http://nlp.stanford.edu/courses/NAACL2013/
A very useful assignment for getting started with deep learning in NLP is to implement a simple window-based NER tagger in this exercise we designed for the Stanford NLP class cs224N. The zip file includes starter code in Java and the pdf walks through all the steps:
Stanford NLP class cs224: http://www.stanford.edu/class/cs224n/syllabus.html#lec14
Two course slides are here:
- http://www.stanford.edu/class/cs224n/handouts/CS224N_DeepNLP_lecture1.pdf
- http://www.stanford.edu/class/cs224n/handouts/CS224N_DeepNLP_Week7_lecture2.pdf
The topic was covered in the above NAACL 2013 Presentation.
- Video1: http://techtalks.tv/talks/deep-learning-for-nlp-without-magic-part-1/58414/
- Video2: http://techtalks.tv/talks/deep-learning-for-nlp-without-magic-part-2/58415/
The same tutorial was also on youtube (lower quality): http://www.youtube.com/watch?v=IF5tGEgRCTQ