There are two methods from computer science that I think are relevant to help understand French.
Syntax Grammar
First having a syntax grammar similar to what is used to build compilers [1]
An interesting project for a grammar is here: http://alpage.inria.fr/frmgwiki/book/export/html/10600
Here is a parse tree for “ce tremblement de terre a fait beaucoup de dégats”
Also, here is a paper discussing “A context-free grammar of French” [2]
NLP Representation
In “Neural Machine Translation By Jointly Learning To Align and Translate” [3] Bahdanau introduces the concept of soft attention which provides a soft alignment between words and phrases which details what words from the source utterance influenced the target utterance.
Works Cited
- Java 1.8 Grammar https://docs.oracle.com/javase/specs/jls/se7/html/jls-18.html[↩]
- M. Salkoff and Morris, “A context-free grammar of French,” in Proceedings of the 8th conference on Computational linguistics -, 1980.[↩]
- D. Bahdanau, K. Cho, and Y. Bengio, “Neural Machine Translation By Jointly Learning To Align and Translate,” Iclr 2015, 2014.[↩]