LinYun:林业领域智能问答与决策支持的大型语言模型研究OA
LinYun:a domain-specific large language model for intelligent question answering and decision support in forestry
林业知识具有高度的专业性,涵盖范围广泛,并对法规及实践标准高度敏感.针对目前通用大语言模型在林业场景中存在的知识缺口、术语歧义以及事实性错误等问题,提出"数据合成-模型训练-系统化评测"一体化框架,基于通用基座模型进行领域指令微调,得到面向林业领域适配的模型林云(LinYun).实验结果表明,LinYun在林业相关任务上显著优于同规模的通用模型,在部分任务上接近甚至超过更大规模模型的表现.
Forestry knowledge is highly specialized and broad in scopeand is particularly sensitive to regulations and prac-tical standards.To address the knowledge gaps,terminological ambiguities,and factual inaccuracies that general-purpose large language model(LLM)exhibit in forestry scenarios,an integrated framework of"data synthesis-model training-systematic evaluation"was proposed.Based on a general base model,domain-specific instruction fine-tuning was con-ducted to obtain LinYun,a domain-adapted model for forestry.Experimental results demonstrate that LinYun significantly outperforms general-purpose models of comparable scale in forestry-related tasks,and in some tasks approaches or even surpasses the performance of much larger models.
赵维康;王浩宇;华净;刘耀兵;王秀娟;康孟珍
中国科学院自动化研究所多模态人工智能系统全国重点实验室,北京 100190||中国科学院大学人工智能学院,北京 100049中国科学院自动化研究所多模态人工智能系统全国重点实验室,北京 100190中国科学院自动化研究所多模态人工智能系统全国重点实验室,北京 100190北京江恒数智生态科技有限公司,北京 102628中国科学院自动化研究所多模态人工智能系统全国重点实验室,北京 100190中国科学院自动化研究所多模态人工智能系统全国重点实验室,北京 100190||中国科学院大学人工智能学院,北京 100049||澳门科技大学创新工程学院,澳门 999078
信息技术与安全科学
林业大模型指令微调数据合成LoRA人工智能
forestry large language modelinstruction fine-tuningdata synthesisLoRAartificial intelligence
《智能科学与技术学报》 2026 (1)
83-93,11
国家自然科学基金项目(No.62076239)The National Natural Science Foundation of China(No.62076239)
评论