I'm looking to the different methods of representing a text into a machine-readable format. However, until now, I only found "Bag of Words" approachs with a lot of variations (boolean BoW, weighted BoW, Bag of Topics...).
What is troubling me is that these methods don't take context into account (some approach this goal by counting n-grams, wich seems a bit light to me).
I would like to know if other approaches are also used (I found no mentions of graph representation, for instance). Papers/books/websites presenting these methods would be a huge bonus.