"Quality inspection is a meticulous job. You can stare at the glare for 8 hours, your eyes are spent, and the quality and efficiency of your work drop significantly in the last period of time. In the shuangtu new material production workshop under Hengyi Chemical Fiber, one of the world's largest fiber manufacturers, Wang Lina, the head of the section, said, "No change, no." She has been doing quality inspection of chemical fiber products for 15 years, and her normal work in the past was to take a glare flashlight to see with the naked eye, and at least 2500 spindles were manually detected every day.
In the past, workers held a glare flashlight to check the product with the naked eye, which was easy to cause damage to vision
Nowadays, the intelligent quality inspection equipment jointly developed by Hengyi Chemical Fiber and Baidu Intelligent Cloud has been deployed on the production line, and the traditional optical imaging and artificial intelligence algorithms are combined to achieve in-depth analysis of the complex surface of the product, which can shorten the detection time of a single silk ingot by 70%, and the work of quality inspection women has become easier. Wang Lina also transformed into an AI data annotator, helping Baidu's intelligent cloud quality inspection engineers mark various defects on product pictures, transforming the quality inspection experience of quality inspectors into data assets of enterprises, and allowing AI to learn to distinguish product defects such as tripwire, hair, and oil pollution.
Nowadays, with intelligent quality inspection equipment, the original general inspection has become a re-examination after machine screening
In 2021, the phenomenon of AI empowering traditional industries is happening in manufacturing, energy, finance, cities and other thousands of industries, helping countless enterprises to reduce costs and increase efficiency and enhance competitiveness.
For cloud vendors, when industrial intelligence sweeps the entire market, the competition between head cloud vendors has changed from the competition of cloud "computing" to the competition of AI "services". Almost all vendors have added smart elements to their services. "Understanding the cloud, understanding AI, and knowing how to do it" has become a new standard for enterprises to choose cloud computing vendors.
Wang Haifeng, CTO of Baidu, described the market's new demand for cloud computing platforms as "a cloud suitable for running AI" and "AI that understands scenarios". He said at the "Intelligent Economy Summit Forum" in July 2021 that the "cloud" provides a secure, stable and flexible digital base for digital transformation, and the "intelligent engine" provides leading innovative technologies and platforms for intelligent upgrading. That is, the cloud and AI are not simply superimposed, and the integration of cloud and intelligence is deeply integrated, opening up the path of "growth on the cloud" for enterprises.
"We recommend that customers do digital transformation and intelligent upgrading together, one step in place, so that the efficiency is higher and the effect is better." Wang Haifeng said.
Baidu CTO Wang Haifeng
Yunzhi integrates industry and finance and promotes rural revitalization
For science and technology enterprises with new technologies such as artificial intelligence, cloud computing, and big data, driven by national policy support and industrial structure transformation, it is the industry consensus in 2021 to widely participate in the real economy and lay out traditional industries.
The "14th Five-Year Plan" clearly states that it is necessary to give full play to the advantages of massive data and rich application scenarios, promote the deep integration of digital technology and the real economy, empower the transformation and upgrading of traditional industries, spawn new industries, new formats and new models, and strengthen the new engine of economic development.
At the same time, at a press conference in April 2021, the Ministry of Industry and Information Technology proposed that Internet companies should be guided to increase the intensity of scientific and technological innovation, encourage head enterprises to participate in major national strategies and major projects, increase investment in 5G, cloud computing, artificial intelligence, industrial Internet, etc., promote the development of digital industrialization and industrial digitalization, make up for shortcomings, forge long boards, and help enhance the core competitiveness of China's science and technology.
Macro trends superimpose the transformation demands of traditional industries, and technology companies have a new role positioning and mission of the times - to release the potential of artificial intelligence, cloud computing and other technologies in various forms such as platforms, soft and hardware products, and industry solutions, and leverage the productivity of larger traditional industries.
But intelligent upgrades are easier said than done. According to a boston consulting group survey, because the general AI model is difficult to meet the needs of all scenarios, 86% of the intelligent practices in the market require customized AI services. But for most industry companies, not having enough experts proficient in AI algorithms is the first problem on the road to intelligence.
Similarly, as AI is deeply applied in more scenarios, AI has to process more and more data, models are getting bigger and bigger, accuracy requirements are getting higher and higher, and with the explosive growth of computing power demand, the cost is getting higher and higher. This also makes many enterprises prohibitive from intelligent upgrading.
(Baidu Intelligent Cloud releases "Cloud-Intelligence Integration" Architecture 2.0)
This means that when cloud computing vendors provide services to industry users, they need to have three major capabilities: in terms of infrastructure, they must dock and manage AI computing resources with good performance, high utilization and high cost performance; in terms of research and development efficiency, they must provide the optimization of the whole process and the optimization of development efficiency; and at the same time, they can provide overall solutions for the actual business scenarios and customized needs of enterprises.
In response to the new requirements in the process of AI application sinking, in July 2021, Baidu Intelligent Cloud released the "Cloud-Intelligence Integration" architecture 2.0, which was comprehensively upgraded from four aspects: strategy, architecture, product and ecology. With "cloud suitable for running AI" and "AI that understands scenarios" to jointly construct the infrastructure of the intelligent era, we will solve the two core demands in the intelligent upgrading of enterprises: "cloud suitable for running AI" eliminates the burden of computing power of enterprises; "AI that understands scenarios" reduces the development threshold of enterprises and builds its own AI capabilities in the cloud.
As a result, Baidu's leading AI technology has penetrated deep into the field, factory workshop, urban and rural areas through Baidu intelligent cloud... It takes root and sprouts in the vast land of China and blossoms everywhere.
In August 2021, at the Baidu World Congress jointly held with CCTV News, CCTV reported on the landing application of a number of cloud-wise integrated empowerment industries such as Quanzhou Water Brain, Xinjiang State Grid Power, Lijiang City Brain, and Suzhou Intelligent Manufacturing.
Taking the drinkable water available at home in Quanzhou, Fujian Province, as an example, a drop of tap water must be taken from the water source through the pumping station, transported into the water purification plant, and experienced conventional treatment and chlorination disinfection such as flocculation tanks and sedimentation tanks, as well as the deep treatment of ozone and bioactive carbon, before it can be turned into directly drinkable water, which can be transported to various communities and streets through urban pipelines. In the chlorination and disinfection link, in the past, it was mainly based on human experience to add chlorination to the water, which hid potential safety hazards. Nowadays, Quanzhou Water Affairs Group, Baidu Intelligent Cloud, and Ehrendy Company jointly built the Quanzhou Water Brain, which realizes the automatic monitoring, calculation and dosing of intelligent algorithms; in the inspection link, the staff used to need to run in the equipment room every day, always a body of water and mud. After having a water brain, as long as the central control room can complete the accurate management of the whole process of the water plant covering an area of more than 240 acres and 17 factories, the management efficiency and accuracy are greatly improved, and even the phenomenon of "the whole water plant has only two staff".
CCTV celebrity Sabine sighed, "It can be said that in Quanzhou, the tap in your home, this city understands you." Some time ago Quanzhou became a World Heritage Site, I hope you have time to see, to that side do not need to bring water, take an empty bottle to go, pass by the street which faucet opens to drink, too happy. ”
Specific to the intelligent transformation practice of enterprises, its path can be roughly divided into four stages: application exploration, vertical deepening, comprehensive upgrading, and continuous innovation. For large enterprises and small and medium-sized enterprises, due to the different technical thresholds and R&D expenditures, there are also different ways to implement the integration of cloud and intelligence into practice.
Large enterprises started using digital applications earlier, and most of them are currently in the first and second stages of application. For example, the access control gate of face recognition has been popularized in offices, construction sites, and factory parks.
However, solutions in single-point AI application scenarios are often built separately, and the models used may come from third parties or independent research and development, and it is difficult to form a sharing of core resources in data, computing power and algorithms. This will affect the precipitation of enterprise core experience and data assets; when sharing and unified upgrading of AI capabilities across business scenarios, some purchased third-party models are also facing the problem of independent optimization; finally, models from different vendors and different sources also have difficulties in unified operation and maintenance and resource allocation, so that the overall technical capabilities are difficult to reuse, affecting the process of intelligent transformation and upgrading of the entire enterprise.
When large enterprises evolve to the third stage, it is an inevitable problem for large enterprises to build an intelligent middle platform for unified AI capability production, application, operation and maintenance.
As a large central enterprise in China, China Southern Power Grid is facing the above problems in the process of digital transformation. On the one hand, massive and important data, especially non-structural data such as graph file data, machine patrol data, 95598 customer service audio, and production surveillance video generated by various business systems, is difficult to form a unified storage and management; on the other hand, how to tap the value of data to assist decision-making has also become an urgent problem to be solved by China Southern Power Grid.
"China Southern Network Smart Search" came into being, and has become a powerful grasp for revitalizing enterprise data assets by introducing the internal business data, office information, asset information, collaborative materials, industry information and various content operation data of China Southern Power Grid into the knowledge middle office. A person in charge of the data platform business department of the China Southern Power Grid Digital Grid Research Institute said, "Since june 2020 in cooperation with Baidu 'China Southern Power Grid Smart Search', it has integrated various types of multi-source and heterogeneous data resources within the company, providing low-threshold, high-credible knowledge services for enterprise employees, and solving the 'three noes' problem of data and information often encountered within enterprises in the past. ”
At present, the "dual middle platform" combination of AI middle office and knowledge middle platform, as an important drop of Baidu's intelligent cloud "cloud-intelligence integration", has been deployed in leading enterprises in energy, manufacturing, finance, cities and other industries, including the State Grid, State Power Investment, and China National Building Materials. Through docking with resources and systems in the enterprise, the "dual middle platform" can quickly build intelligent applications for more than 500 scenarios, helping large enterprises to intensively manage AI capabilities and knowledge, and accelerate the collaborative innovation of intelligent upgrading.
Unlike large enterprises, small and medium-sized enterprises have insufficient technology accumulation and generally low investment in research and development. However, small and medium-sized enterprises are the largest and most innovative and dynamic enterprise group in Chinese society, and they are the main force of "Made in China". In order to promote small and medium-sized enterprises to take the road of "specialized and specialized" development, the relevant departments have also continued to issue a number of policies to support the digital and intelligent innovation and application of small and medium-sized enterprises.
According to the "Survey Report on the Cloud Use of Small and Medium-sized Enterprises in China" (hereinafter referred to as the "Report"), jointly sponsored by the Southern Weekend Science and Technology Innovation Research Center and the China Association of Small and Medium-sized Enterprises, through the investigation of more than 600 enterprises, it is found that small and medium-sized enterprises have a strong motivation to "use data to empower wisdom on the cloud", but unfortunately they lack corresponding technical personnel and technical ability reserves. At the same time, under the pressure of market survival and development, small and medium-sized enterprises pay more attention to the input-output ratio of cloud services.
Therefore, in the process of small and medium-sized enterprises going to the cloud, cloud service providers and third-party service integrators play a vital role. Combined with industry application scenarios, the complex digital and intelligent technologies are handed over to enterprises in the form of specific products/tools, helping small and medium-sized enterprises to open the "cloud transition" from artificial mode to artificial intelligence mode, better "cloud-intelligence integration innovation", and participating in market competition, which has been verified to be one of the best practice models for cloud manufacturers to empower small and medium-sized enterprises.
Southern Weekend Science and Technology Innovation Research Center and the Association of Small and Medium-sized Enterprises jointly launched the "Research Report on the Cloud Use of Small and Medium-sized Enterprises in China"
"In the past, the company's safety management was routine monitoring, and when something happened, it was only checked. Today, once a violation is discovered, the system corrects it in a timely manner. A person in charge of Guizhou Jinjia New Packaging Materials Co., Ltd., located in Guiyang Economic Development Zone, said, "When there is no access to the 'network' of the industrial Internet, the company also wants to make a similar platform, but considering the huge investment, the lack of information talents and technology has not been realized." After integrating into the industrial Internet platform, the packaging manufacturing workshop is more secure and intelligent. ”
Baidu Intelligent Cloud took the lead in insight into the intelligent interconnection needs of industrial enterprises, based on the open platform, to create the country's first regional-level AI industrial Internet platform, the platform has been put into use in Guiyang Economic Development Zone, many enterprises have enjoyed the platform empowerment "Guizhou intelligent manufacturing" brought benefits.
"Equipment maintenance is indispensable for the development of industrial enterprises, but many companies do not have professional teams, and many industrial enterprises have to hire engineers with high salaries, or contact equipment manufacturers and find professional maintenance teams in other provinces, which is very inconvenient." An engineer from Guiyang Xinghang Electromechanical Technology Co., Ltd. said.
"After we join the Industrial Internet, enterprises on the platform can contact us immediately when they need maintenance, and upload equipment alarm photos through the platform. We carry out precise maintenance based on early warning situations. He said, "Since joining the Guiyang Industrial Internet Platform, Xinghang Electromechanical has added a number of new cooperative customers." In the future, the company will provide inexpensive and efficient maintenance services for more small and micro enterprises on the platform. ”
Xinghang mechanical and electrical engineers carry out equipment maintenance for enterprises
At the same time, for the current widely concerned green finance topic, Baidu Intelligent Cloud also dressed up as a "red bride" and pulled a "red rope" for industry and finance.
Lei Ming, secretary of the party committee and president of Jianxin Financial Technology Co., Ltd., believes that "what we are facing today is to use new technologies to rebuild the financial "water conservancy project" to ensure that financial resources are fairly distributed. ”
Many financial institutions have branches in villages and towns. However, the operating capacity and service capabilities of these outlets have a big gap compared with the headquarters and large cities, and they cannot better meet the needs of rural finance. The use of artificial intelligence technology to enhance the radiation capacity and service quality of village and town-level outlets through certain sorting and extraction of knowledge and service capabilities has become one of the main intelligent practice directions of financial institutions.
At present, Baidu Intelligent Cloud is cooperating with many financial institutions to sink high-quality and inclusive financial services to counties, villages and towns through AI technology. "Baidu through the construction of a characteristic industrial Internet platform, with artificial intelligence + industrial Internet as the industry digital position, we will serve more than 100,000 enterprises in the digital archives and financial institutions in various places to cooperate," Baidu vice president Li Shuo said, "on this basis of financial product innovation can support a series of financial needs of enterprises in the capital, transactions, supply chain, this two-way innovation for technology and financial empowerment of the real industry to achieve the goal of double carbon has formed a better boost." ”
In addition, Baidu Intelligent Cloud has also expanded the experience of AI-empowering the management and construction of cities such as Haidian in Beijing and Lijiang in Yunnan to new urbanization and rural revitalization.
In Shouguang, Shandong, the "hometown of Chinese vegetables", the local government and Baidu Intelligent Cloud jointly created the Shouguang Brain, which accurately transformed vegetable planting, irrigation, fertilization, harvesting and other links through AI technology, and only 2 staff members plus an APP can manage dozens of vegetable greenhouses. Use artificial intelligence technology to help vegetable cultivation change from "experience planting" to "intelligent planting".
AI-native cloud helps enterprises chase the digital golden five years
The many commercial value and public service value brought by AI to enterprises and governments are obvious to all, so the number of AI applications has also shown an explosive growth trend. Deploying AI applications on cloud platforms has become a natural choice for cloud computing vendors.
The growth of artificial intelligence application demand has put forward new requirements for computing power and the underlying architecture of cloud computing. This is not only reflected in the network with higher computing density and greater transmission capacity, but also in the development and deployment of AI applications that are better carried. The rapid development of AI is inseparable from the support of computing power, and in the post-Moore era, the general computing architecture has been difficult to meet the demands of the new era.
Driven by demand, the computing field ushered in three new trends. First of all, the computing power is heterogeneous, and the diverse computing scenarios require the support of heterogeneous computing power, including CPUs, GPUs and customized chips in various fields; secondly, the computing deployment location is distributed, and a large number of computing resources are distributed in large data centers and various edge ends, such as smart parks, smart factories, etc.; finally, the application is becoming more and more universal, and it has a very wide range of applications in finance, smart cities, video, agriculture and other fields.
"In the era of intelligence, AI native cloud is the core driving force to promote the intelligent upgrading of enterprises", said Hou Zhenyu, vice president of Baidu Group, "Baidu Intelligent Cloud can provide highly elastic and high-performance heterogeneous computing power for AI scenarios at the computing power level, and at the application development level, it can provide a series of low-threshold development platforms for AI application scenarios, thus helping enterprises to make the AI application architecture more concise and agile." ”
From cloud to cloud native, from cloud native to AI native cloud, cloud computing is no longer limited to IT computing itself, but a large number of intelligent transformation needs, including AI, are integrated into the capabilities needed for intelligent transformation, so that cloud computing is more and more in line with the real needs of enterprise intelligent upgrading, making the value of cloud computing further burst forth.
Wu Lianfeng, vice president and chief analyst of IDC China, believes that from full-stack technology to hot applications, the trend of AI nativeism is becoming more and more prominent, mainly reflected in six aspects: cloud native computing power, large-scale data, factory of AI algorithms, API of functional standards, agile development mode, and invisible application of scene applications.
ICBC is one of the largest commercial banks in China and one of the first financial companies to see the potential of cloud natives. Facing a wide range of customers and flexibly handling large-scale transactions is a challenge that ICBC faces every day. In addition, in order to adapt to the competition among commercial banks and the penetration of Internet applications, ICBC must also achieve rapid innovation.
Serverless can help ICBC achieve rapid on-line services. By working closely with Baidu's intelligent cloud AI cloud-native technology, ICBC used to take at least two to three months to develop a new business, and it took several weeks to deploy and go live. Now, the development of a business is online, and ICBC only needs a few weeks, or even a few days, to accelerate the speed of business innovation.
In order to meet the business development needs of the global layout, Geely, the leader of domestic car companies, also sees the value of cloud services. Through the "1+6+N" Geely Cloud platform jointly built with Baidu Intelligent Cloud, Geely meets the needs of external rapid response service users and internal data security protection with a Baidu hybrid cloud architecture model that includes public cloud + private cloud. It is worth mentioning that the Geely cloud platform uses the architecture model of Baidu public cloud from the bottom layer, so it can follow the development of Baidu public cloud and upgrade in real time, which greatly reduces the difficulty of upgrading services in the later stage.
Geely Holdings said that relying on this platform, Geely has planned a clear four-stage digital strategy of "consolidating the foundation - business cloudification - business intelligence - empowering business innovation", that is, first moving the original business to the cloud and unifying the resource base; then using the cloud native architecture to create a new application system on the cloud; then integrating data and AI capabilities into the business; and finally forming a basic platform and capabilities to support the business innovation at the upper level. The two major trends of cloud-native leading change and AI-native driving innovation are obvious.
"The next five years will be the golden five years of China's digitalization," Wu Lianfeng said, "if more investment can be made in digitalization, then companies will have a better competitive starting point in the next three to five years." ”
A better 2022
In the "10 Trends in the ICT Market in 2022" released by IDC, the development of artificial intelligence in new scenarios was listed as the first trend. IDC predicts that global business budget spending on AI technology will grow from $287 million in 2021 to $365 million in 2022, a growth rate of up to 27%.
On the basis of 2021, when we look forward to 2022, we will find that AI technology will continue to accelerate the sinking and inclusiveness. For enterprises that have already gone to the cloud, exploring new intelligent application scenarios, accelerating the use of wisdom on the cloud, and improving production and decision-making efficiency through AI is an important way to achieve transformation and upgrading; for scenario-oriented enterprises with relatively weak digital level and considering one-step in place, when choosing cloud vendors, they will pay more attention to AI strength, and support business innovation with stronger heterogeneous computing power and more mature AI technology.
On January 12, 2022, the State Council issued the "14th Five-Year Plan for the Development of the Digital Economy", which proposes that a new intelligent infrastructure integrating intelligent computing power, common algorithms and development platforms should be built, and systematic artificial intelligence services should be provided for key emerging areas such as government services, smart cities, intelligent manufacturing, automatic driving, and language intelligence.
For cloud computing vendors, cloud-intelligence integration will become an important indicator to test their technical and practical capabilities. While providing basic capabilities such as computing, storage and networking, how to better exert the potential of computing power and support intelligent application scenarios has become the core element of the second half of the competition in the field of cloud computing.