Introduction to Machine Learning

Introduction to Machine Learning(alex.smola.org)

181 points by chorola 13 years ago | 21 comments

maxjus 13 years ago |

CloudFlare is saying that the site may be down due to "excessive load." I was under the impression that this is one of the main reasons CloudFlare exists, to handle big surges in traffic like this one. Their intro video leads me to the same conclusion. Can anyone who uses CloudFlare shed some light on this?

bkanber 13 years ago | |

Cloudflare isn't too much more than a wrapper around a CDN for your static assets. The request still gets sent to your server, they just re-host your images and stuff on their edge servers. They also do some "security stuff", and can help you with minifying assets and things like that. So excessive load can still be a problem.

On top of that, the free account doesn't really come with any guarantees.

osteele 13 years ago | |

CloudFlare doesn't proxy HTML pages, by default, because so many of them are dynamically generated. You have to add a page rule to proxy HTML too.

gwern 13 years ago | | |

Thanks. I was actually under the impression that CF would do HTML as well, but I searched and you seem to be right: https://support.cloudflare.com/entries/22036831-How-do-I-cac... So I'll try that out and see if it helps performance/bandwidth.

consta 13 years ago |

Thanks for sharing. I can also recommend the CS229 Machine Learning course from Stanford, available on Coursera: https://class.coursera.org/ml/lecture/preview

dvdhsu 13 years ago | |

One caveat though: the Coursera ML course is CS229A(pplied), which focuses on applying machine learning, rather than the mathematics behind it. The "real" CS229 lectures are on YouTube [1], and go much more in depth.

1. http://www.youtube.com/watch?v=UzxYlbK2c7E

alexholehouse 13 years ago | |

Minor plug - if you're following the CS229A course you may find my notes useful

http://www.holehouse.org/mlclass/

jervisfm 13 years ago | | |

Curious: what did you use to take the notes ?

dtld 13 years ago | |

awesome!

chorola 13 years ago |

Alexander J. Smola's teaching the 10-701 Machine Learning class at CMU in the spring of 2013. All slides and video will be available online at the course website.

kumarharsh 13 years ago |

On a related note, check out

https://news.ycombinator.com/item?id=5424493

We built a A nodejs native extension which allows you to use the popular CLIPS system to build expert systems. CLIPS is a productive development and delivery expert system tool which provides a complete environment for the construction of rule and/or object based expert systems. Developed at NASA in 1985, CLIPS is now widely used throughout the government, industry, and academia. Check out the online CLIPS documentation: http://clipsrules.sourceforge.net/OnlineDocs.html

virmundi 13 years ago | |

Why did you use a dead rules engine? That system hasn't been maintained since 2007, as far as I can tell. I am serious. My client used it recently too. Why not use a news, faster engine like Drools or whatever JBoss is calling it now?

Trufa 13 years ago |

Google cache copy: http://webcache.googleusercontent.com/search?q=cache:http://...

fatjokes 13 years ago |

Is it just me, or are half the resources (i.e., slides for most lectures) missing?

shmageggy 13 years ago |

I think this is an instance where HN's rigid title policy is a negative. It took me a minute to realize why this page, ostensibly a syllabus page for a university course, is relevant to a general tech audience. You have to scroll down past three subsections to "Resources" to realize that all of the lecture slides and problem sets, as well as videos of the lectures, are available for free. I just hope people aren't missing out on this great resource because the title can't be editorialized.

_delirium 13 years ago | |

I like the Metafilter solution to that: rather than just submitting links, users submit links and a short, 1-2-sentence explanation of why they think the link is interesting. I find it much easier to browse the submissions there as a result. Maybe the explanations could be hidden on the front page, but I'd find them very useful on the 'new' page.

SeanDav 13 years ago | |

Wow. Exactly what you feared happened to me. I tool a quick look at the link, couldn't see the point and promptly closed it.

Sometimes, on a quiet day, some quite dubious content can reach the front page and I thought that this was a such a case.

Thank goodness I saw your comment.

BTW. What is HN title philosophy? Could the OP not have said: Lectures available online for Introduction to Machine Learning?

yebyen 13 years ago | |

I upvoted you not because I think you're exactly right, but because I think that the submitter should actually leave that comment and it gets upvoted/downvoted like all the other comments. I've never submitted anything to HN and had it posted on the front page. But I get the idea that you have plenty of time to leave a comment before it gets there.

Personally have really enjoyed reading submissions where the author is the poster, and he's there waiting in the comments to address your concerns and taking suggestions.

chorola 13 years ago | |

Thank u for yr advice.Unfortunately I find no way to edit my submit.Would u pls show me how to edit it?I tried to add some sentences to give more information,at last it just became a new comment.