news

The 9th CAAI National Conference on Big Data & Social Computing (BDSC2024), hosted by the China Artificial Intelligence Association and co-organized by the Social Computing and Social Intelligence Committee and Harbin Engineering University, was recently held.

The conference invited heavyweight scholars in interdisciplinary fields such as artificial intelligence, social computing, information science, ethics, social science, systems science, and policy science to give keynote speeches. Liu Yong, co-founder and executive vice general manager of Zhongneng Shibei, was invited to attend the conference and give a report entitled "The Application of Big Model Technology Empowers New Quality Productivity of Energy and Power" as an expert in energy and power AI+ big model application research.

Rather than pursuing the “size” of the model, the Shibeiyun Energy Large Model focuses more on application value

The advent of the big model era has greatly improved the performance of artificial intelligence in natural language processing, image recognition, sound analysis and other fields, and has also brought a series of impacts and opportunities for the development of intelligent society. In the energy and power sector, AI big model technology has also been vigorously promoted, and a series of industry-level applications have emerged.

As an industrial intelligent product and service provider focusing on "Industrial AI+", CyberWing helps industrial enterprises, especially energy and power enterprises, to transform and upgrade digitally through its self-developed "CyberWing" brand series of products. Faced with the core pain points of building enterprise-level large models, such as high implementation costs and inability to "grow" on their own, Liu Yong shared the overall architecture and deployment plan for building enterprise-level energy large models through rich industry successful practices.

He pointed out: "The application of Shibeiyun Energy's large model CyberwLLM in the industrial field does not pursue the size of the model, but focuses on the applicability and professionalism of the model. The industrial field prefers to use large language models with small parameters for training and knowledge base enhancement for specific scenarios." He emphasized that the larger the model, the better, but it should be appropriate, and at the same time, attention should be paid to the granularity of the model, which should not be too coarse or too fine, so as to achieve system synergy.

When discussing the costs and benefits of large models, he suggested: “Enterprises should reasonably control costs while maximizing the value of large models.

Large model products are a "powerful assistant" to make energy management smarter

Talking about the changes brought about by large model technology, Liu Yong pointed out: In the field of energy and electricity, the application of Shibeiyun Energy Large Model will not replace traditional knowledge and skills, but as a "powerful assistant" of smart power plants, it can better promote the upgrade and optimization of existing digital systems. He encouraged people inside and outside the industry to embrace AI technology, vigorously promote the application of large model products, and let technology promote the further development of society and industry.

For example, in response to the diversity and specialization of knowledge in the power industry, the scattered storage of knowledge, the poor inheritance of knowledge, the lack of knowledge, and the retention of valuable knowledge and experience in solving problems in professional fields, the power knowledge service engine based on CyberwLLM can build a unified natural language interactive knowledge carrier, provide a question-and-answer application knowledge service engine, and fully apply, share and transfer knowledge with a large model unified knowledge carrier as the core.

In the inspection scenario, when an abnormality is found during the inspection, the inspection intelligent interactive question and answer based on CyberwLLM can make intelligent auxiliary decision-making suggestions based on historical experience and intelligent algorithm support, combined with equipment operation data, including precision inspection suggestions, equipment maintenance decision suggestions, emergency disposal suggestions, and operation mode adjustment suggestions. Based on the system interaction capability of the large language model, through dialogue with the model, inspection route planning, inspection standard query, inspection result query, inspection abnormality problem statistics and retrieval, and statistical chart generation can be completed.

In addition, the capabilities of Shibei Cloud Energy Big Model also play a role in multiple scenarios, including maintenance enhancement based on CyberwLLM, technical supervision, dynamic decision-making of production command, centralized control alarm signal analysis, power station equipment signal modeling, Eventlog event anomaly identification, power station monitoring and early warning processing, energy storage power station fault signal processing, power station safety supervision and management, enterprise general document assistant, smart contract assistant, functional training assistant, meeting minutes assistant...

Overall, the Shibei Cloud Energy Big Model helps enterprises optimize grid operation and energy management through intelligent analysis and prediction, supports the efficient integration of green energy, and drives the industry to develop in a low-carbon, intelligent, and efficient direction. At the same time, it provides technical support and innovative solutions for achieving the “dual carbon” goals.

Source: China.com

(Source: Financial Information)

For more exciting information, please download the "Jimu News" client from the application market. Please do not reprint without authorization. You are welcome to provide news clues, and you will be paid once adopted.