In this outbreak, we once again saw the role of big data and artificial intelligence technology. According to media reports, Canada’s artificial intelligence startup BlueDot is one of many companies that use data to assess public health risks. During the Ebola virus outbreak in 2014, the company used global flight data, urban health management systems, population movements, etc. Cross-analysis of the data provides a new reference for epidemic prevention information.
In addition, artificial intelligence has also been applied to patient consultation and management. On January 29, when doctors from Providence Regional Medical Center in Washington were treating the first confirmed case of new coronavirus pneumonia in the United States, they did not interact face-to-face with the patient, but applied a method called Vici’s robot interacts with patients through the screen. This telemedicine robot looks like a tablet on wheels, which doctors can use to talk to patients and perform basic diagnostic operations, such as measuring body temperature. “Although medical staff provide care in the isolation ward, technology has enabled us to reduce the number of close interactions with infected patients, and can protect medical staff from infection.” Amy Komp, Chief Clinical Officer of Providence Regional Medical Center Dayton-Phillips (Dr. Amy Compton-Phillips) said.
A hospital in Guangdong also started to use self-disinfecting artificial intelligence-driven robots to treat patients during the fight against the epidemic, and told the media, “The robots began to provide medicines and food to patients yesterday, and collect bed sheets and medical waste. “The purpose is also to use automatic driving and automatic chargers to reduce the workload of medical staff and reduce the risk of cross-infection. Technology companies have also joined the ranks of research and development to fight the epidemic, and Alibaba and Baidu have provided artificial intelligence gene sequencing tools. Chinese epidemiologists are currently developing vaccines and other treatments for the new coronavirus. According to a report from the Chinese Center for Disease Control and Prevention (CDC), researchers have isolated virus strains for vaccine development. In Australia, scientists are doing similar work, where they recreated the virus and shared the genome sequence.
Not long ago, Sun Zhengyi of Softbank Group once again emphasized the correctness of the trend of investing in artificial intelligence in the future. He foresaw that the scale of the Internet of Things in 2035 will be 100 million times larger than it is now, and society will further develop from the Internet era to the next era of artificial intelligence. Sun Zhengyi firmly believes in the need to continue to invest in the future of artificial intelligence technology? The future of self-driving cars is safe and efficient, and the accident rate is low; the future of artificial intelligence will help us solve most of the disease problems; the future of artificial intelligence robots will participate in rescue and rescue operations, and accompany the elderly; the future of artificial intelligence will ensure food security, and happiness …… unique
artificial intelligence qualitative breakthrough path
long ago, investors Victor gave me a micro letter forwarding the news, “the next inflection point of artificial intelligence: neural network diagram usher in rapid outbreak. “The Graph Neural Network (GNN) technology he mentioned has received great attention and response in academia, and has been continuously extended in business application scenarios, involving computer vision, 3D vision, natural language processing, and scientific research scenarios. The innovative use of multiple value scenarios, such as the knowledge graph recommendation system, and even financial anti-fraud. It can be said that people maintain a strong scientific and technological curiosity for the research of algorithms based on graph relationships. We continue to discover the rules that exist in nature, and use them to turn them into science. However, although the scientific discoveries made so far are great, they are still not omnipotent. “Intelligence based on statistics and operations research has fatal birth defects.” I replied to Victor in this way, and hope that today we have a more rigorous and rational perspective on intelligence.
Artificial intelligence is still only explaining the “relationship” of two events, and it is still unable to infer the “causality” between the direct events from the correlation between the two events. Simply put, what intelligence can give us is the degree of relationship between the two. As for whether there is necessarily a “causal” conclusion between the two, current intelligence cannot give us. There are still a lot of uncertainties in the field of intelligent innovation, and the world’s top technology companies are making continuous investment, challenges and breakthroughs.
In the era of artificial intelligence, humans have completed the upgrade of their intelligence in their interactive methods. For example, the way of human-computer interaction is no longer a relatively rigid process such as the previous button commands. When voice recognition, text recognition, and image recognition technologies develop to more mature At that time, human-computer interaction has begun to improve in a more humane way. Through the upgrade of the interactive process, we not only allow the machine to communicate with people by voice dialogue, but even in a large number of factories, we have seen “physical” robots, and the system replaces people for process manufacturing through strict rules. The interactive communication method of human-machine intelligence has evolved from human-machine to machine-machine, and even the future transformation of machine-human interaction.
Especially in manufacturing factories driven by Industry 4.0, a large number of industrial robots have become the norm. For example, the establishment of the once glorious Volkswagen Phaeton “transparent factory” and the Jinqiao factory of SAIC GM, where there are more than 300 robot arms, even Globally, there are no more than 5 factories of this level. In the huge workshop, there are only more than 10 workers. They manage 386 robots and cooperate with them to produce 80 Cadillacs every day.
Obviously, human exploration of intelligence is far more than breakthroughs in interactive capabilities. In the next stage, we believe that the qualitative breakthrough of intelligence lies in how to build the thinking ability of machines, that is, imitating the operation of the human brain to help humans deal with a large number of problems. This is another dimension of change. In this dimension, the system also needs to continuously strengthen three aspects of ability, namely, understanding ability, reasoning ability and learning ability. These breakthroughs in capabilities will bring more industrial leaps in human history, but in the process it is bound to find more suitable scenarios to allow artificial intelligence to “train”, and humans will begin to enter the artificial intelligence scene. era.
In this era, more attention should be paid to the landing of artificial intelligence scenarios, to create the best industry cases, so as to form scenario intelligence. In the era of artificial intelligence scenarios, experts who understand both technology and industry are needed, but the mastery of technological capabilities is often only the basis. How to play the value of industry scenarios depends more on how to make industry problems through technology empowerment. Converted away, thereby creating more value.
With the popularization of the basic capabilities of artificial intelligence technology, a large number of companies will slowly shift their attention to specific corporate issues, such as how to establish stronger trust and connection with consumers? How to predict the life cycle of important equipment in advance? How can we more accurately correlate the pattern of events? And through the use of big data to find solutions to more scenarios. Although, at this stage, we cannot completely rely on machines to give us answers, but data-based decisions must be more reliable, controllable and sustainable than business perceptual decisions. Moreover, it is certain that under big data, business forecasts will be more accurate and closer to facts, and this value is sufficient to help companies build their own core competitiveness. This ability is not entirely dependent on one person, but is more controllable, thus forming an industry-based cognitive monopoly based on data.
What kind of consumer products are being consumed faster today? Only Taobao knows. Important business problems will continue to drive the production of technical solutions, so that some business problems will continue to be circulated and improved.
Embrace adaptive intelligent
AI further evolution is adaptive intelligence. The characteristic of adaptive intelligence is to make commercial attempts at a fixed frequency. In fact, we have seen today that some technology companies are investing heavily in the research and development of artificial intelligence-related technologies, and are first used to upgrade their products and service experience. For example, Microsoft’s artificial intelligence and R&D departments have invested more than 8,000 people in the field of artificial intelligence, and the R&D investment has exceeded 10 billion U.S. dollars. It not only unites Bing, Cortana, Microsoft Information Platform Group, environmental computing and robotics, but also strategically acquires natural language. Schedule startup Genee and deep learning startup Maluuba to accelerate its artificial intelligence capabilities. Google and Apple also invest billions of dollars in research and development every year. These AI technology benchmark companies often conduct commercial experiments at a fixed frequency to gain the traction for innovation and transformation, thereby forming adaptive intelligence. These technology companies actively participate in the changes of the times. They are not driven by a specific problem, but hope to build the innovative power that leads the times, so as to better adapt to the next era.
However, we can also see that the cost of adaptive intelligence is often not low. Google’s artificial intelligence company DeepMind continues to lose money, from 341 million U.S. dollars in 2017 to 570 million U.S. dollars last year, and Google sold to Japan Softbank in 2017. Boston Dynamics (Boston Dynamics) is also an expensive luxury.
In fact, the creation of adaptive intelligent capabilities is not a blind pursuit of innovation. The company will automatically adjust the frequency of attempts according to the environment, thereby establishing better intelligent protection and intelligent profitability. However, most commercial companies are actually not good at using information technology to innovate, because they pay more attention to their own business development issues, especially in some industries, more attention will be placed on peer competition and business continuity In terms of development and change, as for technological changes, there is actually a lack of research and patience.
Not long ago, I helped a real estate company to transform. It hopes that it can transform from a traditional real estate company to a smart city operation service company. The most important thing in the transformation seems to be the need to quickly build smart city technology understanding capabilities and use technology to City services are exported, so as to benefit from the value-added services of the city. However, the road of change is not so smooth. When it is necessary to “revolution” a company’s best profit model and turn it into another unfamiliar profit model, any company will resist because of the uncertainty here. Too big. Therefore, in the end, it is still the business model of technology serving real estate, and in the end it will help them acquire land and sell houses.
If an enterprise cannot fully resolve to carry out model changes, then the building of adaptive intelligence can often only rely on external forces. Technological innovation that pays attention to cost-effectiveness is not necessarily cheap. We often see that many commercial innovations do not necessarily rely on their own technological transformation, but rely on external forces to accomplish the right things at the right time and keep up with the trend of the times. And its own intelligence department has become a window to connect innovation-more innovative intelligence that keeps pace with the times depends on professional technology companies to invest, and to ensure quality and effect.
For example, cloud computing technology, if it were 10 years ago, I would think that cloud computing technology is only a research and development innovation direction, but today it has become an inevitable trend of informatization in all industries. If an enterprise hopes to establish an innovation environment for cloud computing technology by itself, and then move its business to the cloud, the result will be either too much investment in the early stage and an abrupt halt, or the cost of later technical operation and continuous innovation will be too high and progress will be slow. . In order to adapt to the environment and establish the adaptive intelligence of the enterprise, it is actually a good choice to connect with highly R&D intelligent technology companies, reduce the uncertainty of innovation, and then continuously adjust the frequency of innovation attempts according to the maturity of the business.
Reinforcement Learning can solve the problem of maximizing the goal of the agent through learning strategies in the process of environmental interaction. This goal can make the basic logic of the agent safe.
Establishing adaptive intelligence is like people need to use external professional medical institutions to continuously develop new antibodies to resist the virus. In fact, when people are not exposed to large-scale overt harm, these professional institutions are constantly trying to ensure the continued health of humans. Service, R&D and innovation. These new drugs and devices that have passed clinical trials and are qualified for the market will help human beings to carry on life continuation and social value innovation more confidently and healthily.
Therefore, when we face the future intelligent era full of uncertainties, whether it is good or not, we need to ensure that we have established strong intelligent adaptive capabilities and can quickly adapt to the requirements of the new era, and companies need to make sure that they are not Will slowly die out along with the original business model.