Re:Deep Learning for NLP without Magic (2013)
by kastnerkyle on Monday August 18, 2014 @07:26AM
(#8192186)
There is a nice tutorial here [1] about max-ent classifiers. Ultimately in neural networks there are usually a number of cost functions you can use for the last layer - softmax or cross-entropy are other possibilities that may be easier to understand, though possibly less performant for NLP tasks.
I thought the introduction to neural networks from Andrew Ng's coursera course (even though it meant writing MATLAB) was quite good, and allows you to implement backprop, cost functions, etc. while still having some other helper code to make things easier. I highly recommend working through that course if you are intersted in ML in general [2].
[1] http://www.cs.berkeley.edu/~klein/papers/maxent-tutorial-sli...
You can check out Machine Learning: An Algorithmic Perspective by Stephen Marsland[0] which takes a less math-driven approach to ML. The code is available online if you want to take a look at it (it's written in Python)[1].
[0]: http://www.crcpress.com/product/isbn/9781420067187
[1]: http://seat.massey.ac.nz/personal/s.r.marsland/MLBook.html
Re:Deep Learning for NLP without Magic (2013)
by davmre on Monday August 18, 2014 @02:13PM
(#8193747)
The Stanford deep learning tutorials (http://deeplearning.stanford.edu/wiki/index.php/UFLDL_Tutori...) have pretty good notes and exercises with detailed guidance / starter code. Depending on where exactly your hangups are, working through some of the coding exercises there might be useful.
Re:Deep Learning for NLP without Magic (2013)
by Houshalter on Tuesday August 19, 2014 @11:23PM
(#8200961)
http://www.metacademy.org/ is the best resource for this.
Any resources to learn that stuff?