On January 22, the annual meeting of the World Economic Forum in Davos, Switzerland was getting better, and the intelligent economy led by artificial intelligence attracted wide attention of attendees. The forum with the theme of "the next decade of the Smart economy: starting with Trust, better than the pattern" was held at the local Ice Village conference center, and heavyweight attendees from around the world discussed the topic of "trusted AI" in depth.
Zhou Bowen, Vice President of JD.com Group, President of JD.com Yun and AI, Chairman of the Technical Committee of JD.com Group, Daniela Rus, Vice Dean of Kay Firth-Butterfield,MIT School of computer Science and Director of computer Science and artificial Intelligence Laboratory, Director of artificial Intelligence and Machine Learning at the World Economic Forum, and Professor Xue Lan, Dean of Schwarzman College of Tsinghua University, attended the forum and shared their unique views. The forum is hosted by Nicholas Thompson, editor-in-chief of Wired magazine.
"trustworthy AI": the Business Ethics Center of Intelligent economy
From a global point of view, a new round of scientific and technological revolution and industrial revolution led by artificial intelligence is in the ascendant, technology is accelerating the integration with the real economy, and a new economic form of intelligent economy is taking shape. In this process, the trust of different stakeholders in the development of artificial intelligence will have a profound impact on the future of this new economic form. In the face of the rapid development of artificial intelligence, many people have a worry: can the future of artificial intelligence be trusted?
Looking back on the process of artificial intelligence being recognized by the public, Zhou Bowen believes that it was the man-machine war between the deep blue supercomputer and the Watson cognitive computing system promoted by IBM in the early years that made artificial intelligence known to the public. On the one hand, it shows the great progress made by artificial intelligence technology, but on the other hand, it also arouses people's concern. Many people will worry that artificial intelligence is becoming ubiquitous, how to manage it? Will artificial intelligence replace people's jobs?
Zhou Bowen believes that people began to do artificial intelligence research 50 or 60 years ago. Artificial intelligence starts from imitating human beings, but it still has to serve human beings after all. Today's artificial intelligence is still in the early stage of development and still depends on a lot of human support to a large extent. "if we think of artificial intelligence as the ocean, humans are still just wandering on the beach, and there is still a lot of work for us to do. Artificial intelligence is not a sprint, but a marathon. " Zhou Bowen said so.
Zhou Bowen is also optimistic about the governance of artificial intelligence: "Don't worry that supervision will not be able to keep up with the pace of technology." Ford produced the first Model T car in 1908, and it was not until 60 years later that safety regulations were introduced one after another, which did not prevent people from enjoying and benefiting from technological change. "
Recently, face recognition has become the most concerned artificial intelligence application. It is this application that has become one of the hot topics among the guests present. Eastern and Western views and disputes mainly focus on the concern and goodness around face recognition technology: Nicholas Thompson expressed concern about the challenge of face recognition technology to non-central database and the problems that may be caused by the difference in recognition accuracy caused by feature differences. Daniela Rus believes that in view of skin color differences, equal accuracy identification of skin color can be achieved by adjusting training data sets and other setting strategies.
At a time of heated discussion on the topic, Zhou Bowen also shared his views. Zhou Bowen believes that Chinese police are using face recognition technology to find lost children, by recognizing adult facial information to match childhood photo information and correctly identify, this is a very good application case. From this point of view, the practical application of artificial intelligence technology has far exceeded people's understanding of technology perception.
But Zhou Bowen also said that artificial intelligence is not only limited to the application of perception layer, but also includes cognition, reasoning and decision-making. people should not only pay attention to the surface of the application of face recognition, but also pay more attention to the in-depth application of artificial intelligence in the industry, and explore how to apply artificial intelligence technology to help enterprises improve efficiency.
Zhou Bowen gave the example of Aunt Qian. Aunt Qian is a small pork shop run by a mother and daughter in Guangdong, but it has been determined not to sell overnight meat from the beginning. This is a challenging goal because decisions need to be made about how much to buy every day and when to start a discount. JD.com artificial Intelligence, in cooperation with the supply chain team, designed an intelligent system for Aunt Qian, using online residential data, traffic, weather and other information, using a pricing strategy based on reinforcement learning. to help the butcher shop expand from Guangdong to the Yangtze River basin, with annual sales of more than 60 million yuan. As a successful case of reliable AI, JD.com helps businesses improve their user experience and further reduce social consumption through technology empowerment.
"We have helped Aunt Qian, but I am also asking myself that there may be millions of Aunt Qian in China, as well as Aunt Li and Aunt Zhang. How can we help them?" It is based on this idea that JD.com established the artificial intelligence platform NeuHub, opening JD.com 's AI technology to all developers. "I may not be able to help every store, but I can help build an ecosystem that brings in more developers, startups and academics and works together to build solutions for more stores."
In Davos, Zhou Bowen held discussions with experts from all over the world. "despite cultural differences, I see more consensus on reliable AI between China, the United States and Europe. The people I come into contact with agree that artificial intelligence needs fairness and responsibility, which is a kind of value consensus. I think this is very important. I think artificial intelligence is more valuable if researchers, developers and users of artificial intelligence have to reach a consensus on value. "
As a technical expert, I think reliable AI should also include robustness, replicability and interpretability. These dimensions are also challenging for us. Reliable AI is not only a slogan, not only a social value of technology, but also a great challenge to technology. " Zhou Bowen said so.
JD.com 's practice of Technology with temperature
Positioned as a supply chain-based technology and service enterprise, JD.com began to invest continuously in technology more than a decade ago, constantly improving customer experience. JD.com 's own rich scenes, coupled with JD.com 's profound technology accumulation and solid infrastructure, enable JD.com to achieve deep integration in the fields of AI technology, big data, cloud computing, Internet of things, 5G, etc., with AI and big data as capabilities and assets, JD.com Cloud and JD.com IoT cloud + end combination mode, the use of 5G technology, on the one hand, to provide technical and service support for JD.com 's own retail, logistics, financial and other scenarios. On the other hand, the accumulated experience and results will be empowered and exported to external partners.
When it comes to JD.com 's technology strategy, Zhou Bowen can be summarized as "ABCDE", that is, AI (artificial Intelligence), Big Data (big data), Cloud (Cloud Computing), Devices (IoT) and Exploration (Frontier Exploration). "there is a symbiotic relationship between businesses. If you think of the whole technology as a person, AI is the brain, IoT is the end of the nerves, clouds are the torso, muscles and blood vessels of the body, and big data is blood and oxygen." Zhou Bowen said, "JD.com 's AI, IoT, Cloud and big data combine to form an organism with stronger ability, more vitality and competitiveness."
Zhou Bowen believes that JD.com 's technology has temperature. He cited JD.com 's emotional intelligence customer service as an example. During November 11, 2019, JD.com NeuHub artificial intelligence platform accumulated 29.8 billion calls, emotional intelligence customer service accumulated more than 34.16 million consultations. On November 11, intelligent customer service independently solved more than 6.59 million consultations, covering the whole scene of pre-sale, in-sale, after-sale and logistics. The application of artificial intelligence in the field of customer service not only improves the response speed and service quality, but also frees the artificial customer service personnel from the heavy workload and can focus on providing more personalized services for users.
In the field of big data tagging, JD.com specially developed a "JD.com micro-worker" data tagging system for the disabled, so that the disabled can easily participate in the intelligent tagging work of AI machine learning at home. The micro-engineering project can bring an income of 100-200 yuan to the participants every day, which is very helpful for the employment and placement of the disabled in poor areas. Since 2017, JD.com micro-engineering data indicate that the poverty alleviation project has taken the lead in Ningxia. So far, JD.com micro-engineering project has landed in Ningxia, Shandong, Hebei, Heilongjiang, Guangxi and other places, attracting more than 600 jobs.
During his stay in Davos, Zhou Bowen also attended the Deloitte Davos Live event. Zhou Bowen believes that in view of the current hot discussion and practice of industrial 4.0 or artificial intelligence applications and other new technologies, industry leaders should first make clear their goals and expectations, and the most important thing is to establish top-level design. Zhou Bowen introduced JD.com 's experience, giving priority to three levels of issues when applying new technologies such as artificial intelligence, big data, cloud computing, and the Internet of things: the first is to reduce costs for users; the second is to help improve the efficiency of workers in production, to help managers have a larger body of knowledge and the ability to make better decisions; and the third is to enhance user experience.
Zhou Bowen said that the essence of JD.com 's technology strategy "ABCDE" is to promote the superposition of values brought about by technology integration, to further integrate with industry Know-how, to build a good technical cooperation ecology with partners, and to always focus on use cases to solve real scene problems, so as to ultimately create more value for the society.
Sina stated: all the minutes of the meeting are arranged in shorthand on the spot and have not been reviewed by the speaker. The publication of this article on Sina for the purpose of transmitting more information does not mean agreeing with its views or confirming its description.