中国科技术语 ›› 2025, Vol. 27 ›› Issue (2): 109-117.doi: 10.12339/j.issn.1673-8578.2025.02.015

• 实践应用 • 上一篇    下一篇

大语言模型技术在口译术语管理中的应用

王华树(), 白华()   

  1. 北京外国语大学,北京 100089
  • 收稿日期:2024-09-05 修回日期:2024-11-08 出版日期:2025-03-05 发布日期:2025-03-06
  • 作者简介:

    王华树,男(1980—),博士,北京外国语大学高级翻译学院教授,博士研究生导师。研究方向为翻译技术、术语管理、语言服务管理。通信方式:

    白华,女(1999—),北京外国语大学高级翻译学院在读研究生。研究方向为翻译技术、术语管理、口译理论与实践、翻译与跨文化研究。通信方式:

  • 基金资助:
    国家社会科学基金项目“数字人文视域下译者数字素养研究”(22BYY043)

The Application of Large Language Models in Terminology Management for Interpreting

WANG Huashu(), BAI Hua()   

  • Received:2024-09-05 Revised:2024-11-08 Online:2025-03-05 Published:2025-03-06

摘要:

传统的口译术语管理方式存在信息孤岛凸显、及时响应滞后、管理效率低下、智能功能较弱等问题。大语言模型 (LLMs) 的出现为口译术语管理带来了新机遇。它具备强大的文本处理能力,能够自动提取术语、提供术语翻译、分析术语使用情况等,可提升术语翻译质量、提高术语管理效率、促进口译员的专业化发展。尽管如此,大语言模型在口译术语管理中的应用仍面临模型训练数据偏差、数据隐私安全堪忧、训练维护成本较高等诸多问题,未来研究需要进一步探索大语言模型在口译术语管理中的应用场景,优化算法与模型,提升其准确性与可靠性,并开发更加人性化和智能化的口译术语管理工具。

关键词: 口译, 术语管理, 大语言模型, 人工智能

Abstract:

Traditional methods of terminology management in interpreting have some problems, including information isolated island, real-time insufficiency, low efficiency, and intelligence insufficiency. Thanks to the powerful text processing capabilities, large language models (LLMs) offer new opportunities for terminology management in interpreting. LLMs can automatically extract terms, provide term translations, analyze term usage, thus improving translation quality, increasing management efficiency, and supporting the professional development of interpreters. However, the application of LLMs in terminology management for interpreting still faces challenges such as bias in training data, concerns over data privacy and security, and high costs for training an maintenance. Future research should further explore the application scenarios of LLMs in terminology management for interpreting, optimize algorithms and models to improve accuracy and reliability, and develop more user-friendly and intelligent terminology management tools.

Key words: interpreting, terminology management, large language models, artificial intelligence