fail all three times you have to wait eight hours, and I failed the first time I took it. And there wasn’t any way to really go back and look, and there wasn’t any explanation. Because I think a bunch of us, right… So, all together there’s four of us that are taking the class that we know of, just within the BigDataBeard community, and we felt…we would have meetings and talk about the course and everything along those lines in separate, just, Hangouts and FaceTimes and things like that. And there was no explanation after the tests about what you got wrong or why you got it wrong or what you were kind of learning along those lines or kind of anything to give you hints, “Hey, it’s in this area. This is what you should practice or read up on this.” And I didn’t like that because I kept feeling like if I were in a real course, I would be able to ask somebody. I would be like, “Okay, this didn’t make any sense to me. Can you help explain it to me? Can you provide some greater insight?” And we don’t have that.
And I went back in the forms to look and somebody had specifically asked for it, and the mentor that’s on that for week one was like, “We don’t do that,” and I said, “I think you guys need to rethink that. That’s a huge failure, I feel, on your part because you’re basically either…you’re going to fail.” And other people will be like, “Okay, well this is the side information that I got. I had to do some Googling. These are how these formulas work. This is along those lines. It wasn’t explained the best way, so check it out.”
Thomas Henson: Yeah. So, first off, you have an EE and you’re complaining. I mean, wow. [Laughs]
Erin K. Banks: Well, it was so long ago, though. Like, I don’t… And it was a joke. I haven’t done electrical engineering in forever, even when I first got out of college. So, it’s just a little…it’s just not… Electrical engineering was great, but it’s just not something that I’m always doing in my day-to-day life, especially now.
Thomas Henson: Yeah. But I’ll tell you one thing, so, same problems there. I think I never got more than 80 percent on any of the tests. I didn’t have to take each one multiple times, but I started finding out — let me see, let me look at my notes. So, I’m a notetaker. So, anybody that’s watching, I’m a crazy notetaker in my, also, Moleskine. You know, I’m a snob when it comes to notebooks.
Erin K. Banks: I hear you.
Thomas Henson: But one of the things that I started figuring out, I guess after I got through the cost model — like the model and cost function — when I started getting into the linear algebra parts, I think the last two sections, what I started doing was, instead of watching the video first, I jumped ahead, took the notes, and went through some of that. Even though I totally didn’t understand some of it, it was good. You’re pretty much just practicing your writing a little bit there it feels like, but you’re trying to figure it out in your head by reading it, then go back and watch the video and it makes a whole lot more sense. But I will say it was still hard when I would go through and take those quizzes. But it was kind of weird because I felt like through part of the linear algebra I did really well in the end course quizzes. So, I don’t know if you remember as you were going through and you had to do some of the…
Erin K. Banks: Yeah.
Thomas Henson: …pretty much it was just like some fun, simple math once you figured out the equations and stuff. So that’s my tip for anybody taking it, if you’re a crazy notetaker. Because a lot of times, like what I was doing through the first two sections is, I was missing what Andrew was saying because I was sitting there frantically trying to copy down what he was doing on the slides, and then I was kind of losing it and having to go back and reference in my notes. So, what I started doing was taking those notes beforehand, because at the end of each section there’s kind of like a summary after the video, and so, honestly, for me, I think it should be switched. And so, if anybody out there is a notetaker, just try that for one section and see how you do. It seemed to work a little bit better for me.
Erin K. Banks: Yeah, and you can download the transcripts, right. You can download the slides. You can download a lot of the information that’s happening on that course, and even when you’re taking the final exam, you have all of that information available to you which was great, and I did appreciate that. But I think that’s really important to understand. I think I hopefully will do better now that I understand how the course is kind of laid out and how the tests are done. Because you get a totally, again, kind of a nice feeling when you’re taking those sample quizzes along the way like, “Yeah, I get this. I understand this. This is making sense. I’m really excited,” and then that last five questions that they ask, and you have to get four of them right or you fail, just brought a little tear to my eye [INAUDIBLE 00:06:18] I’m like, “Oh my god, I’m going to… This is… I thought I was better than…” not better than this, but I thought I had it, right. I had a little bit of confidence and then it was squashed. So, I love your points about reading it and then going back. And especially, that’s certainly what helped me taking it again, because when I realized, “Alright, well, you failed the test. You have to really sit down and review it,” I’d already taken the notes — I use Evernote to take my notes. I use them both, either typing it or actually writing it — but I knew that I was going to need all that information to easily reference it back.
So, I took all those notes and then watched it where I could consciously look at him and understand. And I guess in some sense it was also nice because I knew what the questions were around, so then I could be like, “Alright, this is…” I wasn’t looking at it the way I guess they were looking at it, maybe. Maybe. I don’t know if that’s kind of correct, but the way that they were asking it, I was kind of viewing it a little bit differently. Now I see what they’re saying. Now I completely understand this formula means this, and this is how I have to use it, and this is kind of like…it made a lot more sense. Finally, by the time I did all that, then I was able to just take that test perfectly, thank god. So that worked out for me.
Thomas Henson: Yeah. I mean for me, like I said not having a lot of the deep math, it was kind of interesting as we were going through, I think it was the gradient descent section where he was like, the first few slides he was kind of going through, and I thought it was pretty deep. And then at the end, he kind of gave us the hook for what we’re going to learn in the next video. He’s like, “And now we’ll go and get deeper.” And I’m like, “Wait! Wait a minute. [Laughs] We’re already pretty deep, so.” [Laughs]
Erin K. Banks: Yeah. Yeah, I totally agree. I mean, I’m definitely… I’m looking forward to week two. I think it’ll be good to kind of see how it goes. Again, I can’t stress enough how nice it is, a little bit, to have a support system, especially for week one. That’s just been great to just kind of get some clarity in my brain, and try to like, “Okay, is this how I make it work?” “No.” So that certainly is helpful. And I know they talk about that in the actual class, right, that you need the group to talk to one another and kind of do the introductions. That tends to be difficult, so it’s nice if you’re able to…if you know some of the people and you’re able to use some of the technology that we have nowadays like a Google Hangout or a FaceTime or anything along those lines; get a couple people together even if they have to leave and kind of vent about it. It’s kind of almost like a support group to help you kind of get through the class. I’m not sure if I would be able to do it without everyone, just to kind of, “Hey, did you take that quiz yet? What did you think? What were your thoughts?” and really just be able to bounce it off of one another, which is nice.
Thomas Henson: Yeah, kind of struggle together. Maybe that’s one thing we should do is maybe we should start posting a link to the YouTube videos once we get these pushed out, internally, even inside of our course. See if some of the other students, maybe it’ll kind of help, kind of like a sounding board.
Erin K. Banks: Yeah. Yeah, that’s great. And certainly, if they can add comments or anything along those lines, because I think the more information, and to be prepared. Because I had no…I’d never taken a course before through this company, so I didn’t really know what to expect. I’ve taken online classes, I was working on my MBA, so I was trying to do those kind of things, and you just don’t know. And certainly, classes like this are, I think, somewhat difficult when it’s…everything’s online.
Thomas Henson: Yeah. And I think as we kind of go through this course it’s going to be a lot more fun. And you were talking about week two. I’m excited about week two when we get to start using, I think it’s Octave, versus Mathlab, [Phonetic] and just kind of be able to go through and get hands-on a little bit more. So, I’m excited to see what that brings, and maybe I’ll be able to shine a little bit more there. But, for everybody that’s watching, I think this is great, just with you and I being on. So, Erin who has a major math background with EE and some other pieces but maybe hasn’t used those math skills in a while.
Erin K. Banks: [Laughs] Yeah.
Thomas Henson: And then me who hasn’t used those math skills in a while, but never had them, too. So, for people watching along or really just contemplating whether to take the course or not, we’ll be your guinea pigs and you can kind of follow us along.
Erin K. Banks: Yeah, I love that. I certainly would wish I had that to kind of begin with. So, I think it’s great, great capabilities. And I look forward to week two. So, hopefully everyone will subscribe to the YouTube channel so that they can follow along with us as we do the course. And certainly, if you’re preparing to take it or you’re thinking about it, just review them. These are going to be really quick snippets just to get your, kind of, opinions, best practices, and what we’ve learned along the way so that you can be successful yourself. So, thanks for your time.