On Building a Bilingual Terminology Corpus from the Perspective of Human-machine Collaborative Translation

Abstract

The rapid advancements in artificial intelligence and large language models have significantly improved the quality of machine translation, profoundly influencing not only professional translation workflows but also driving pedagogical innovation in translation teaching at the higher education level. However, the ongoing issue of terminological and conceptual inaccuracies in current machine translation systems highlights the need for further refinement. Given the centrality of terminology in academic discourse, ensuring accuracy in terminological translation is essential for facilitating effective scholarly communication. This paper seeks to propose a methodological framework for the systematic extraction of bilingual terminological data to support the development of specialized corpora, with a particular focus on the social sciences and humanities. The primary aim is to maintain terminological precision and conceptual consistency while ensuring strong contextual alignment. Additionally, the paper aims to design an innovative teaching model to equip translation students with the technical skills required to create discipline-specific bilingual terminology databases, addressing a critical competency gap in the contemporary language services industry.