Text|Xiaoxiang Lee
Source|Intelligent Relativity (ID: aixdlun)
The black swan of the epidemic and the rapid pace of new infrastructure has become a consensus. In the first half of 2020, unlike the previous "self-purge", AI has truly ushered in its own bright moment.
In the past six months, AI companies represented by Baidu have quickly and intensively implemented a large number of in-depth applications that reflect "how much AI technology has accumulated". Regardless of the construction of the AI system or the penetration of AI scenarios, the first half of 2020 is a milestone. Baidu recently even specially released "Baidu AI's first half of 2020", from the three aspects of AI war "epidemic", AI new infrastructure, and industrial intelligence. Summarized this not-so-trivial AI in the first half of 2020.
Some people say that this is Baidu's "prosperity". It has been dormant AI for many years and finally waited for its own era. From the perspective of the entire industry, the industry reality of AI "rich accumulation" has also been revealed in the first half of 2020.
"Accumulate a lot" to give AI companies a fundamental advantage
"Save a lot and release slowly." We often praise those who have experienced long-term accumulation and finally have several high-quality results.
For AI entering 2020, accumulating and accumulating is no longer enough to describe the characteristics of its development. Accumulation and development, rapid and intensive implementation of a large number of high-value applications after long-term accumulation are the current status of AI development.
1. In the era of centralized landing, accumulating and developing become the best option
After summarizing the development of the first half of 2020, Baidu CTO Wang Haifeng said: “As an artificial intelligence platform company, Baidu will use the artificial intelligence technology achievements and practical experience accumulated over many years to help speed up the new infrastructure, promote the faster implementation of artificial intelligence, and enable all walks of life. All can get unprecedented new kinetic energy from it."
Although this passage has the intention to replace Baidu’s AI platform, it also implies another logic behind it, that is, a platform that has accumulated a deep accumulation of AI technology and meets a market with comprehensive and in-depth needs. In the past, it was step by step, one by one. The approach of landing is no longer suitable for the times; in an era where technology is changing rapidly and opportunities are emerging, AI must focus and release value quickly to seize opportunities of the times and meet the needs of reality.
And this is the "rich accumulation" of the AI industry. From the large part of "Baidu AI's First Half of 2020", it can be seen that the industry representative player Baidu AI is using this "rich accumulation" method to achieve AI. Landing.
2. Only thick accumulation can make thick hair
Regardless of the overall performance of Baidu AI in the first half of 2020, in the last month alone, Baidu AI has frequently and intensively signed several "big orders", including cooperation with Shanghai Pudong New Area and comprehensive cooperation with Shanghai Pudong Development Bank. Strategic cooperation, all-round cooperation with the aircraft carrier-level central enterprise China National Building Material Group, and participation in the State Grid "digital new infrastructure" contract.
This is a typical "thick hair", but it stems from Baidu’s long-term technology accumulation that can truly create value and can be recognized by large customer companies. On the other hand, it also stems from Baidu’s cooperation with these heavyweight companies. There have been "pre-order actions". For example, the cooperation with Shanghai Pudong Development Bank has been for three years, and there have been a large number of results in many sub-fields. The first "financial digital person" has once become a hot spot in the industry.
Obviously, for AI, only "thick accumulation" can achieve "thick hair". In other words, Baidu AI's "hard accumulation" is a result, not a choice. In the industry, many companies are still stepping up their adjustments to AI under the dividends of new infrastructure, and Baidu's more than ten years of accumulation is the fundamental reason for the rapid and intensive deployment of AI in the first half of the year.
This is a threshold that is difficult to cross in a short time.
3. Only thick hair can maximize the value of thick products
In addition to the wide range of impact of the epidemic black swan, the biggest feature is also to test the society's response ability in a very short time, including the response ability of technology.
AI's performance this time is not bad. Take Baidu AI as an example, in virus research (accelerating virus analysis and drug, vaccine research and development, etc.), spatio-temporal big data (analyzing population flow), AI temperature measurement, intelligent outbound call (serving cluster community management), AI consultation ( In response to a large number of new crown-related inquiries) and epidemic prevention and control robots (transporting supplies and meals), Baidu AI quickly launched and implemented a series of smart products and services.
The power of science and technology makes epidemic prevention and control more accurate, efficient, and warmer, and this in itself is a kind of "thick hair" of AI. This also shows that only thick hair can reflect the value of "rich accumulation" to the greatest extent. An epidemic has allowed the accumulation of AI technology to show great social value over the years.
Looking beyond the epidemic, more industries with urgent needs will also have similar rapid and in-depth landing needs, but they are not as urgent as the epidemic.
In the six months that AI has "accumulated and developed", the four major tactics of the giants emerged
In terms of specific practices, in the first half of 2020, the rapid and intensive landing of AI is not a brainstorm. From the perspective of Baidu AI, it shows four levels of tactical characteristics, which are also the embodiment of the overall pace of Baidu AI's landing.
1. Take advantage of the "systematic framework"-march on the whole
In the first half of the year, Baidu completed the important "framework" while implementing AI applications, that is, what structure is the AI business developed in. This systematic "overlook" made the rapid and intensive deployment of AI more rapid. There is "organization".
This is first manifested in that Baidu has completed a more complete and systematic depiction of the "Baidu AI Layout", showing the layout of Baidu AI to the outside world, and has also become an action map of Baidu AI's future play:
Then, in the context of the new infrastructure, Baidu sorted out the "territory" of how Baidu AI and the new infrastructure will be integrated. On the one hand, it shows the main entry scenarios for Baidu AI to promote the development of the new infrastructure. On the other hand, it also expresses the relationship with other requirements. Different overall promotion methods for companies that eat new infrastructure dividends:
As a result, the accumulation of AI has become a process of "whole march" rather than "stretch soldiers". Whether it is the AI war epidemic from basic technology to direct application, or the coverage of the entire scene of the new infrastructure, multiple attacks have shown a certain "order".
Taking the new infrastructure as an example, Baidu AI has firmly occupied the C position in AI infrastructure construction. The official said it will become "the largest provider of AI services in China's new infrastructure", in Baidu brains, flying paddles, smart clouds, chips, data centers, etc. On top of "AI technology infrastructure", Baidu AI has accelerated the upgrading of industrial intelligence in the fields of smart transportation, smart cities, smart finance, smart energy, smart medical, industrial Internet, and smart manufacturing.
Judging from the large amount of data listed on the "territory", Baidu's new AI infrastructure has not only achieved completeness, but also ensured that the application of various scenarios is sufficiently deep on the basis of technological accumulation.
2. Take the most urgent needs as the starting point-short-handedly
Under the "whole march", Baidu AI also showed the characteristics of "short-handedness" in colliding with the pain points of the industry, and most of them selected the most urgent market needs to cut in.
Needless to say, anti-epidemic. For example, Baidu’s AI temperature measurement system quickly built in railway stations and other places with high traffic flow uses AI image recognition technology and infrared thermal imaging technology to solve the problem of fewer facial recognition features caused by wearing masks and hats , A total of 27 million people were tested within two months of landing, which played a direct value in epidemic prevention and control.
In more industrial fields, Baidu AI is solving the most urgent and realistic problems on important tracks. Not long ago, Baidu Smart Cloud completed the strategic upgrade of "based on cloud computing, with artificial intelligence as the starting point, focusing on important tracks", and it focused on several important tracks in the early stage.
For example, in the financial track, Baidu Smart Cloud relies on AI middle station and knowledge middle station to build "future bank" solutions, helping financial companies to provide customers with a completely consistent financial service experience online and offline, and solve the digital advancement process Each channel in the experience different frequency problems. Public data shows that Baidu "Smart Finance" has served nearly 200 financial customers and built an ecosystem of more than 30 partners;
On the smart city track, Baidu has built a solution system with the "1+2+1" framework as the core. These solutions have been implemented in many cities. Coordinated decision-making basis supports management and services; in May, it signed a strategic cooperation agreement with Chengdu High-tech Disaster Reduction Institute to empower emergency management, build smart cities, etc.;
In the energy track, Baidu’s Smart Energy has reached cooperation with the two giants of the State Grid and China Southern Power Grid, completing more than 20 business scenarios such as drone inspection, “swiping face for electricity”, and integrated energy, and covering two Chinese Intelligent monitoring of UHV intelligent lines, more than 150 intelligent substations, and more than 40,000 transmission lines;
In addition, there are more than 300 hospitals and more than 1,500 primary medical institutions in the field of smart medical care, assisting tens of thousands of doctors and serving more than 25 million people. The smart customer service product has undergone a new upgrade, successfully landing more than 80 customers, and developing 40 With multiple partners, intelligent manufacturing successfully landed in 14 sub-industries and more than 50 customers, etc.
The focus of these important tracks all show Baidu AI's intention to solve key problems tactically by "accumulating and developing."
3. Improve the landing effect with mutual coordination-formation echo
Baidu AI's "thick hair", under the overall march, its tactics also reflect the characteristics of "formation echo", that is, to split different dimensions to interact with each other to improve the speed and quality of the overall landing.
The most typical one is the industrial application of Baidu's new AI infrastructure, which not only has the implementation of purely external ecological cooperation projects, but also uses Baidu's existing technology to promote the in-depth application of AI, showing the characteristics of internal and external collaboration.
For example, the effectiveness of Baidu AI's application in the Baidu mobile ecosystem continues to emerge. Typically, the satisfaction rate for the first search result of Baidu App has increased to 58% in the past year, and more than half of the people have met the demand for the first search result. .
In addition, with the support of Baidu's AI capabilities, Baidu Maps has upgraded industry solutions such as smart private network maps, smart logistics, and smart parking, and intensively deepened cooperation with industry leaders such as Shuanghui Logistics, China UnionPay, and Tesla.
The same is true for Baidu Apollo. In terms of autonomous driving, it has currently tested 150 licenses, landed in 24 cities, and tested over 6 million kilometers. In March 2020, NR, a world-renowned research organization, named the world’s top four “leaders in autonomous driving” ", Apollo ranks among them; in terms of intelligent transportation, Baidu released the first domestic and international full-stack intelligent transportation solution "ACE Traffic Engine" that integrates vehicles, roads and lines. It has been installed in more than 10 cities in Changsha, Beijing, Chongqing, Nanjing, etc. City landing practice.
In addition to internal and external dimensions, Baidu AI's "thick hair" can be found, and it also presents the characteristics of soft and hard collaboration. For example, Baidu Xiaodu has built a complete ecosystem of software and hardware integration, with both market leadership in hardware products and industry-wide empowerment of voice solutions;
Another example is that with a large number of AI "soft" solutions, Baidu's "hard technology" is also accelerating. Baidu Kunlun chips have been widely used in the industrial field. The second phase of the Yangquan Supercomputing Center has been completed. In the future, Baidu Smart Cloud 5 million A server plan is also on the agenda.
4. Accelerate application production with multiplier mode-quick replenishment
If the landing of AI is regarded as a war against the pain points of the industry, in addition to the overall march, short-term engagement, and formation response, the rapid recruitment of soldiers has also become an important tactical content that affects the course of the war.
Baidu AI's thick launch in the first half of 2020 also shows that it has completed a certain multiplier model application acceleration at the technical and business levels.
At the technical level, Baidu Feidao faces the sudden increase in demand for AI applications in the first half of the year, and its value is further highlighted. Currently, it has 1.94 million developers, serves 84,000 companies, and has created 233,000 models. Platform-based AI industrial production is better than Baidu. The creation efficiency of one application by oneself is too fast;
At the business level, as mentioned above, Baidu Smart Cloud has completed the strategic upgrade of "based on cloud computing, with artificial intelligence as the starting point, focusing on important tracks". Simply put, Baidu AI has the business of unifying resources and sharing capabilities The structure not only allows enterprises to quickly build their own applications, but also enables Baidu AI to deploy intelligent applications or solutions quickly and flexibly.
After the first half of the year, can AI's "thick hair" momentum continue?
A realistic question is whether the “rich accumulation” demonstrated by the landing of AI will be a transient prosperity. After the “new infrastructure” has been on fire for a while, will the momentum of “thick development” stop and become the past That high-quality but not too frequent "thin hair"?
At least from the information currently available, the accumulation of AI will not be a transient phenomenon. This is because it was originally the result of a series of basic conditions rather than a choice. Giants such as Baidu AI have completed the preliminary technical accumulation and structure combing. Is the norm.
Moreover, AI giants are also continuing to deepen their accumulation and provide more and lower-level guarantees for AI development.
In terms of technology, Baidu's basic technical reserves are still advancing. On May 28, 2020, Baidu's global AI patent applications have exceeded 10,000, of which more than 7,000 Chinese patents, ranking first in China. In the major core competitions and top conferences in the first half of the year, Baidu also demonstrated this technical ability. For example, 28 papers were included in AAAI 2020, 22 papers were selected as the "Oscar" CVPR 2020 in the field of computer vision, and 8 challenges were won. Race the world champion.
More content will not be listed one by one. Under the technical background of the strong Hengqiang, the technological accumulation of AI giants has also shown a significant acceleration.
In terms of the most important guarantee for the development of AI and the construction of the talent team, Baidu has taken more future-oriented actions.
In mid-June, Baidu announced that it is expected to train 5 million AI talents in the next five years. Specifically, Baidu will provide teaching resources, training platforms, ability assessment certification, corporate talent training, industrial ecological cultivation, competition activities, etc., these actions and basic technologies Like the accumulation of AI, it has become the fundamental guarantee for AI's continued “thick growth”.
Conclusion
Looking back from the future, the first half of 2020 will be a magnificent and extraordinary half year regardless of the macro environment or industrial development.
As far as AI is concerned, represented by the large-scale implementation of Baidu AI, the era of “rich accumulation and development” has arrived. On the one hand, the process of AI development has been double-accelerated by the background of new infrastructure construction and the efforts of giant companies themselves, on the other hand , The only feature of deep accumulation and development also makes Baidu one step closer to "the largest provider of AI services in China's new infrastructure", and the only C position is set.
It can be expected that the first half of 2020 will only be a preview of "accumulation and development", and there will be even more exciting performances in the era of vigorous development of new AI infrastructure.
Dig the smart well, and add vx: zhinengxiaoyan
This content is the original [Intelligent Theory of Relativity],
It only represents personal views. Without authorization, no one may use it in any way, including reprinting, extracting, copying, or creating mirror images.
Some pictures are from the Internet, and the copyright ownership has not been verified. They are not for commercial use. If there is any infringement, please contact us.
Intelligent Relativity (WeChat ID: aixdlun):
• AI industry new media;
• Today’s Toutiao Qingyun Project winners TOP10;
• The paper's monthly top5 in the news technology list;
• Articles "occupies" the TOP10 list of popular articles in titanium media for a long time;
• Author of "One Hundred Thousand Whys of Artificial Intelligence"
• [Key areas of focus] Smart home appliances (including white appliances, black appliances, smart phones, drones and other AIoT devices), smart driving, AI+ medical care, robotics, Internet of Things, AI+ finance, AI+ education, AR/VR, cloud computing , Developers and the chips and algorithms behind them.