2024 listening to the digital age is a joint in-depth interview box of the cio era and the new capital innovation institute, focusing on exploration and sharing in the area of digital transformation. An expert from the industry will be invited to share best practices on digital transformation in various fields and to discuss new developments in digitalization
This time, 2024 listening to the digital age column, a series of interviews with the 360 group's technical vice-president, technology centre, and cloud applications officer yin ying ui, focused on the opportunities for change, change strategies, core technologies, innovation products and the opportunities and challenges facing the new generation of technological transformations. Through interviews, the challenges and opportunities of corporate knowledge management in the age of large models were explored, as was how the 36 billion square cloud smart document cloud, through technological innovation, helped enable firms to harness knowledge assets in big amounts of information to achieve efficient, secure, intelligent knowledge management。
Interviewer: yin woo, technical vice-president, technology centre, cloud applications officer, 360 group fai
Presenter: co-founder of the cio era and secretary-general of the new capital innovation institute
Liu jing: first of all, please introduce yourself and the company profile of the 360 group。
Yin woo fai: i am the technical vice-president of the 360 group and head of technology centre, cloud applications, responsible for the entire technology medium and digital operations. As a pioneer in the digital transformation of the country, the 360 group operates in various areas throughout the country. Of these, 36 billion are fully-owned subsidiaries acquired by the 360 group, initially based on enterprise cloud dishes, and have now developed into leading firms providing smart solutions for enterprise-level documentation。
Liu jing: what kind of knowledge management programmes do you think businesses need in the age of big models
Yin woo-hui: in today's age of large models, corporate knowledge management is becoming an indispensable “navigational system” that leads enterprises forward in the knowledge ocean。
The enterprise's quest for km is no longer limited to upgrading a single tool, but to a transformation at the corporate strategic level. Individuals believe that knowledge management programmes adapted to the age of large models should have four core features: big data carrying, intelligent knowledge extraction, accurate knowledge insight and supply, and rigorous knowledge security. These four core features are not only about the upgrading of tools but also about the inherent requirements of corporate strategic transformation。
Therefore, firms need to actively embrace the power of change brought about by large models, moving from passive acceptance of knowledge to active stewardship of knowledge, from fragmentation to seamless alignment, from static to continuous innovation, in order to remain ahead of intense market competition. This series of shifts will better enable firms to remain competitive and lead in a changing business environment。
Liu jing: please introduce the intelligent documentation cloud scheme, design concept and core functions。
Yin woo fai: in the current age of large models, enterprises need to cope with massive information flows and rapid and iterative rates of knowledge updating. As one of the first nationally certified large models in the country, the “360 brains” (i. E. The smart document cloud program) is the solution that has been introduced to meet the needs of this era。
The design concept of smart document clouds is to help enterprises improve knowledge management and better respond to the challenges of the age of larger models. We expect to help clients introduce large model capabilities in a less costly and safer way to maximize the value of ai。
The smart document cloud programme, launched from 36 billion square clouds, is a knowledge management programme for industry-wide clients, centred on helping enterprises to aggregate and process non-structured data, such as documents, pictures, audio and video, to facilitate efficient operations. In order to achieve this goal, 36 billion square clouds have created a series of “smart weapons” for businesses, including document cloud knowledge management platforms, document cloud megamodels, ai file assistants, ai cloud files, ai knowledge questions and answers, and ai knowledge search。
The document cloud model is more capable of layout analysis of document-type content, such as paragraphs, tables, pictures, and is more capable of identifying complex content more quickly and accurately than the same product on the market. It is a leading industry model. We believe that the threshold for smart applications will be lowered and the value will be greatly enhanced by the addition of document cloud models。
At the programme level, for small and medium-sized clients, the smart document cloud provides a saas version, which is operational and helps sme clients experience ai at minimal cost. The saas version is also one of the group's strategies to support small and medium-sized enterprises (msmes) in moving up and down the mountain, helping to secure them while increasing the efficiency of their business collaboration。
For medium-sized and large-scale privatized clients, smart document clouds provide versions of different privatization-level parameters, integrated on demand. A unified inventory of in-house non-structured data and asset value extraction, supported by the capability of intelligent document clouds, helped to digitize operations and increase productivity。
Liu jing: what are the unique features of intelligent document clouds compared to other smart knowledge management solutions in the market
Yin yonghui: in the wave of digital transformation, enterprises face the challenges of overloading information, inefficient access and poor decision-making support. To deal with the pain, we launched the smart document cloud program. The programme combines large file models and technology for mass knowledge management to provide efficient and intelligent document management solutions for enterprises。
The smart document cloud program has three main advantages:
Readable: through the capability of large models, in-depth interpretation and summary of document contents。
In a fast-paced business environment, enterprises need to quickly understand and process a large amount of information. Smart document clouds automatically analyse document content through large-model technology, extract key points and generate concise and clear summaries, significantly saving enterprise staff time to read and understand complex documents and improving decision-making efficiency. For example, in the context of contract review, the system has been able to speed up the contract approval process by translating professional legal terms into easily understandable language and ensuring that key provisions are not ignored。
Searched: quick access to answers from mass knowledge files, using the capability of large models。
Enterprises often face the challenge of big data when looking for specific knowledge. Smart document clouds combine text and image search functions through large model techniques, enabling enterprises to quickly locate the required documents or information. This cross-model search capability is very useful for the r & d teams in finding scenarios such as technical information, market sector analysis competitions, etc. The user uploads a picture containing key information, which allows the system to identify and recommend relevant documents, significantly enhancing the efficiency of the work。
Correct: provide accurate matching results using large models。
In the area of client services and technical support, quick and accurate responses to client needs are essential. Smart document clouds learn about user history queries and behaviour through large modeling techniques, allowing for a more accurate understanding of user needs and providing highly matched search results. This intellectual optimization not only increases client satisfaction, but also ensures that enterprises remain ahead of competitive markets through continuous self-learning to improve service quality。
In sum, 36 billion square clouds have improved not only readability and comprehensibility of document contents but also the ability of users to search and obtain information in big files, as well as the accuracy and user satisfaction of search results, through the application of large model technologies. These innovations and breakthroughs have provided strong support for enterprise-level markets and digital transformation。
Liu jing: how can smart document clouds guarantee customer data security and privacy in the complex information security environment of the current generation of ai technologies
Yin woo fai: in the digital age, information security is one of the most important concerns of business clients. The 36 billion square clouds, as industry-led provider of smart document cloud solutions, are aware of the importance of enterprise data security and strictly control business data from three dimensions to ensure the security of enterprise data。
First, we rely on the powerful security model of the 360 group, starting with data cleansing, systems security architecture, controlled training modules, and comprehensive network security, to provide a solid security base for document cloud models. This initiative is designed to ensure that enterprises are able to protect their data in a comprehensive manner in the use of smart document clouds and effectively protect against various security risks。
Secondly, with regard to the control of staff file privileges, we have adopted a sophisticated management strategy to ensure the security of large model output in the rag model. When using ai for dialogue or question-and-answer purposes, a large model is used to respond strictly to documents that are accessible to current users, thereby avoiding unnecessary disclosure and safeguarding the core data of the enterprise。
In addition, we use the ai content identification technology for intelligent content detection and control. Improved security and administrative efficiency of enterprise data through smart labels and intelligent classification capabilities and the achievement of classification and confidential control of documents. We are committed to using the power of science and technology to defend the data assets and privacy interests of every user, every enterprise and provide strong safeguards for the digital transformation of enterprises。
Liu qing: what are the breakthrough techniques used by smart document clouds as the industry's first knowledge management plus large model programme
Yin yonghui: in the wave of digital transformation, enterprises face the twin challenges of information security and efficient knowledge management. The 36 billion square clouds provide enterprises with a secure, efficient and intelligent document management solution based on their advantages in large language models, ocr technology, document generation and customization applications。
Chinese understanding and context capability
“360 brains”, with its strong linguistic understanding and advanced context token processing techniques, is in the first tier of the chinese language model, with a high level of understanding and accuracy of both chinese texts and languages。
Document identification and extraction techniques
Thanks to the research and development power of the 360 ai institute, 36 billion clouds have the country's leading document layout recognition and content extraction techniques. Through the accumulation of advanced ocr techniques and a wealth of practical experience, key information in the document can be accurately identified and extracted, providing efficient information retrieval and knowledge management services to enterprises。
Document generation and customization applications
Using advanced rag technology, 36 billion square clouds were quantified, combining business knowledge and expertise within the enterprise, improving the quality of documentation generated. At the same time, a variety of ai applications have been developed to respond to the needs of different industries and clients and to meet the needs of enterprises for format specifications, specialized documentation, such as official papers, redheads, etc。
Industry experience and privatization deployment
A wealth of industry experience has been accumulated in the 36 billion clouds in the business sectors of business, finance and securities, and the ability to privatize large models of bottom training for deployment. Individualized customization and integration, based on client-specific industry knowledge needs, will help enterprises to apply knowledge deposition and transformation applications in their professional domains and to enhance the relevance and professionalism of overall solutions。
In sum, 36 billion square clouds provide a safe, efficient and intelligent document management solution for many businesses. The programme not only contributes to raising the level of information security of enterprises, but also helps them to achieve efficient knowledge management and provide strong support for the digital transformation。
Liu jing: how many stages have the intelligent document cloud developed
Yin yong-hui: the development of intelligent document clouds has gone through three stages, from enterprise webs to document clouds to intelligent document clouds, each with different functions and features. Technically, we focus on user experience and ease of use, providing simple, visual interfaces and flexible customized services to meet the needs of different enterprises。
In developing intelligent document clouds, we focus on data security and privacy protection. Our entire system was fully privatized and deployed to ensure that data were not available. On the application side, we use the corresponding watermarking and traceability techniques to prevent data leakage from the flow, distribution, sharing and downloading of data. At the same time, through operational techniques, we have ensured that the different departments, organizations and personnel have access only to some of the documents to which they have been assigned, so that they can monitor and monitor the back office. This is not what a big model can do。
In terms of technological breakthroughs, we focus on the integration of large models with business, scenes and tracks. There are problems with the capabilities of large models, which need to be addressed in terms of short-term and long-term memory. Through the use of rag technology, external knowledge is used as the outer brain of a large model, enabling larger models to remember more knowledge. In addition, we have used ai technology to link large models to applications, systems, software and hardware, and to achieve the deep integration of large models with scenarios。
Knowledge management is an important component of the digital transformation of enterprises. Past knowledge management has issues of recollection, difficulty sharing, and unstructured documentation within enterprises is difficult to use. By using business search techniques, we can retrieve documents from titles and abstracts, but search techniques and large models are the product of both times. With big models, we can produce content more efficiently. There are real pains and difficulties, such as the poor controllability of large models and the risk of intellectual hallucinations. We need to address and avoid these problems through technical means such as providing scenarios and rag aids to larger models。
In general, we have a wealth of technological accumulation and industry experience in the area of intelligent document clouds, which, through continuous innovation and refinement, provide enterprises with efficient and intelligent document management solutions to help them better manage and utilize document resources。
Liu jing: how can your company maintain the competitiveness of its products in the face of increasing market competition
Yin yonghui: with regard to the sme client community, we note that they may have relatively limited needs for computing resources. Therefore, we have launched saas version of the smart document cloud product, which allows users to experience and use our services easily and quickly by simply registering and activating accounts. In doing so, we attach great importance to the privacy protection of users, with all privately owned data being encrypted and automatically deleted upon completion of testing to ensure safety. In addition, we provide this segment of the user with an integrated, large model solution for software and hardware integration, i. E., on-site, without complex configuration。
For large and medium-sized enterprises and for specific industries, such as zongxi and government agencies, smart document clouds provide customized service programmes. Depending on their specific needs, we can deploy models of different scales, such as 1 billion parameters, 10 billion parameters or even hundreds of billions of parameter levels, with a range of smart application tools such as ai documentation assistant, ai knowledge query and answer system, ai cloud document management and ai knowledge search engine。
At the same time, in order to meet the requirements of the country's innovative development of information technology, we fully support the adaptation of national production, particularly in the area of mobile offices, with seamless access to the newly launched national production system。
Liu jing: what are the market promotion strategies prepared by the 36 billion clouds in the face of intense competition for large model products, as the technologies of the age of large models are changing
Yin woo fai: one of our strengths is that 360, as the country's largest producer of pc terminal applications and search products, is closer to the user side. In addition, our accumulation in the search field and the copyrighted content of our libraries provide us with a wealth of material and templates that help us generate high-quality content。
With regard to the pc tools, we have provided a series of small tools and ocr tools that can be combined with large models, using them as a gateway to understanding, while integrating them through ai processing scripts or processes, making them integrated and user-friendly。
In terms of marketing, our client base is relatively static, dominated by large businesses such as the party and the zong state enterprise. For msmes or individual markets, we will rely on traditional internet agents for outreach。
In 2024, we plan to launch the 360-internet-based ai congresser service, which will package the services of text, pictures, videos, storage spaces, etc., to users, as well as member services to enhance user experience。
Overall, 360 strategies in enterprise knowledge management and marketing are constantly seeking innovation and differential advantages to meet user needs through the provision of quality services and experience。
Liu jing: what do you think of the future of knowledge management as a market trend? What are the plans for the smart document cloud product
Yoon yoon: with regard to future trends in the knowledge management market, i believe that knowledge management will become more intelligent and efficient as large model technologies evolve and apply. Enterprises will rely more on intelligent knowledge management systems to enhance operational efficiency and innovation. At the same time, the application of knowledge management in the vertical sector will emerge, with more professional knowledge management solutions tailored to different industry characteristics and development needs。
As important participants in the knowledge management market, we have a clear planning and development orientation. First, we will continuously optimize large model technologies, upgrade the intelligence of products and provide more accurate and efficient knowledge services to users. Second, we will actively explore deeper integration with industries and introduce more knowledge management solutions tailored to industry needs, including but not limited to finance, health, education, manufacturing, etc., to meet diversified market needs。
Finally, we will strengthen ecological cooperation and look forward to closer links with a wider range of partners in different areas, building an open and mutually beneficial ecosystem, sharing resources and complementing strengths to cover a wider range of applications, collectively contributing to the flourishing of the knowledge management market, and promoting enterprise knowledge innovation and social value creation。
In summary, the future development of the knowledge management market will be a process of deeper integration of large model technologies with industry applications. Smart document cloud products will continue to support innovative, professional and cooperative ideas, provide better quality knowledge management services to users and enterprises, contribute to digital transformation and intellectual upgrading。
The smart document cloud of 36 billion clouds is a knowledge management product with leading technologies and advantages, with broad market prospects. It is believed that through further iterative and development, smart document clouds can bring value to more businesses and contribute to the digital transformation of enterprises。
