APOLLON

The study of ideas of historical texts require an expertise identifying the context, the precise
 meaning of the statements and to establish their genealogy. Statistical counts cannot take
 into account the nuances induced by the style, the presence or the absence of a word 
than may change the meaning of a sentence, the context.
 
Our project aims at automatizing such a task by combining the reinforcement learning 
technique with human expertise to create the first lexicon of ideas from the corpus of 
Aristotle's Politics, then to apply such techniques to other corpuses and the to automatic 
text generation.