Subject matter experts ensure that the translated text fulfils the business requirements. As they did before for human translations. Only that now Semiox guides them through the parts which need revision and supports them with suggestions and terminology.
Multilingual knowledge systems are the base of the Semiox approach. Multilingual Knowledge Managers build and maintain the customer’s knowledge graph and terminology. They agree on the knowledge structure and fill-up the asset with terms in all languages. Their main objective is cross border semantic interoperability.
Neural machine translation leverages like content recycling previously translated texts. Linguistic Assets Curators organize terminology, translation memories, and translated documents. They train the NMT and select the content to be recycled. They update and also delete linguistic assets.
The Semiox workflow is customized to the type of data and the balance between cost, time, and quality. Translation Workflow Engineers select NLP components and NMT engines. They define criteria for quality and the deployment of human resources. They deliver business intelligence and audit trails.