SynHuaAI:面向华商数智校园建设的人工智能大模型
DOI:
CSTR:
作者:
作者单位:

1.广州华商学院&2.amp;3.#160;4.人工智能学院;5.广州华商学院 

作者简介:

通讯作者:

中图分类号:

基金项目:

国家自然科学基金项目(面上项目,重点项目,重大项目)


SynHuaAI: Artificial Intelligence Large Language Model for the Construction of? Huashang Smart Campus
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    高校使用大模型时面临着大规模数据训练和数据隐私保护的挑战。提出面向华商数智校园建设的本地人工智能大模型SynHuaAI。SynHuaAI具有数据隐私保护性、高效性、泛化性、回答问题准确性的特点。SynHuaAI的数据隐私保护性得益于其本地部署与大规模本地数据训练的模式;高效性与准确性得益于其采用的参数微调技术和面向大模型的知识库语义本文分析技术。泛化性通过支持调用云上的ChatGPT-4 API接口或其他的开源大模型接口访问。描述了SynHuaAI的部署过程和典型应用示范场景。通过原型系统实验结果表明:SynHuaAI大模型可以准确的回答问题,提高了高校师生的工作效率。

    Abstract:

    Universities face challenges in large-scale data training and data privacy protection when using large models.An Artificial Intelligence Large Language Model called SynHuaAI which was constructed for Huashang Smart Campus was proposed in this paper.SynHuaAI has the characteristics of data privacy protection, efficiency, generalization, and accuracy in answering questions.The data privacy protection of SynHuaAI benefits from its local deployment and large-scale local data training mode.The efficiency and accuracy of SynHuaAI benefit from the parameter fine-tuning technology and semantic analysis technology of the knowledge base for large models adopted in this article.Generalization of SynHuaAI is achieved by supporting access to the ChatGPT-4 API interface on the cloud or other open-source large model interfaces. The deployment process and typical application demonstration scenarios of SynHuaAI was also described in detail.The experimental results of the prototype system show that the SynHuaAI large model can accurately answer questions and improve the work efficiency of university teachers and students.

    参考文献
    相似文献
    引证文献
引用本文

徐胜超,吕峻闽,朱文,鲁健恒,黎祥远,廖青,曾志区. SynHuaAI:面向华商数智校园建设的人工智能大模型计算机测量与控制[J].,2025,33(9):237-244.

复制
分享
相关视频

文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2025-02-27
  • 最后修改日期:2025-04-11
  • 录用日期:2025-04-10
  • 在线发布日期: 2025-09-26
  • 出版日期:
文章二维码