todo.todo

CODE:
   ✔ Run the plot_letters on "Undirected" methoods and compare to "directed" @OHAD @due(19.05.2016) @done (16-05-17 07:56)
   ✔ Run on "Half words" (first 2 letters) and compare to LLCRF @TOPAZ @due(19.05.2016) @done (16-05-17 08:43)
 
PRESENTATION: @due(19.05.2016)
   ✔ Theoretical background on training (Tutorial) @TSACHI @done (16-06-02 14:45)
     ✔ Inference: FrankWolfeSSVM, https://pystruct.github.io/user_guide.html @TSACHI @done (16-06-02 14:45)
     ✔ Data presentation - see Whatapp PDF @TSACHI @done (16-06-02 14:45)
     ✔ Graph presentation - node for each letter in word, etc.. @TSACHI @done (16-06-02 14:45)

   ☐ Daphny's Paper: http://papers.nips.cc/paper/2397-max-margin-markov-networks.pdf @TSACHI
   ✔ Add reulsts from code/outputs @TSACHI @done (16-06-02 14:45)

COMMON: @due(22.05.2016)
   ☐ Describe limitations and future work for improvements (Part 3) @OHAD @TOPAZ @TSACHI
     ✔ General CRF - "before and after" letters @done (16-06-02 14:45)
     - RNN as CRF: http://www.robots.ox.ac.uk/~szheng/papers/CRFasRNN.pdf @OHAD
     ✔ SVM as CRF @done (16-06-02 14:45)


Important dates:
    ✔ 26-28.05.2016 @OHAD and @Topaz are in vacation @done (16-06-02 14:46)
    ✔ 02.06.2016 - Class presentation @OHAD @Topaz @TSACHI @done (16-06-02 14:46)

Final stage:
  ☐ Train CRF with SVM
  ☐ 
  ☐ 
  ☐ 
  ☐ 
  ☐ 
  ☐