检索规则说明:AND代表“并且”;OR代表“或者”;NOT代表“不包含”;(注意必须大写,运算符两边需空一格)
检 索 范 例 :范例一: (K=图书馆学 OR K=情报学) AND A=范并思 范例二:J=计算机应用与软件 AND (U=C++ OR U=Basic) NOT M=Visual
作 者:周艳秋 毛宇凯 陈蕾[2] 句芳[3] ZHOU Yan-qiu;MAO Yu-kai;CHEN Lei;JU Fang
机构地区:[1]清华大学经济管理学院,北京100084 [2]首都经济贸易大学财政税务学院,北京100070 [3]内蒙古农业大学经济管理学院,内蒙古呼和浩特010011
出 处:《东南大学学报(哲学社会科学版)》2025年第1期65-73,151,共10页Journal of Southeast University(Philosophy and Social Science)
基 金:国家社会科学基金青年项目“数据资产课税税基评估理论与实践运行体系优化研究”(24CJY031)成果之一。
摘 要:多模态大语言模型(MultiModal Large Language Models,MM-LLMs)作为生成式人工智能发展的新方向,正在快速变革税收征管领域,其所具备的跨模态、专业化、高效化等优势,可显著提升税务数据分析与预测精准度、税收风险识别与管理水平、税务咨询与服务质量等。税务部门须顺应新趋势,探索如何推进MM-LLMs在税收征管领域的深度实践应用,从而推动税收征管现代化。基于MM-LLMs在税收征管中的可行性和关键技术,构建和提出税收征管领域专用MM-LLMs架构和具体实现路径,并以自然人办税服务平台的优化升级为例,分析该模型架构在税收征管中的应用场景,以期能够为MM-LLMs在税收征管领域的应用提供参考。Multi-Modal Large Language Models(MM-LLMs)is a new development of generative artificial intelligence that is rapidly transforming tax administration.Their advantages in cross-modal capabilities,specialization,and efficiency can significantly enhance the accuracy of tax data analysis and forecasting,tax risk identification and management,as well as the tax consulting and services.Tax authorities need to adapt to this new trend by exploring how to promote the application of MM-LLMs in this field,thereby advancing tax administration modernization.Based on the feasibility and key technologies of MM-LLMs in tax administration,we construct a specialized MM-LLMs architecture and propose pathways for the tax administration.Using the optimization and upgrading of individual taxpayer service platforms as a case study,we analyze this model application in the scenarios of tax administration,aiming to provide a reference for the application of MM-LLMs in this field.
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在链接到云南高校图书馆文献保障联盟下载...
云南高校图书馆联盟文献共享服务平台 版权所有©
您的IP:216.73.216.172