машинний переклад

The paper addresses the problem of designing a machine translation service for variable data that operates at the level of document segments rather than monolithic text. In contemporary publishing, marketing and educational workflows, documents are increasingly represented as sets of structured segments associated with content keys, language attributes and, in many cases, personalization markers inherited from CRM or content management systems.

The article summarizes the author’s contribution for translation Ukrainian sign language. We describe the model of translation used for foreign sign languages. Among these statistical models of translation, the rule-based translation models and ontology-based models are described. These models are compared with the alternative approach for translation Ukrainian sign language using grammatically-augmented ontology. The method of filling ontologies using domain-specific languages, allowing us to develop a grammatically- augmented ontology is described.