In the rapidly evolving landscape of artificial intelligence technology, how will large-scale models enhance urban governance and services, empowering various industries? Wave of the Sea Cloud has provided its own answer.
On April 2nd, the "Cloud Horizon: Born for 'Data'" Digital Innovation Ecological Conference, guided by the Cyberspace Administration Office of the Shandong Provincial Committee of the Communist Party of China, supported by the Qingdao Big Data Bureau, and hosted by Wave of the Sea Cloud, was successfully held in Qingdao. Focused on serving the new quality of productivity, Wave of the Sea Cloud officially launched its business strategy for the "Wave of the Sea" large model, aiming to embrace the wave of large models collectively through digital ecological innovation communities and to foster new industry formats in collaboration with ecological partners.
As the digital economy continues to evolve, data elements have been placed on par with traditional production factors such as labor, capital, and land, and are now unleashing incalculable value. Enabling data elements to unleash multiplier effects across various industries is the direction pursued by numerous cloud computing enterprises.
In her address, Xiao Xue, the CEO and Chief Engineer of Inspur Group, and Chairman of Inspur Cloud, outlined the company's commitment to a fusion strategy represented by the "cloud-network-edge-end, cloud-intelligence, construction-management-operation" principles, alongside their next-generation industry cloud strategy termed "MEP". They aim to construct Hai Ruo industry mega-models, which are among the first batch of model products in the province to be filed for record. Currently, they have authorized and applied for over 400 patents related to large models and data elements.
Leveraging 7 core cloud centers, 113 regional cloud centers, 502 regions, and 254 data cloud nodes, Inspur Cloud has established the largest distributed computing network nationwide. This network operates vast amounts of heterogeneous data. It not only facilitates data trust circulation based on data but also provides diverse computing power services around models, supporting distributed training and inference of numerous large model nodes.
Furthermore, Inspur Cloud integrates its "cloud intelligence" products to drive digital and real-world integration. They have already introduced practical business scenarios such as official document composition and electronic medical record generation, focusing on addressing pain points to create an integrated industry mega-model plus industry intelligence system. This initiative aims to enhance efficiency and quality across various industries.
Yan Liang Shares Insights
WaveCloud's Precision Targeting in Industry Mega Models
Yan Liang, the General Manager of WaveCloud, asserts that industry mega models, compared to general mega models, can swiftly deliver scenario value, thus enabling rapid implementation within specific industries. WaveCloud's Hai Ruo mega model is tailored for industry application, initially focusing on six major sectors: government, transportation, emergency services, manufacturing, healthcare, and agriculture. It boasts three core advantages: reliability, ease of implementation, and sustainability.
Mega models bring immense value to data elements and are presented in an infrastructure-like manner. WaveCloud plans to invest 5 billion CNY, leveraging the digital innovation ecosystem, to expedite the rapid implementation of the Hai Ruo mega model in 100 cities. This initiative aims to continuously unlock the value of data elements and facilitate the development of new quality productivity.
Sun Siqing's Keynote Share
Currently, Inspur Cloud is accelerating the development of a tripartite open ecosystem with its ecological partners. Sun Siqing, Chief Technology Officer of Inspur Cloud, introduced that the Inspur Hai Ruo large model has achieved key technological innovations by adopting MoE architecture, comprehensive optimization of RAG, and end-to-end security protection. These measures address the concerns of industry users regarding operational costs, accuracy, and security in the deployment of large models. Sun Siqing summarized and refined the "Three-Step Method" for industry deployment of Hai Ruo, aiming to create a tripartite digital innovation community. This community fosters a co-building, co-winning, and innovative operating model, where partners and the ecosystem continuously enhance integrated solutions, products, and delivery capabilities. This approach accelerates industrial upgrading and innovation in formats.
The Wave Cloud Digital Ecosystem Innovation Community Officially Established
In the governmental sector, Wave Cloud actively explores the application of government affairs big models in scenarios such as document writing and grassroots burden reduction, achieving a 100-fold increase in efficiency in grassroots material organization and improving response accuracy to 75%. In the medical industry, Wave Cloud relies on the Hai Ruo Medical Big Model to assist doctors in generating electronic medical records, reducing the time to generate an admission record to just 15 seconds with an accuracy rate of over 90%, thus reducing workload by over 70%. In the manufacturing industry, in promoting new product development, with the help of the Hai Ruo Manufacturing Big Model, the accuracy of target identification has been improved to 95%, and the weekly analysis of patent quantity has increased to 3000 pieces.
In the future, Wave Cloud will work hand in hand with its partners to provide industry customers with trustworthy, easy-to-implement, and sustainable industry big models, accelerating industrial upgrading and business model innovation.
