政务大模型
Search documents
青岛数据集团“四位一体”体系领跑全国政务大模型建设
Xin Lang Cai Jing· 2025-12-28 07:00
Core Insights - Qingdao Data Group, a subsidiary of Huatuo Group, is a key player in the digital economy development of Qingdao, creating a replicable model for national government AI initiatives through its "computing power-data-model-service" integrated operational system [1][9] Group 1: Computing Power - The company has established a computing power scheduling service platform that integrates diverse national computing resources, creating a one-stop platform and a full-stack empowerment system [1] - Currently, it has connected 13 computing centers with over 24,000 PFlops of computing resources, serving more than 300 enterprises, significantly lowering the innovation threshold and application costs for AI companies [1] Group 2: Data Activation - Qingdao Data Group has developed a public data operation platform that collects 5,201 public data directories, covering social data from 50 million enterprises nationwide [3] - The platform has launched nine scenario zones and 157 data products, enabling 247 application scenarios such as financial risk control and rapid claims processing, creating a sustainable data circulation ecosystem [3] Group 3: Model Development - The company has independently developed the "Tongcheng Government Large Model" to meet concentrated demands in the government sector, achieving a 93% accuracy rate in knowledge-based intelligent Q&A and a 70% efficiency improvement in rapid model construction [5] - This model supports over 50 practical scenarios, including "similar case judgment," leading the nation in government intelligence [5] Group 4: Service Implementation - Based on the government large model, Qingdao Data Group has created over 160 practical application scenarios, enhancing urban event handling efficiency by over 30% through collaborative analysis of "small algorithms + large models" [7] - The results have been promoted to over 30 cities nationwide, transforming technical capabilities into urban governance effectiveness, benefiting urban development and improving people's livelihoods [7] Group 5: Recognition and Future Plans - The company has achieved notable recognition, including four software copyrights, two patent applications, and awards such as the third prize in the 2025 National Data Element × Competition [9] - Future plans include further refining the "four-in-one" system, enhancing data supply, technical support, and scenario empowerment capabilities, aiming to be a pioneer in data value and a facilitator of urban upgrades [9]
“彭小捷”成长记
Xin Lang Cai Jing· 2025-12-25 22:30
依托深度优化的政务大模型,"彭小捷"构建了"数据智能中枢+行业智库引擎"双核体系,知识库覆 盖企业开办、业务审批、社保医保等高频事项并实时动态更新。当用户输入"想开个小饭馆""公司 要交社保"等口语化问题时,"彭小捷"能同步进行语义识别与意图判断,从海量政策文件中提取关 键条款,并生成条理清晰、步骤明确的个性化指南,识别准确率高达99%。"这不仅仅是回答一个 问题,更是完成一次服务导航。"徐宁说。 □四川日报全媒体记者 底伊乐 "想在夜市摆摊,咋办手续?"不久前,"90后"彭山人王先生在眉山市彭山区政务服务中心打造 的"彭山服务"智能客服对话框里,发出了这句询问。 大段的回复迅速出现在屏幕上。它准确判断出王先生的需求属于"临时便民服务摊点设置手续", 还细致区分了"集中摊区"与"住商出店"的不同要求。按照指引,王先生在线提交了材料。预审通 过后便有工作人员与其确定相关事项,当天缴费后就拿到了摆摊许可证。"就像有个明白人在手把 手教。"王先生感慨。 这位24小时在岗、秒级响应的"明白人",是眉山市首个基于DeepSeek大模型开发的AI数智人"彭小 捷"。今年2月底上线以来,这位特殊的"政务员"已推动实现智能客 ...
华为云AI CITY战略落地,武汉、石家庄等成智慧城市典范
Guan Cha Zhe Wang· 2025-11-17 15:04
Core Insights - The 27th China International High-Tech Achievements Fair was successfully held in Shenzhen, focusing on the transformation of smart cities through AI and computing power [1] - Huawei Cloud received multiple awards, including "2025 Leading Solution Provider for Smart Cities in China" and "2025 Leading AI CITY Solution Provider" [1][3] - The development of smart cities in China is transitioning from "cloud as the norm" to "AI as the standard" [3] Huawei Cloud's Contributions - Huawei Cloud's AI CITY concept aims to drive urban modernization by integrating AI technology with infrastructure and data resources [3] - The company has partnered with various cities to implement smart city solutions, resulting in numerous awards for cities like Wuhan, Shijiazhuang, and others [1][3] Case Studies of Smart City Initiatives - Wuhan has developed a digital public infrastructure called "Wuhan Cloud," serving 124 units and integrating over 400 business systems [4] - Shijiazhuang's "Kunpeng Government Cloud" has been operational for over 1,000 days with zero incidents, providing services to over 30 units [5] - Xiangyang has implemented an AI traffic light system, improving traffic efficiency by over 15% [5] - Yancheng has created a city management platform that enhances governance efficiency through AI and data integration [6] - Karamay has achieved 95% of its government systems on the cloud, introducing a digital government assistant named "Xiao Ke" [7] - Bao'an and Qianhai are establishing an innovation center for embodied intelligence robotics, focusing on cutting-edge technology [8] - Huadu District in Guangzhou is leveraging "digital twin" technology to enhance governance and public services [9] - Changping District in Beijing is developing a deep-thinking AI governance model, integrating various AI capabilities [10] Market Position - Huawei Cloud has served over 800 government cloud projects in China since 2012 and holds a 23.3% market share, ranking first in the government cloud market [10][11]
思特奇:公司积极参与政务大模型统一服务平台建设与部署
Zheng Quan Ri Bao Zhi Sheng· 2025-11-11 10:13
Core Viewpoint - The company is actively engaging in the construction and deployment of a unified service platform for government large models, focusing on core needs such as government services and social governance [1] Group 1 - The company is concentrating on intelligent Q&A, assisted processing, and direct policy services to enhance government service efficiency [1] - The initiative aims to promote resource reuse across different levels and regions, contributing to inclusive and shared government services [1] - The company emphasizes its role in supporting government governance through technological advancements [1]
致远互联签约政务大模型平台及应用场景建设项目
Zheng Quan Shi Bao Wang· 2025-10-31 06:20
Core Insights - Zhiyuan Huitong has signed a project for the construction of a government large model platform and application scenarios in a certain district, aiming to create the first nationwide "government large model service platform + application scenario ecological co-construction" innovative paradigm [1] Group 1 - The project is a benchmark initiative for government large models [1] - The goal is to establish a comprehensive service platform and ecosystem for government applications [1]
浪潮科技一项目入选国家级政务大模型标杆案例
Qi Lu Wan Bao· 2025-06-30 08:09
Core Insights - The "Shanxi Provincial Taxation Bureau Tax and Fee Analysis Model Project" was recognized as a typical case of government large models by the China Academy of Information and Communications Technology, supported by Inspur Technology [1] - This project is the only solution focused on tax governance among the 22 benchmark cases, demonstrating a successful transition from laboratory to practical application in government services [1] - The project utilizes Inspur's "Yanyu" industry large model as its technical foundation, achieving a semantic conversion accuracy rate exceeding 95%, significantly enhancing response efficiency to taxpayer inquiries [1] Data Integration and Efficiency - The project addresses the issue of fragmented tax inquiry information across multiple channels, creating the first provincial-level tax "data lake" in China for real-time aggregation of all inquiry information [2] - The system can transcribe voice calls into structured text and automatically label key emotional tags, reducing manual review and data entry workload [2] - A dynamic heat map monitors inquiry distribution in real-time, and a trend prediction engine can identify potential issues seven days in advance, improving taxpayer satisfaction in Taiyuan [2] Competitive Advantage and Policy Optimization - The project's core competitiveness lies in the deep collaboration between industry knowledge and AI reasoning capabilities, addressing potential issues with general models in specialized fields [3] - A comprehensive knowledge service platform incorporates a vast amount of tax regulations and judicial precedents, enhancing the model's professional capabilities [3] - The project has identified 15 ambiguous policy clauses, leading to optimized regulations and a reduction in inquiry repetition and case transfer rates, while promoting a shift from passive response to proactive governance [3]