Shulou(Shulou.com)11/24 Report--

In recent years, artificial intelligence has become one of the most active frontier fields of global digital technology innovation, a new track of digital economy and a new hot spot of international competition. At the same time, the public still have doubts about artificial intelligence, so how to control the speed and direction while stepping on the throttle, so as to build a safe and credible artificial intelligence industry ecology, and explore AIGC (generative artificial intelligence) based on trusted data has become an important direction of artificial intelligence breakthrough in the future.

The recent release of the world's first business search model-enterprise search "know the other alpha" is both a useful exploration of trusted artificial intelligence. This large model is based on the results of large-scale pre-training after the accumulation of trusted data in the business query field in the past ten years. The innovative products that will be launched later will be jointly built through the large model and enterprise credit database, and AI technology will be used to enable enterprises to trust big data's innovation, create secure and reliable artificial intelligence products, and provide users with more convenient and accurate business information query services.

(photo: enterprise search business search big model "know the other Alpha")

Why is the first business search model in the industry an enterprise search?

It is reported that during the Jiangsu Development Conference on May 20 this year, Chen Deqiang, the founder of the company, brought the world's first business search model, "know the other Alpha", and launched a live roadshow, which attracted the attention of the industry.

(photo: at the Jiangsu Development Conference on May 20, 2023, the company checked and released the world's first large-scale business search model, "know the other Alpha")

Chen Deqiang said that the company has an enterprise trust database covering nearly 500 million enterprises around the world. The core resources of business information inquiry service are related data resources, such as industrial and commercial information, litigation information and so on. Moreover, the data coverage should be broad enough to have not only all the data of China, but also the data of other countries in the world. In addition, in the highly professional fields such as enterprise standard information and enterprise dispatching, it is also necessary to establish a special database. In order to introduce AIGC technology into the field of business information query service, the most critical action is to use enterprise trust data set to train related models. The larger the amount of data and the higher the quality of the data, the more accurate the final training model. Compared with other large language models which mainly use open Internet data sets, it is more difficult to obtain professional data such as industry and commerce, judicature and so on. It can be said that the credit data resources of 500 million enterprises covered by enterprise inspection are the core barriers to business search services, bringing together 80 industrial chains and 8000 industries in the domestic market. and a large amount of real-time market industrial and commercial information, risk publicity, intellectual property rights, credit reports, equity relations and other 300 + dimensions of enterprise credit data. These data not only build a high competition barrier for enterprise inspection, but also lay a solid data foundation for its training to know the big model of alpha business search.

It is understood that the AI algorithm model checked by enterprises is the leading in China, and it won the "highest Award of China's Intelligent Science and Technology" Wu Wenjun artificial Intelligence Science and Technology Award in May this year. At the algorithm level, enterprise search has a profound accumulation of AI technology. Over the years, enterprise search has realized automatic and intelligent data analysis and text mining in massive global multilingual texts through the use of AI technologies such as deep learning and natural language processing (NLP). It can further achieve deep semantic analysis and provide users with more accurate semantic retrieval services. In the domain of the pre-training model, which is rich in data resources, enterprise search has a strong technical accumulation, and the Zhibi Alpha business search model released by the enterprise has realized completely independent intellectual property rights.

A variety of products, services and application scenarios. Users are directly faced with products, not technology. Therefore, if we want to apply the AIGC underlying large model technology to the business inspection service, we must package the technology into a product that is convenient for users to use. Moreover, according to the needs of different users, it is necessary to build targeted products, and then form a relatively complete product matrix. In the way of product matrix, to provide one-stop service for business inquiry users. At the product level, enterprise check has constructed a perfect product matrix for different user groups, specifically: for enterprise users, enterprise check provides solutions in scenarios such as accurate expansion, enterprise rating, due diligence, risk control, judicial investigation, public opinion monitoring and supply chain management through customized services to assist enterprise users to improve enterprise portraits, information cross-verification and find partners. For individual users, enterprise check provides solutions for investment and financing, job recruitment, risk determination and other scenarios through cloud platform integrated multi-dimensional data, so as to see through the enterprise equity structure for individual users and avoid the credit risk in the process of enterprise identification. For public sector users, enterprise search data is not only an important supplement to the central bank's official credit channels, but also an important reference for local government policy formulation, social credit system construction, investment attraction, screening policy support objects, and enterprise credit supervision.

Why hit the manufacturer to check the version of ChatGPT?

With the accumulation of data and the increase of product functions, there are many problems and demand pain points in the field of commercial inspection, mainly in two aspects:

Users are still in the keyword search stage, the platform can not well understand the complex business needs of users. At present, most of the business search platforms are essentially search engines in the field of enterprise credit information. Users search the corresponding enterprise or risk information in the enterprise credit database through keywords, so it is difficult to express the complex and structured needs of users. For example, when a school canteen wants to know the food safety risk of the supplier in the bidding process, the previous commercial inspection product is usually to enter the name of the company participating in the bid, and then jump to the main page of the corresponding company. then, by looking up the administrative penalty in the company's "business risk" or the "food safety" content in the "business information", we can determine whether the company has food safety risks, with many steps. And not friendly to new users. If the user also wants to know whether there have been any major security incidents in other catering companies run by the enterprise legal person? Is there any affiliated enterprise accompanying this bid? Similar to these more in-depth needs, it is difficult to quickly satisfy users by simply retrieving information. Even if the demand can be achieved, users also need to have relevant professional knowledge, and can skillfully master the search functions of the business search platform. After many steps of relatively tedious operation, they can finally get more in-depth and perfect information about the enterprise. However, for the vast majority of new users, the threshold for use is too high, which is not conducive to opening up information access in a wider range.

In addition, the search engine model, business search platform feedback to users is a large number of basic data, rather than direct answers. If it is a large group company, there may be thousands of basic enterprise trust data, such a huge number of page views is not a small burden for users. This kind of business information query service is still in the tool stage and can not be called a powerful business assistant. At the same time, Enterprise Checkout has hundreds of products and services, such as checking enterprises, checking bosses, checking risks, checking tenders, credit big data, risk big data and so on. It is very difficult for users to learn and master systematically. Based on the alpha model, enterprises can skip the tedious retrieval steps and fully release the ability to check various products by way of dialogue. To provide users with "integrated, easy to understand, with high value" business information services.

However, if users use ChatGPT and other large language models to search for high-value corporate credit data, they will find obvious problems: due to the lack of professional database support, the industrial and commercial and credit data searched through ChatGPT all come from public Internet data, which can not guarantee the accuracy of the data. At the same time, some high-quality and commercial corporate credit databases are not open to them. Without the support of professional database, the analysis of ChatGPT and other large language models in the field of commercial inspection is "making bricks without rice", or even "making something out of nothing". On the other hand, the Alpha business search model is an in-depth training based on searching all the reliable data, which can provide users with professional enterprise trust data and diversified analysis results.

AIGC + enterprise credit database, business search service mode evolves again.

As artificial intelligence enters the "AIGC era", Chen Deqiang believes that this will completely change the mode of querying and using data. "after the scale of industry data has reached a certain order of magnitude, the way of data query has changed. AIGC + enterprise credit database can make full use of the data." According to reports, the Zhibi Alpha business search model currently released by the company is the world's first large business query model, which is trained based on the global enterprise credit data covered by the enterprise query, compared with the traditional business search platform. the subsequent products based on the knowing Alpha business search model have achieved changes in the following three aspects:

In the aspect of human-computer interaction, the use of natural language dialogue can be realized, that is, complex query steps can be completed. If users want to query a certain business data, it is no longer limited to keyword search, but can use a natural language description to put forward their own needs and reduce the user threshold. For example, when users want to make a superficial adjustment to a company, they can ask, "what is the industry status of a company? what are the competitors?" The knowing Alpha model understands the requirements from the user's description and "deconstructs" the requirements into corresponding instructions. In this way, when users know a company, they are faced with an alpha model that is like a professional business research assistant, rather than an unintelligent tool.

In the technical aspect, the second-level response is realized according to the needs of users. According to the semantics of the user, the alpha model can retrieve and check the enterprise trust data, and then present the results of "collation and induction" to the user. In this case, the user no longer gets a pile of piecemeal information, but a complete answer. In order to improve the response speed, knowing that the alpha model has been fully connected to the enterprise search supercomputing platform, we can complete the cumbersome steps such as query, browsing, summary, structured output and so on in seconds.

A new "multi-round dialogue" function has been added to make the business inquiry platform have the ability of logical thinking. The "multi-round dialogue" function of knowing the other alpha model is a bright spot that this model is superior to the previous check dialogue AI. With this ability, the Zhibi Alpha model can guide users to conduct more in-depth analysis step by step through multiple rounds of dialogue. In multiple rounds of conversations, users can put forward new instructions on the basis of the results already obtained, so as to ask more in-depth questions. In this way, the Zhibi Alpha model can not only become the user's assistant, but also be the "guide" to guide the user to find the answer step by step.

According to the big data Industry Development Research report on Enterprise Credit of China Institute of Information and Communication, enterprise inquiry is in the first place in the industry in terms of scale index such as user scale and data volume. Product user stickiness, enterprise check per capita use time, per capita daily start-up times are also higher than the general level of the industry. In addition, in terms of industry penetration, the penetration rates of the legal profession and pan-financial industry are 50.1% and 41.2% respectively, both of which maintain the leading position in the industry and rank first.

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.