摘要
当前大语言模型在口语和书面语的理解与生成方面已展现出卓越能力,但基于“视觉‒空间”模态的中国聋人手语智能仍处于发展起步阶段,其瓶颈在于高质量手语资源供给不足。当前,中国手语智能资源建设的根本问题是语言资源建设理念、语言信息规范体系与跨领域协作机制相对滞后,主要表现在5个方面:宏观标准断层,标注体系缺口,语料资源瓶颈,建模技术局限,应用生态滞后。单一领域的突破无法补齐体系建设的短板,必须多层面协同建设,以期建成“语言‒技术‒认知‒社会”四维联动的手语智能资源体系:手语语言学理论层面,以系统化描写促成规范化标注;工程技术层面,依靠标准化平台强化管理;认知神经层面,从行为模仿走向神经对齐;社会应用层面,建设聋人主位的伦理化体系。
Abstract
Large language models have demonstrated exceptional capabilities in understanding and generating both spoken and written language. However, AI for Chinese Sign Language—based on the “visual‒spatial” modality—remains in its early stages of development, constrained by a shortage of high-quality resources. Resource development is hindered by inadequate planning, linguistic information standards, and interdisciplinary coordination. These weaknesses manifest in five areas: fragmented overarching standards, incomplete annotation systems, limited corpora, modeling constraints, and an underdeveloped application ecosystem. Breakthroughs in a single domain cannot address the shortcomings in system development. Coordinated development is therefore needed to build an integrated resource system spanning language, technology, cognition, and society. Priorities include systematic linguistic description to support annotation guided by explicit conventions, standards-based platforms for resource management, a shift from behavioral imitation toward neural alignment, and an ethical framework grounded in Deaf perspectives and agency.
关键词
手语 /
聋人 /
语言智能 /
语言数字资源 /
语料库
Key words
sign language /
deaf people /
language intelligence /
digital language resources /
corpus
王子晗, 郑 璇.
中国手语智能资源建设的现状与思考[J]. 语言战略研究. 2026, 11(6): 63-73 https://doi.org/10.19689/j.cnki.cn10-1361/h.20260606
Wang Zihan and Zheng Xuan.
Sign Language AI Resource System in China: Status and Strategies[J]. Chinese Journal of Language Policy and Planning. 2026, 11(6): 63-73 https://doi.org/10.19689/j.cnki.cn10-1361/h.20260606
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
基金
国家社会科学基金青年项目“残障群体的数字化生计研究”(25CSH021),国家语委“十四五”科研规划2022年度省部级重点项目(中青班项目)“语言服务视域下的手语数字人能力评估与治理研究”(ZDI145‒67),北京外国语大学2023年度基本科研业务费学生科研创新项目“面向计算的多语言短文本话题结构研究”(2023JX083)。