Smart Things reported on July 8 that this morning, the 2021 World Artificial Intelligence Conference (WAIC) opened in Shanghai! Li Yanhong, Ma Huateng, Hu Houkun, Dong Mingzhu, Zhou Hongyi, Mei Hong and other industry, academic and political leaders from all over the world, as well as two investment circles, Shen Nanpeng and Sun Zhengyi, shared content and talked freely about their intellectualization from all angles. And thinking about digital development.
As an annual landmark event in the artificial intelligence industry, WAIC has experienced the “ cloud exhibition ” held last year affected by the epidemic , and finally returned to offline this year. The number of guests and exhibition area of the conference doubled, and the number of exhibitors and exhibitors also exceeded 300 Home.
At the opening ceremony, Baidu founder, chairman and CEO Robin Li launched a series of discussions on the social value of AI; Huawei Rotating Chairman Hu Houkun, who participated in the WAIC speech for the first time , put forward three suggestions for accelerating the development of the AI industry. At the same time, Mei Hong, academician of the Chinese Academy of Sciences, and Zhou Hongyi, founder and chairman of 360 Group, shared the challenges of digital transformation in the current society.
However, this year’s opening ceremony did not show the famous scene of Musk, Jack Ma, and Ma Huateng’s “three horses” gathering like in 2019 and 2020. Except for the absence of the former “two horses”, Ma Huateng is still connected via remote voice. Way cloud participation.
In addition, this year WAIC site B station invited virtual idol Ling kite, millet little love students, Microsoft wheatgrass as AI virtual host, while radio and television host and Shanghai hosted the same stage, performing their new skill.
The highest award of the World Artificial Intelligence Conference-“Super AI Leader” (“SAIL Award”) was also announced at the opening ceremony. The 5 SAIL award winners are: Robotic Scientist from the University of Liverpool and Shanghai Chinese Academy of Sciences. Institute of Microsystems (no craniotomy and minimally invasive implantable high-throughput flexible brain-computer interface), Kunyun Technology (high-performance data flow AI chip CAISA), Zhejiang University (digital creative intelligent design engine), as well as Zhu Mingjie, Wu Fei, Lu Cewu’s thesis “China Towards a New Generation of Artificial Intelligence”.
01 Yanhong Li: See you in 2 to 3 years after achieving carbon neutrality in 2023
Before the official lecture, China’s first Mars rover digitizer Zhu Rong, released by China’s Mars Exploration Project and Baidu, was the first to debut. Not only did he share his check-in photos taken on Mars, but also invited Baidu’s founder, chairman and CEO Robin Li Come to the stage to talk and take a group photo.
“Artificial intelligence as a national strategy has been deeply rooted in the hearts of the people.” Li Yanhong said that in the past few years, everyone’s discussion on AI has mainly focused on two directions. One is the help that AI provides for future economic development, and the other is AI’s contribution to the future. Risks and unpredictable changes, or ethical issues related to AI.
However, there are still few discussions on the value of AI to other aspects of human society in the industry. Li Yanhong believes that the social value of AI is very worthy of attention and is closely related to the people’s yearning for a better life.
For example, in intelligent transportation, people pay more attention to the improvement of traffic efficiency and the growth of urban GDP, but in fact, about 1.35 million people are killed in traffic accidents every year, and 94% of traffic accidents are caused by man-made causes. At the same time, the world’s annual carbon emissions are as high as 51 billion tons, of which 16% comes from the transportation industry.
“Therefore, the development of smart transportation systems through smart cars and smart roads can not only improve the efficiency of safe traffic, but also reduce carbon emissions and make travel more green and environmentally friendly.” Li Yanhong said.
On the other hand, as our country enters an aging society, it also poses new challenges to the entire society. Li Yanhong said that in the field of intelligent assistance to the elderly, AI has played an obvious role in providing convenience for the elderly, economic rescue, and mental crisis.
“When the elderly fall at home, how to find out, notify family members, and provide assistance in time, machine vision can do a good job.” He said.
At the same time, AI is continuing to expand the living space and freedom of human beings. For example, the Mars rover digital person Zhu Rong can not only achieve new breakthroughs in machine translation and autonomous driving technology, but also make cars no longer look like cars, but more like robots.
Li Yanhong mentioned that many people now think that a smart car is like a huge smart phone, but Baidu believes that the smart car in the future is more like a robot. “Conversely, the mainstream robots in the future will look like smart cars.”
On the whole, Robin Li believes that AI will undoubtedly affect the process of human development in the next 40 years. In the transportation, finance, industry, energy, media and other industries, the application of AI technology will provide new ideas and new solutions for the digital upgrade of the industry, which will affect human society. s future.
In order to better meet this future, Robin Li revealed that Baidu has just launched a new shared unmanned vehicle that will make travel cheaper. In the future, it plans to open shared unmanned vehicle services to more than 30 cities across the country, and new unmanned vehicles are also available. In the simultaneous research and development, it is expected to meet with you in the next 2-3 years.
At the same time, Baidu will reduce carbon emissions through AI innovation, and the group is expected to achieve carbon neutrality at the operational level in 2023.
02 Ma Huateng: 145 seconds of voice speech, release of “Star Exploration Project”
This year , Ma Huateng, Chairman and CEO of Tencent ‘s board of directors, still failed to come to the scene. Like last year, he sent remote greetings to everyone in the form of voice chat. Compared with the last 60-second voice, this time I did not deliberately. Four messages were sent. The total 145 seconds should be the shortest speech in the audience… And from the interface, the enterprise version of WeChat was used.
Ma Huateng emphasized that in the past few years, Shanghai has gathered talents, opened up scenes, and built platforms to create a good innovation ecosystem for technology companies including Tencent. At this conference, Tencent and the National Astronomical Observatory will jointly release the “Star Exploration Program”, which will use the AI technology of Shanghai Youtu Lab to find pulsars and explore the universe.
Not only that, Tencent will also bring the e-sports of the glory of the king, so that everyone can appreciate the highest level of AI competition, and at the same time stimulate young people’s interest in AI research.
Ma Huateng said that in the past few years, AI has brought more and more convenience to people’s lives in the fields of medical care, urban treatment, and non-contact services, but people still know more about AI than they know. And Tencent’s pursuit of technology for good is to realize AI for good, making AI technology usable, knowable, and reliable.
Therefore, Tencent has also incorporated sustainable AI innovation into the company’s core strategy this year, hoping to use Xiangshan to guide the company’s business technology development and technological innovation, step by step to achieve technological improvement.
03Two internationally renowned professors: AI faces three challenges, two insights into the development of smart cities
The 2007 Turing Prize winner, Joseph Sifakis, a professor at the Verimag Lab of the International Embedded R&D Center, shared his views on weak AI to strong AI and then generalized AI through video.
He believes that weak AI is just a cornerstone, and in order to realize AI in a broad sense, it may not be progressive machine learning. In fact, people can develop from a single task, a single area, a single system, such as personal assistants, natural language translation, etc., to fast and rapid intelligent systems, working in unpredictable electronic and mechanical environments, such as smart cars. , Smart grid, smart factory, etc.
Based on this, Joseph Sfarkis mentioned three of these challenges.
One is neural network. The advantage of neural network is that it can quickly distinguish objects such as cats and dogs, which means that it can be from 1 to 0. But in terms of input independent variables, when the odd and even numbers of 0 and 1 change, more or less changes in the input data may cause independent variables.
The second is the interpretability of neural networks. People have a great deal of trust in neural networks. This requires a model that can learn by itself and can continuously learn in an adaptive environment. Joseph Sfarkis thinks this is the most difficult problem.
The third is the definition of intelligence. “Intelligence is not something that is known, intelligence is something you use when you don’t know what to do,” he said.
“Father of big data,” Oxford school network network College Professor of Internet Governance and Regulation Schoenberg (Viktor Mayer-Schönberger) then shared two points drive him forefront of the development of the city for insights.
The first is the megatrend. As of 2020, more than half of the world’s population will live in cities or urban areas. This is an important milestone in the history of human development. As more and more people live in cities, it has brought tremendous pressure to cities. Cities need better management and intelligent systems, which requires the development of smart cities.
But the core of a smart city does not lie in a specific technology or digital infrastructure. The core lies in making better decisions. The foundation lies in rich and diverse data. This requires far more than technological advancement, but also organization and mechanism. Innovation and progress.
At the same time, the concept of smart city development cannot only be applied to large cities, it must also be applied to small cities to have a real impact. However, small cities have limited resources and few implementation experience, so there are many challenges. Therefore, managers should base their decisions on facts rather than compromise policies.
The second insight is that people need to know what data to make the smart city concept effective, and how to properly use indirect methods to collect data. Generally speaking, complex pattern recognition and deep learning methods can find potentially useful proxy data, but this usually does not require particularly novel technologies, and relies more on innovative ideas and breakthrough thinking, so it will be more difficult to execute .
“We need to develop the concept of a smart city so that it is no longer limited to a narrow view of data infrastructure, and only in this way can it succeed.” Schönberger said.
04 Hong Mei: Digital transformation is a trend, and open source has become an indispensable strategic resource
ACADEMY Shimei macro site to share a few Understanding word transformation and open innovation.
He emphasized that digital transformation has become a trend, and the development of a digital economy is a historical opportunity for the country. Now, building a digital China and digital economy has become a national strategy. This is the inevitable choice of the times. The inevitable way is the digital transformation of various industries, especially the transformation of manufacturing.
“This round of change is the general trend, and its core driving force is the extensive connection of people, machines, and things brought about by the Internet and its extension.” Mei Hong said that he has two basic observations on this:
One is that all business formats will be deeply coordinated and integrated around the main line of informatization to complete their own transformation and transformation, and give birth to new business formats. Many traditional business formats will also die out during this transformation; second, transformation will be a long-term process, he believes This stage may last for decades.
But at the moment when my country’s digital transformation is in full swing, there are also several phenomena: not wanting, not daring, and not going.
“Don’t want to know it is not in place, dare not to be a problem of unknown path, it will not be a problem of technology.” Mei Hong explained that digital transformation is facing challenges in systems, management, talents, and concepts, of which concept is the most important factor.
Therefore, he believes that digital transformation requires ideas first, through emancipating the mind to achieve key transformations, and establishing Internet thinking and digital thinking. At the same time, education is a long-term process. We need to cultivate qualified talents that we need in the future. At the same time, we must have overall planning and promotion at the government and enterprise levels.
In addition, Mei Hong also mentioned his understanding of open innovation. What is open innovation? He believes that openness means that anyone can freely access, use, modify and share. The essence of it is the change from mechanized thinking to big data thinking, and the change from zero-sum game to cooperation and win-win.
“Open source software is the earliest and most successful practice of open innovation in the technical field.” Mei Hong believes that open source software has completely changed the global software landscape, allowing people to achieve popular coordination, open sharing, and continuous evolution. With changes in the entire industrial structure, open source has become an irreversible trend.
At the same time, the rapid development of AI is also inseparable from the open source and data. High-quality open data has promoted the rapid advancement of deep learning algorithms, and the framework of deep learning has greatly improved the efficiency of the algorithms, complementing each other.
“Open source is conducive to competing for more user resources and building its own ecology. At the same time, open applications can quickly obtain a large amount of data, which in turn can further optimize the model algorithm and achieve important feedback.” Mei Hong said, therefore Open source has become an indispensable and irreplaceable strategic resource for enterprises.
05 Hu Houkun: Like Shanghai hotline 12345, and propose a new model of industrialized AI development
This year is the first time that Huawei’s rotating chairman Hu Houkun has come to speak at the World Artificial Intelligence Conference.
He shared a few findings during his visit to the exhibition hall. In the past few years, when you watched AI exhibitions and demonstrations, you could always see cool technologies and applications, but this year I feel that there are not so many cool things.
“This also means that these technologies have gradually penetrated into various industries and have been applied. This seemingly unremarkable change reflects the transformation of AI from tangible to intangible, changing all walks of life like a spring rain moistening things silently.” Hu Houkun Exclaimed.
He mentioned that the current Shanghai citizen service hotline 12345 has a very fast response speed. The reason behind this is that in recent years, the service hotline has begun to introduce AI technologies, such as speech recognition, emotion perception, and automatic order dispatch. With the blessing of these technologies, the entire The response speed of the service hotline has been greatly improved.
He believes that AI’s changes to various industries are now in full swing, and AI is now at a critical point in time. In order to speed up the implementation of AI, Hu Houkun made three suggestions:
The first is to bring together the forces of all parties to vigorously develop a new type of urban infrastructure represented by the AI computing center, so that AI computing power will become a new urban public resource like water and electricity, making intelligence within reach .
“In the era of digital economy, computing power is like water and electricity. It must become an available and affordable resource.” Hu Houkun mentioned that not having sufficient AI computing power is like not having water and electricity, which will greatly restrict cities. The process of digitization.
In the early stages of reform and opening up, people often said that “build roads before you want to get rich”. At the time of the development of the digital economy, the construction of a new type of urban infrastructure represented by the AI computing center is particularly important. Huawei must make such efforts. Let computing power become a widely available and affordable public resource.
The second is that the technology must be rooted in order to grow strong. The industry must vigorously develop technologies such as processors and AI computing frameworks to enhance the development resilience of the AI technology industry.
In the past two years, Huawei has built a relatively complete AI industry ecosystem relying on Shengteng’s basic software and hardware platforms. For example, in terms of hardware, Huawei has more than a dozen partners that have launched AI system hardware based on Shengteng module boards; in terms of application development, Huawei has more than 500 partners and developed more than 600 based on Shengteng. The application of AI.
The third is to boldly apply technological means to break through the current bottleneck of AI inclusiveness. In Hu Houkun’s view, the current bottleneck of AI popularization is not technology or application, but development efficiency.
“The current development efficiency is too low, and the AI application development is too slow, which seriously hinders the combination of technology and demand.” Hu Houkun explained that the current AI development model is still more traditional and requires a high degree of reliance on experts and data, and it not only takes a lot of time to collect data. , Processing data, it also consumes a lot of manpower to optimize the parameters, it is difficult to meet the requirements of productization in a short time.
Huawei calls this model a manual workshop model, and Huawei hopes to change this method through technical means to improve development efficiency.
To this end, Huawei has developed a “new industrialized AI development model”. Based on a super-large-scale pre-training model, developers can quickly develop AI models with higher accuracy and greater generalization capabilities with only a small amount of industry data.
In order to support this new development model, this year Huawei and its partners have launched a large pre-training model called the “Pangu Series”, which includes the industry’s first Chinese ALP model with both production and understanding capabilities, and a general vision model.
“We hope that more industry partners and companies can use such a large model to realize the rapid development of AI applications, so that AI can truly benefit thousands of industries, so as to achieve the prosperity and development of the AI industry.” Hu Houkun said.
06 Two major manufacturing giants talk about AI reform, Dong Mingzhu talks about the perfect combination of the Internet and industry
Roland Busch, Chairman of the Board of Directors, President and CEO of Siemens AG, shared his views on digital transformation and AI development via remote online.
He believes that AI plays an important role in this change, but to succeed in the AI field, the industry must have three conditions. The first is to have industry knowledge and experience, the second is to establish trust and network information security, and the third is to develop collaboration and build an ecosystem.
Dong Mingzhu, chairman and president of Gree Electric, shared the three relationships between manufacturing and AI in a simple way.
First of all, in the manufacturing field, Gree Electric has introduced digitalization into the industrial manufacturing field since 2016. Through years of development, it has built a “black light factory”, that is, an unmanned chemical factory, which realizes full-line coverage of unmanned operation, and has a great impact on accuracy, quality and efficiency. Both have brought about disruptive changes. At the same time, a factory with tens of thousands of people now only needs more than 1,000 people.
“The changes brought by Gree Electric’s use of the Internet have allowed us to realize the era of intelligence throughout the entire industrial chain from the beginning of logistics, the entry of materials into the factory to the entry of finished products, and then to the market.” Dong Mingzhu said.
In this structural change, Gree Electric’s personnel structure has also undergone a fundamental change. The R&D personnel has grown from 800 in the initial stage to 1,600 now. At the same time, it also has 16 researchers and more than 1,000 laboratories. These are all supporting Gree. The preparation of electrical appliances in the age of industrial automation. “Intelligence has brought us even more powerful effects for the manufacturing industry.” Dong Mingzhu said.
The second is in the Internet age. “Everyone is talking about the development of rail transit and communication technology. For our manufacturing industry, the meaning of the Internet is more valuable.” Dong Mingzhu said.
For example, in the era when everyone is pursuing zero carbon emissions, Gree Electric’s air-conditioning field consumes 15% of the total consumption, and it also uses digitally developed products and technologies. At the same time, Gree Electric also spent 9 years to develop zero-carbon source technology and won the international gold award.
In Dong Mingzhu’s view, the changes brought by Internet big data to Gree Electric have not only improved their R&D technical capabilities, but also played a very good role in the interaction of various devices. So, what changes can the realization of zero-carbon technology bring about? “The first is to reflect the social effect, but to reflect the cost of use.” She said.
“So Gree in this process, I think the Internet is not not part of our traditional business, but with the traditional enterprise complete beauty combine in order to achieve real digital age.” Dong Mingzhu says.
Finally, in this new era, digitization covers the different needs of various fields. However, in Dong Mingzhu’s view, the industry often talks about intelligent rail transit, which seems to be the highlight of intelligence, but if the industry realizes automation and digitization, the industry will achieve greater take-off and meet the technological revolution. And innovation, let the world fall in love with Made in China.
07 Zhou Hongyi: The higher the digitization, the network security problems will further intensify
Zhou Hongyi, founder and chairman of 360 Group, shared his views on digital security from the perspective of network security.
“Talking about security is to better guarantee the use of new technologies.” Zhou Hongyi mentioned that digitalization has now become a national strategy, among which digital city is the core scenario of digitalization. It has not been developed to build more information and systems. It is to promote the comprehensive digital transformation of cities, and use digital technology to drive cities, empower cities, and reshape cities.
For example, in an extreme smart city, every street lamp, trash can, and manhole cover will be digitized, with sensors, and the process of city management will be transmitted to the cloud and integrated into big data.
Zhou Hongyi believes that, in essence, urban digital transformation is to build a digital twin of the city, with big data as the core, using the Internet of Things, sensors, cloud computing, 5G, AI, transformation computing, blockchain and other technologies, through the production of data , Collection, operation and empowerment, open up the digital space and physical space, and form a digital twin closed loop.
Although the digital city has countless possibilities for imagination, from a security perspective, the higher the digital level, the greater the security challenge. With the future of smart transportation, smart grid, government cloud, big data center, and hundreds of millions of equipment and terminal applications The increasingly complex and diverse scenarios and the huge amount of network and social resources will lead to unlimited expansion of network exposure attacks and exacerbate network security issues.
Therefore, in Zhou Hongyi’s view, when data accumulates to a certain extent and undergoes qualitative changes, data security will directly establish a direct connection with national security.
In addition, Zhou Hongyi also talked about the three characteristics of digital cities: everything is programmable, everything must be interconnected, and big data drives business. “In the final analysis, it is a software-defined city, and the urban architecture is based on software.”
And when the entire city is on the structure, network security is not an optional auxiliary function. It will become an important infrastructure. It is necessary to plan the digital construction of the city and the construction of network security simultaneously to build a digital city and a smart city. Safe base for escort.
Zhou Hongyi revealed that while 360 has spent 15 years doing free antivirus, it has also invested more than 20 billion yuan to build a security capability framework with security big data analysis as the core, and build a city-level capability for digital cities with complete, operational, and operational capabilities. A growthable operating framework.
08 Shangtang CEO Xu Li: If you send a soul torture to AI, will the machine guess it?
In the Pratt & Whitney AI forum, SenseTime’s co-founder and CEO Xu Li brought up an interesting topic-“Can machines guess?” He thought the answer was yes.
In Xu Li’s view, a large part of AI development is based on guesswork. For example, when people solve falling objects from a high altitude, it is impossible to have so many data on falling objects. When solving community care issues such as fires and elderly falls, there are often zero or small data. Machines can only do it with general technology. Derivation and generalization, conjecture about these scenarios.
Therefore, in many cases, machines cannot have complete control and iteration, so SenseTime has launched a controllable, green, sustainable, and people-oriented AI ethical development concept.
“We need to balance different governance frameworks with a development perspective, and choose different policies at different stages, in order to use AI innovation to drive AI inclusiveness and thus affect more industries.” Xu Li said.
Regarding the topic of AI inclusiveness, Shen Nanpeng, Global Managing Partner of Sequoia Capital and founding and managing partner of Sequoia Capital China Fund, also shared his application of AI technology through remote connection, especially in the field of life. Views on the change and development of consumption patterns.
Shen Nanpeng said that if the computing power level and the image of the application scene are regarded as the two legs of AI in the realm of life, then the computing power is very long and strong, and it increases exponentially. However, the legs of the application scenario are relatively short and thin, and are still in linear growth. There are still a large number of subdivision scenarios for food, clothing, housing, and transportation that need to be explored. There is still much room for improvement in data mining in life scenarios.
On the one hand, the closer the integration of AI and home life, the more life scenarios are explored, the greater the industrial value of AI; on the other hand, service AI is also expected to improve future interactive experience.
For example, people often see service robots. At this stage, they cannot fully understand the collective expressions and emotional language of consumers. Therefore, in some scenes such as complaints about customer service and social companionship, users have a sense of rejection and distrust of AI. This requires Explore the potential applications of AI in emotional computing.
At the same time, AI also brings more possibilities to solve the difficulty of seeing a doctor. Shen Nanpeng mentioned that the current relatively mature application of AI in the medical field is the auxiliary diagnosis of AI images, but other fields such as online consultation and health management are still in the active exploration period.
However, the many possibilities brought about by AI+medical care have given ordinary people a lot of room for expectation and imagination when it is difficult to see a doctor, which inspires entrepreneurs and researchers who devote themselves to medical AI to go forward and succeed. “It can be seen that this is an important direction for future AI applications at the application level.” Say.
Sun Zhengyi, the representative president, company executives, chairman and CEO of the Software Banking Group, which is also connected remotely, talked about the impact of AI development and revolution from the perspective of investment.
He believes that with the application of AI, a large number of industries will usher in a transformation in the future, making the entire human society better and improving the happiness of all mankind.
However, the development of AI and the advent of industry revolution cannot be driven by inventors alone, but also by capital. Someone needs to provide capital to bear the risk of possible investment losses. “Especially under the premise that the technological prospects are still unclear and the technology can be successful, there is a risk that someone is needed to provide capital.” Sun Zhengyi said.
He mentioned that among the companies invested by the SoftBank Vision Fund, more than 90% or even 95% of the companies have not yet achieved profitability or are even losing money. Despite this, Softbank still insists on providing capital for the high-risk stage of the company’s entrepreneurship, supporting outstanding technology and outstanding entrepreneurs, sharing risks together , and jointly creating a better vision and future. “This is also my original intention to create the Vision Fund.” Sun Zhengyi said.
09 Conclusion: AI is accelerating into the homes of ordinary people
The World Artificial Intelligence Conference will continue for 2 days. This morning the opening main forum is the most interesting and the highest standard. From it, we can not only see the unique insights from all walks of life on the development of artificial intelligence and the digitalization of the industry. , You can also see the technological progress of related companies.
With the continuous innovation and development of various industries in our country, the digital and intelligent transformation has gradually penetrated into the strategic thinking of every technology enterprise. With the efforts of the entire industry, AI is stimulating the development of the entire social economy from all dimensions, bringing new momentum to technological innovation and integrated applications in thousands of industries.
However, it is undeniable that the current development of the AI industry still needs to break through the efficiency bottleneck of application landing and development. How to make better use of new technologies to achieve cost reduction and efficiency increase in production and operation has become a constant need for every player on this track. Problems to explore and solve.
In the future, we also expect AI to truly fly into the homes of ordinary people and become a blossoming existence in our lives.
Posted by:CoinYuppie，Reprinted with attribution to:https://coinyuppie.com/ma-huateng-talked-about-the-space-program-robin-li-took-a-photo-of-the-mars-rover-and-4-digital-people-were-invited-to-host-the-waic-opening/
Coinyuppie is an open information publishing platform, all information provided is not related to the views and positions of coinyuppie, and does not constitute any investment and financial advice. Users are expected to carefully screen and prevent risks.