The Yangtze River Delta Artificial Intelligence Industry Chain Alliance is bursting this bubble

Due to the popularity of Chatgpt, big models have sprung up in the past period. According to incomplete statistics, at present, 79 large models with parameters of more than 1 billion have been released in China, and the "100-model war" has sounded the horn.

What exactly are the capabilities of these various large models? At the Yangtze River Delta Entrepreneur Alliance Industrial Digitalization Summit held today, Xu Jiajia, vice president of iFLYTEK, suggested in an interview with reporters that it can be evaluated through the general cognitive intelligence big model evaluation system. His other identity is the vice chairman and secretary general of the Yangtze River Delta Artificial Intelligence Industry Chain Alliance.

"The manufacturer said that how his large model ability is, which can be measured through the evaluation system." Xu Jiajia said, "This can better regulate and promote the healthy and orderly development of our entire industry, so that we can solve problems on a down-to-earth basis and prevent the emergence of bubbles in the large model industry." ”

At the summit, the "Yangtze River Delta Artificial Intelligence Industry Ecological Report" jointly researched and compiled by the Yangtze River Delta Entrepreneur Alliance, the Yangtze River Delta Artificial Intelligence Industry Chain Alliance and iResearch was released. The "Report" shows that the development of artificial intelligence industry in the Yangtze River Delta region ranks first in the country, artificial intelligence enterprises account for more than 30% of the national total, and the artificial intelligence industry layout of the three provinces and one city has its own characteristics: Shanghai gathers the world's top artificial intelligence enterprises and scientific research institutions to create key industrial clusters such as smart chips and smart hardware; Anhui has formed a development pattern of "one core, two places and multiple points" led by "China Sound Valley"; Based on the advantages of the Internet of Things and manufacturing, Jiangsu has created an intelligent manufacturing "AI+" industrial ecology; Zhejiang's "AI+" industrial chain structure featuring e-commerce, security and commercial intelligent terminals is becoming more and more perfect.

There is no point in asking "which big model is stronger"

At the Yangtze River Delta Entrepreneur Alliance Industrial Digitalization Summit held today, the National Key Laboratory of Cognitive Intelligence, the Chinese Academy of Sciences Artificial Intelligence Industry-University-Research Innovation Alliance and the Yangtze River Delta Artificial Intelligence Industry Chain Alliance jointly released the general cognitive intelligence large model evaluation system, and Wang Shijin, executive vice president of iFLYTEK Research Institute and deputy director of the National Key Laboratory of Cognitive Intelligence, invited the Yangtze River Delta entrepreneurs present to participate in the scenario application of the large model.

On May 6, at the "Spark" conference of artificial intelligence cognitive large model held by iFLYTEK, Chairman Liu Qingfeng mentioned that iFLYTEK took the lead in designing a general cognitive large model evaluation system through the National Key Laboratory of Cognitive Intelligence, and discussed with the Chinese Academy of Sciences Artificial Intelligence Industry-University-Research Innovation Alliance and the Yangtze River Delta Artificial Intelligence Industry Chain Alliance to form 481 subdivided task types covering 7 categories, and promote the development of large models in the scientific evaluation system.

In response to the hotly discussed question of "which domestic model is stronger" on the Internet, Xu Jiajia believes that it is meaningless. "Today's cognitive big model is just starting to start, and it is still in the process of rapid growth and iteration, even ChatGPT and GPT-4 have many questions that are not well answered, and if we just find some single examples to prove which system is strong and weak, I think it is meaningless."

What kind of comparison is meaningful? At the press conference on May 6, Liu Qingfeng said that he would pay tribute to and learn from OpenAI, and at the same time, he should quickly catch up and strive to surpass.

"When we pay tribute to and learn from OpenAI, while catching up quickly and striving to surpass, we believe that we first need a scientific and systematic evaluation system." Xu Jiajia said.

"What we especially want to say is that if we want to make this intelligent emergence of cognitive intelligence really solve the rigid needs of society, rather than simply making a publicity, we must establish a real, down-to-earth, scientific and systematic evaluation system." Xu Jiajia explained, "We must use this scientific and systematic evaluation system to tell us what extent our technology has come, where the next development should go, tell us whether the technology is far from practical and practical, and what work remains to be done." ”

It is necessary to be vigilant against high-level duplicate construction

In the interview, Xu Jiajia mentioned that the field of artificial intelligence is experiencing a high level of repeated construction. Artificial intelligence itself is an emerging industry, and infrastructure such as intelligent computing centers also has high scientific and technological content, which also means that the investment in high-level repeated construction is greater, and the consequences are naturally more serious.

In September 2021, the "Research Report on the Development of a New Generation of AI Computing Power Infrastructure" released by the National Industrial Information Security Development Research Center also pointed out that the construction of artificial intelligence computing power infrastructure in mainland China faces a series of problems and challenges, including the lack of clear market planning, duplicate construction, and confusing pricing by participants, etc., which urgently needs the overall guidance of the state from the source and the establishment of unified industry standards.

"The problems brought about by marketization must be solved in a market-oriented way." "Some areas may have a high level of repeat investment in a short period of time, especially in its infancy, but through fierce market competition, it will eventually precipitate the leading companies in the industry," Xu said. ”

However, Xu Jiajia also said that a major reason for the high level of repeated investment is information asymmetry. When talking about this, Xu Jiajia mentioned the Yangtze River Delta Artificial Intelligence Industry Chain Alliance. Up to now, the alliance has held a number of industry summits, selected outstanding figures and excellent cases, and released the Yangtze River Delta artificial intelligence industry ecology report, strengthening exchanges and cooperation among entrepreneurs. He believes that mutual exchanges within the industry can level out information.

The frequently mentioned Yangtze River Delta Artificial Intelligence Industry Chain Alliance was established in 2021, led by iFLYTEK, with the participation of major enterprises and academic institutions in the field of artificial intelligence in the Yangtze River Delta, with four groups: chip computing power group, core algorithm group, intelligent scene group and industrial ecology group, including Suning.com, Hangzhou Hikvision, Fudan University, Zhejiang University, Nanjing University, University of Science and Technology of China, a total of more than 70 enterprises and academic institutions.

Previously, the Yangtze River Delta has established 10 industrial chain alliances in two batches, including artificial intelligence, digital health, superconducting industry, high-end medical care, green transportation, cable and optical cable, smart retail, software information, smart city, and digital creativity.

Today, the industrial chain alliance has "expanded" again, adding 4 new ones, including the Yangtze River Delta Digital Technology Application Service Industry Chain Alliance led by Shanghai, the Yangtze River Delta Transparent Commercial Ecological Industry Chain Alliance led by Jiangsu Province, the Yangtze River Delta Digital Drive Innovation Biomedical Industry Chain Alliance led by Zhejiang Province and the Yangtze River Delta Industrial Internet Industry Chain Alliance led by Anhui Province.

Column editor-in-chief: Kong Lingjun Text editor: Kong Lingjun Title picture source: Interviewee Courtesy photo

题图：长三角企业家联盟联合发布《通用人工智能合肥宣言》。      

Source: Author: Chen Shuyi