معالجة المستوى التركيب-دلالي اللغة : بين المقاربة الإحصائية و المقاربة المبنية على نظام القواعد

Abstract EN

At first glance, it seems that the statistical approach to natural language processing is in complete contradiction to the rule-based systems approach .

Namely, that the latter is based on an explicit declaration of syntactic and semantic rules as a necessary preliminary phase for machine translation .

Hence the legitimacy of asking the following question: why the statistical approach is by far more effective even for non-selected texts in comparison with the results from the rule-based approach designed by a multitude of linguist experts in several years?