自动生成词典:我们接近了吗?  

Automating the Creation of Dictionaries:Are We Nearly There?

作　　者：Michael Rundell 赵翠莲 于文雍 Michael Rundell;Zhao Cuilian;Yu Wenyong

机构地区：[1]Lexical Computing Ltd. [2]四川外国语大学,重庆400031 [3]复旦大学出版社,上海200433

出　　处：《辞书研究》2024年第4期1-14,I0001,共15页Lexicographical Studies

基　　金：2018年度国家社会科学基金重点项目“中华文化信息在新时期汉英词典中的凸显表征模式研究”(项目编号18AYY026)的研究成果之一。

摘　　要：就在十余年前,多篇论文评介了计算技术在词典编纂中的应用研发情况(尤见于Rundell&Kilgarriff 2011)。这些论文展示了词典编纂过程如何在某种程度上实现了自动化,并对完全自动化道路上可能取得的更多进展进行了预测。文章首先简述2011年的前沿技术,然后梳理迄今所取得的进展。对早期论文所做出的预期进行了回顾。在被称作“后期编辑词典编纂”的模式中,人类词典编纂者的角色是后期编辑,即对自动生成并转入词典编写系统的词典初稿进行评估优化。但这些已取得的进展目前皆受到怀疑,因为ChatGPT等大型语言模型似乎有望绕过眼下所使用的技术。文章通过ChatGPT生成词典文本的诸多实验,探讨了这些人工智能工具取代目前词典编纂前沿技术的可能性。Over a decade ago,many papers reviewed developments in the application of language technologies to dictionary making.They showed how the dictionary-making process had been automated and speculated on the prospects for further advances towards full automation.Now,it is time to assess what progress has been made.This paper first reviews the state-of-the-art in 2011,and then looks at developments between then and now.Predictions made in earlier papers are reviewed.Several semi-automated projects are reported,showing gradual progress towards a new approach to dictionary compilation.In the model known as“postediting lexicography”,the role of human lexicographers is to post-edit the first draft of a dictionary which has been generated automatically and transferred into a dictionary writing and editing system.All these developments have been called into question by the recent arrival of ChatGPT and similar large language models.Through several experiments using ChatGPT to generate dictionary texts,the potential for AI tools to replace the current state-of-the-art is investigated.

关 键 词：自动化 SKETCH ENGINE 后期编辑词典编纂 ChatGPT 大型语言模型 

分 类 号：H061[语言文字—语言学]