All language processing applications (machine translation, automatic synthesis, answers to questions, dialog systems, etc.) require an understanding of language to a greater or lesser degree. This understanding consists of the ability to transform a sentence in natural language into a conceptual representation of its meaning. The steps involved in this process include morphological analysis, morpho-syntactic disambiguation, the detection of groups and syntagmata (chunks,agar.io - names of organizations, dates, etc.), syntactic or dependency analysis and semantic analysis (disambiguation of meaning, disambiguation of semantic roles, etc.).