Recently, the 2023 Inno China China Industrial Innovation Conference and Peking University Innovation Review Annual Forum was held, which was attended by Ma Chunquan, founder and CEO of Hesi. Peking University Innovation Review, founded by Peking University, is an industry-university-research innovation platform with authoritative academic background and deep industrial resources in China. "INNO CHINA" China Industrial Innovation Conference focuses on data-driven industrial transformation and upgrading, industrial intelligence and industrial transformation and upgrading, enterprise digital transformation and upgrading, industrial service ecological construction and business intelligence integration and development.
At the event site, Ma Chunquan said, "the large amount of data in the field is suitable for the application of large models, but the calculation cost of the current large models is too high, which may affect its performance-to-price ratio in the process of industrialization." There is still a gap in the research of domestic large language models, and it is necessary to strengthen the cooperation of the underlying platform manufacturers. There is still a lot of room for innovation in the fields of informatization, digitization and automation, which can improve productivity in these areas. Industrial chain reconstruction and business model innovation are the core for enterprises to enhance their competitiveness, which will bring more benefits for enterprises. "
The following is Ma Chunquan's speech (with excerpts):
In the domain, the application of the large model is more common, the essential reason is that the large amount of data in the domain gives the large model the opportunity to show its ability. Many areas are still in the process of digitization and do not produce huge amounts of data. Small-size algorithms have been able to support business decisions, and the application of large models will incur a lot of overhead. When we communicate with various big manufacturers, especially when we are doing AI and large language model manufacturers, we have a prominent feeling that the computing power is too expensive, and the problem of low performance and price will be exposed in the process of industrialization of the large model. The large model can be combined with practice in many scenarios, but there is still a big difference in terms of algorithm effectiveness and economy.
The recent launch of OpenAI showed its strong ability, but at present, there is still a considerable gap between domestic manufacturers and OpenAI in the research results of large language models. although many manufacturers have carried out relevant tests and practices, the results are not satisfactory. Therefore, in this issue of the Innovation Forum, we focus on large models and the future application space of digitization and automation.
Five or six years ago, some experts pointed out the problems of big data and small data in the application of large models, and many business areas have not used methods such as linear programming and data model analysis. these methods have great potential to improve in the process of enterprise management. With the development of technology, we are fully committed to the research and application of large model and AI. In the process of business development, we also deeply feel the innovation opportunities brought by the restructuring of the industrial chain. For example, we are engaged in financial revenue and expenditure management, but when the business extends to corporate consumption management, the amount of data is also increasing rapidly. At this time, the demand for data processing and processing becomes more prominent and urgent. In order to reconstruct the industrial chain, we combine the information processing which used to be limited to the enterprise with transactions and e-commerce, so as to transform one business into two or even three businesses. "Management Software + transaction + Service platform" has become the core element of our business model.
Both technological innovation and model innovation can bring benefits. American economist Schumpeter stressed the importance of new elements, new technologies and productivity links, this business model is not a small innovation, but also can bring huge benefits. For example, Apple has reconstructed the value chain of the industrial chain by turning mobile phones into a combination of computers, phones and cameras, turning a few thousand yuan into a business of 20,000 yuan.
In the field of financial digitization, technological innovation and mode innovation are also particularly critical. Enterprises have entered the era of digital tickets, which means that capital flow, negotiable instrument flow and financial information flow urgently need three-in-one and real-time online. In this context, Hesi innovates on the basis of a widely connected ecosystem, a two-wheel drive business model and a self-developed full-link L4 "no reimbursement" solution. Further build an agile financial revenue and expenditure management platform to provide enterprises with intelligent aggregate consumption, expense control reimbursement, receipt and payment management, financial income and expenditure analysis and electronic accounting files and other omni-directional financial digital services Liberate the hands and release people's creativity to help enterprises realize the integration of karma and wealth.
Welcome to the official account of "Hesi" Wechat to download the report.
Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.
CTOnews.com, October 14 (Xinhua) Peter Molineuz (Peter Molyneux), the producer of the "Fables of Ghost" series, has returned to the game field and is currently developing the Project MOAT project, which will be the work of his life
CTOnews.com, December 5 (Xinhua)-- Apple's Neural rendering Research team (Neural Rendering Research Team) issued a recruitment announcement last week, indicating that it is recruiting a neural rendering research scientist.
Beijing, January 13 (Xinhua) the illegal use of cryptocurrencies in 2022 reached a record $20.1 billion (135.4 billion yuan), according to data from blockchain analysis firm Chainalysis on Thursday. Chainalysi
According to CTOnews.com news on December 6, at the 2023 supercomputing innovation application conference held today, the National Guangzhou Center officially released the new generation of domestic supercomputing system "Tianhe Xingyi", but did not disclose specific performance details. At noon today, Xinhua News Agency released the announcement.