6 ms·

Some Notes on the “Andrew Ng” Coursera Machine Learning Course

62 points by harveynick 9y ago

lee101 9y ago
Ended up founding a ml startup predicting crypto currency prices https://bitbank.nz https://bitbank.nz after finishing this course! such a good skillset to have I didn't know what I would do when I started the course just that I should learn this ml stuff
- dalacv 9y ago
  serious question: how well was bitbank able to predict the recent bitcoin selloff?
  - kerberos84 9y ago
    given that he gives only 1 day free trial, clearly he is not confident of his system.
- harveynick 9y ago
  Post author here :-) Oh wow, this is right in the wheelhouse of my site... have you written anywhere about the process of setting it up / the progress you’ve made?
jcadam 9y ago
I took the 11-week Coursera course last year (and even paid for the 'verified' certificate). It was worth it, I think. I may not have a deep understanding of the different ML algorithms that were presented, but at least I know of their existence, and to what problems they are suited (and then I can look up the details when needed). Granted, that's not enough to actually find work doing anything ML-related, as there are now plenty of experts with years of experience out there on the job market, but hey, I like learning stuff for the sake of it. I'm thinking of taking one of the follow-up courses on NN.
- kregasaurusrex 9y ago
  What made you decide to take the paid certificate route? I've browsed job boards for the past few months; but have seen that employers are increasingly looking for degrees as a minimum requirement versus having a Coursera certificate. Especially with regards to anything involving the ML space, it seems like employers are more interested in it as a supplement to a degree rather than being useful on its own.
  - jcadam 9y ago
    Well, I have a BS in Comp Sci as well as an MBA. Moment of weakness, I guess. Besides, since I paid for it up front, I had that much more motivation to finish the course :) In any case, it's on the resume. No idea if it looks good or not. Of course, if I were really serious about it, maybe I ought to go for a PhD in CS. But at 37 years old, I'm approaching the end of my useful life as a software engineer. I understand I have to report to something called "Caoursel" on my 40th birthday, whatever that is. EDIT: Allright, I took it because at the time I was working as a software development "SME" on a program in which AI, ML and data science terms were being bandied about by the PhDs and I wanted to be able to sit in on meetings and at least follow the conversation (And not look completely stupid. Adopting a stern look and nodding along will only get you so far, occasionally you have to speak). The course largely fulfilled its purpose on that score.
    - kregasaurusrex 9y ago
      I'm a bit younger on the age range at 23, and taking some time out of school since my grades and finances weren't quite where I wanted them to be. Would you say a person who has a handful of relevant Coursera certificates and does well on an interview would be able to compete against a degree holder who also did above average? For your case of staying current in newer tech I can see the market reason for taking it. People like Knuth who're well into their 70's still seem to find ways to remain relevant as well. What is the "Caoursel" you're referring to?
      - jcadam 9y ago
enraged_camel 9y ago
I have tried to get into Machine Learning/Deep Learning a few times, but I found it difficult to get motivated because I feel like I don't have any immediate use cases in my own personal or professional life to which I can apply them. Even areas in which my company can potentially benefit from Machine Learning, we don't have access to vast troves of data to use as training sets. So our brainstorming sessions have gone nowhere.
- lee101 9y ago
  So many use cases! Take the course! You can use regularization to learn from small datasets by penalizing learning extreme biased feature weightings. For finding a use case or at work just think of any normal problem like UI or ux design, or showing someone an enjoyable loading screen or funny picture, anyone can make something okay but to solve a problem close to optimally you start to need to gather data and do ml/ai
  - enraged_camel 9y ago
    Well, for example, can I look at the company's sales performance over the past 24 months (along with various factors such as number of sales staff on the field) and accurately predict its numbers next month? Where would I start with this? Is this even an ML problem?
    - FlyingLawnmower 9y ago
      Time series data is hard, especially without lots of training examples. Consider looking into traditional forecasting methods as a baseline. AR(I)MA based approaches with seasonality might actually be fine for you. However, it all changes if there are structural breaks in your dataset. Look into Rob Hyndman's free online book (https://otexts.org/fpp2/ https://otexts.org/fpp2/) for an excellent introduction into timeseries forecasting.

62 points by harveynick 9y ago

Ended up founding a ml startup predicting crypto currency prices https://bitbank.nz https://bitbank.nz after finishing this course! such a good skillset to have I didn't know what I would do when I started the course just that I should learn this ml stuff

dalacv 9y ago
serious question: how well was bitbank able to predict the recent bitcoin selloff?
- kerberos84 9y ago
  given that he gives only 1 day free trial, clearly he is not confident of his system.
harveynick 9y ago
Post author here :-) Oh wow, this is right in the wheelhouse of my site... have you written anywhere about the process of setting it up / the progress you’ve made?

jcadam 9y ago

I took the 11-week Coursera course last year (and even paid for the 'verified' certificate). It was worth it, I think. I may not have a deep understanding of the different ML algorithms that were presented, but at least I know of their existence, and to what problems they are suited (and then I can look up the details when needed). Granted, that's not enough to actually find work doing anything ML-related, as there are now plenty of experts with years of experience out there on the job market, but hey, I like learning stuff for the sake of it. I'm thinking of taking one of the follow-up courses on NN.

kregasaurusrex 9y ago
What made you decide to take the paid certificate route? I've browsed job boards for the past few months; but have seen that employers are increasingly looking for degrees as a minimum requirement versus having a Coursera certificate. Especially with regards to anything involving the ML space, it seems like employers are more interested in it as a supplement to a degree rather than being useful on its own.
- jcadam 9y ago
  Well, I have a BS in Comp Sci as well as an MBA. Moment of weakness, I guess. Besides, since I paid for it up front, I had that much more motivation to finish the course :) In any case, it's on the resume. No idea if it looks good or not. Of course, if I were really serious about it, maybe I ought to go for a PhD in CS. But at 37 years old, I'm approaching the end of my useful life as a software engineer. I understand I have to report to something called "Caoursel" on my 40th birthday, whatever that is. EDIT: Allright, I took it because at the time I was working as a software development "SME" on a program in which AI, ML and data science terms were being bandied about by the PhDs and I wanted to be able to sit in on meetings and at least follow the conversation (And not look completely stupid. Adopting a stern look and nodding along will only get you so far, occasionally you have to speak). The course largely fulfilled its purpose on that score.
  - kregasaurusrex 9y ago
    I'm a bit younger on the age range at 23, and taking some time out of school since my grades and finances weren't quite where I wanted them to be. Would you say a person who has a handful of relevant Coursera certificates and does well on an interview would be able to compete against a degree holder who also did above average? For your case of staying current in newer tech I can see the market reason for taking it. People like Knuth who're well into their 70's still seem to find ways to remain relevant as well. What is the "Caoursel" you're referring to?
    - jcadam 9y ago

enraged_camel 9y ago

I have tried to get into Machine Learning/Deep Learning a few times, but I found it difficult to get motivated because I feel like I don't have any immediate use cases in my own personal or professional life to which I can apply them. Even areas in which my company can potentially benefit from Machine Learning, we don't have access to vast troves of data to use as training sets. So our brainstorming sessions have gone nowhere.

lee101 9y ago
So many use cases! Take the course! You can use regularization to learn from small datasets by penalizing learning extreme biased feature weightings. For finding a use case or at work just think of any normal problem like UI or ux design, or showing someone an enjoyable loading screen or funny picture, anyone can make something okay but to solve a problem close to optimally you start to need to gather data and do ml/ai
- enraged_camel 9y ago
  Well, for example, can I look at the company's sales performance over the past 24 months (along with various factors such as number of sales staff on the field) and accurately predict its numbers next month? Where would I start with this? Is this even an ML problem?
  - FlyingLawnmower 9y ago
    Time series data is hard, especially without lots of training examples. Consider looking into traditional forecasting methods as a baseline. AR(I)MA based approaches with seasonality might actually be fine for you. However, it all changes if there are structural breaks in your dataset. Look into Rob Hyndman's free online book (https://otexts.org/fpp2/ https://otexts.org/fpp2/) for an excellent introduction into timeseries forecasting.