align-toparrow-leftarrow-rightbackbellblockcalendarcamerachatcheckchevron-downchevron-leftchevron-rightchevron-small-downchevron-small-leftchevron-small-rightchevron-small-upchevron-upcircle-with-crosscrosseditfacebookglobegoogleimagesinstagramlocation-pinmagnifying-glassmailmoremuplabelShape 3 + Rectangle 1outlookpersonplusImported LayersImported LayersImported Layersshieldstartwitteryahoo

Machine Learning Discussion Group - Deep Learning Algorithm

This week Brian will lead the discussion on Deep Learning, an increasingly popular machine learning algorithm in recent years. For those that want to get a jump start before the meeting, please visit the following links:

http://deeplearning.net/
http://ufldl.stanford.edu/wiki/index.php/UFLDL_Tutorial

Also, a shout of thanks to Andrea for his very detailed overview on Support Vector Machines during last week’s meetup.

We will be meeting at Hacker Dojo in a conference room down the hallway to the left of the entrance.  You do not need to be a member of Hacker Dojo to attend.

For those new to this group, the purpose of this meetup is to learn from each other the various aspects of machine learning. Our focus will be on intuitions, theories, and implementations behind machine learning algorithms. The format will be a chosen algorithm or topic reviewed/explained by a pre-assigned member through a highly interactive discussion session, and it is expected that other members will draw from their respective knowledge and experience to contribute to the explanation on the topic. We may also discuss specific machine learning projects should time permits after the chosen topic discussion.

The background of attendees varies from those that have exposure to machine learning, whether through Andrew Ng’s Coursera class or through other learning channels, to those that use machine learning in their day-to-days. However, even if you have no exposure to but have interest in machine learning, you are welcome to attend the meetup. Of course you might get more out of it if you go through the videos in Andrew Ng’s free Coursera class or equivalent.

Join or login to comment.

  • Tom B.

    Quote from Nasa winner on class forum, about image segmentation code he used: Since open source libraries were allowed, I adapted the "Efficient Graph-Based Image Segmentation" from Pedro F. Felzenszwalb and Daniel P. Huttenlocher.
    http://cs.brown.edu/~pff/segment/

    2 · August 4, 2013

  • Tom B.

    The presentation on the neural net was very interesting. Thank you so much for presenting.

    As I mentioned at the end of the meeting, I was reading the ML class forum and found a post by a former student, Sébastien Drouyer who had won a NASA challenge for recognizing instrument boards on the space station. He references using some image processing methods he used. I also found image processing code for Mathlib and Octave by Peter Kovesi at the University of western Australia. http://www.csse.uwa.edu.au/~pk/research/matlabfns

    I haven't had a chance to full absorb all functions there are several categories of them.

    My goal is find a better way of segmenting an image in Octave.

    1 · August 4, 2013

  • Stoney V.

    IPAM deep learning workshop videos 2012
    https://www.ipam.ucla.edu/schedule.aspx?pc=gss2012

    Deep learning tutorials and example code
    http://deeplearning.stanford.edu/wiki/index.php/UFLDL_Tutorial

    Large scaling up of deep learning at a lower cost
    http://stanford.edu/~acoates/papers/CoatesHuvalWangWuNgCatanzaro_icml2013.pdf

    1 · August 3, 2013

    • Stoney V.

      Also, pay attention to the newer Berlin and Kaveri cores with unified memory for CPU + GPU streaming units.

      August 3, 2013

  • John P.

    Great presentation!
    Please use this comments section to suggest topics for the next sessions. we will use topic modeling to group the comments into topics. :)

    August 3, 2013

  • Wenfeng W.

    Excellent intuition! Glad to meet all of you guys.

    August 2, 2013

  • al f.

    Are we going to need a bigger room next time?

    August 2, 2013

  • Tom B.

    I just got the neural net software running on my Android phone, using the Octave port. About 50 x slower than my Core I5 laptop. It takes ~1ms to predict one digit (20x20) with 25 hidden node net.

    August 2, 2013

  • Rishabh S.

    Looking forward to attending tomorrow

    August 1, 2013

  • Julie L

    I've never been to a MLDG meetup or Hacker Dojo before. How long do they last? Is there a place to get food/dinner nearby?

    August 1, 2013

    • Bill

      There is a sandwich place pretty much next door, but it closes at 6:30 so you would need to be there early. It is a combination of Specialty's and Panera Bread.

      August 1, 2013

Our Sponsors

  • Hacker Dojo

    Hacker Dojo provides meeting space for this group.

People in this
Meetup are also in:

Sign up

Meetup members, Log in

By clicking "Sign up" or "Sign up using Facebook", you confirm that you accept our Terms of Service & Privacy Policy