Author: Liu Jia
Suppose you have a robot in your home, and while you are sleeping, the robot turns on the vacuum cleaner and starts working. You wake up and tell it angrily, “You shouldn’t have woken me up!” You’re saying that turning on a loud vacuum cleaner while you’re sleeping is wrong.
At this time, the robot needs to understand: when you are not sleeping, it can vacuum; when no one is at home, it can also vacuum; when the vacuum cleaner is silent, it can still vacuum.
But in fact, it is difficult for the robot to “think” the cause and effect between “wake up” and “vacuum cleaner”, and may even interpret your complaint as not being able to use a vacuum cleaner to clean.
This is what Judea Pearl, Turing Award winner and “Father of Bayesian Networks”, cites in “The Book of Why: The New Science of Cause and Effect” The example of , intuitively illustrates that many passwords that are extremely short for us humans actually contain rich content. For robots to “get” these “hidden” amounts of information, they must learn causal reasoning and understand causal relationships.
Source: “Why: The New Science of Cause and Effect”
1. When AI understands cause and effect, technology has temperature
Nowadays, humans have become more and more accustomed to directing artificial intelligence to serve themselves, such as planning driving routes through in-vehicle voice assistants; directly communicating with artificial intelligence voice customer service during online shopping; issuing voice commands to smart home devices to obtain audio-visual entertainment and weather information , Control home appliances, etc.
Gartner, a well-known consulting firm, predicted in 2019 that in 10 years, the total amount of daily language communication between humans and smart devices may account for one-third of our daily language communication.
From small sweeping robots and intelligent customer service, to medical and financial systems, intelligent products and upgrades are everywhere. Although significant improvements have been made in reducing costs, increasing efficiency and enriching life, most of them still remain in data intelligence and process intelligence. For the service industry with more complex interaction scenarios, more emphasis on process experience, and more need for human thinking, the current The level of artificial intelligence is far from enough. For example, in the result-oriented sales operation transformation scenario, the “intelligence” of artificial intelligence is higher, and lasting and stable effect delivery is required.
Although artificial intelligence has made great progress in perceptual intelligence after more than half a century of development, a large part of the achievements of neural networks with deep learning as the mainstream rely on massive data collection and high-performance Intensive computing hardware. This leads to a saying in the field of machine learning and artificial intelligence – “As much as there are artificial intelligence, there is as much intelligence.”
Moreover, as relevant applied research enters the “deep water area”, the limitations of machine learning for correlation calculation based on statistical data are gradually becoming apparent.
As an authoritative expert in the field of artificial intelligence, Judea Pearl once clearly expressed in his book that he opposed the view that “deep learning only needs correlation and not causation”. He believes that in order to achieve strong artificial intelligence and even transform intelligent machines into morally conscious organisms, machines must learn to ask “why”.
Judea Pearl also proposed a “three-level causal ladder”:
-
The first layer: Association, which involves statistical correlations defined by the data, is where most machine learning systems operate. For example: people after a headache, according to the information collected on the Internet, most people take aspirin to relieve the headache.
-
The second layer: Intervention, which supplements and reprocesses the data by doing some kind of experiment, and puts forward some hypotheses before the experiment. For example: If you take aspirin, will your headache be good?
-
The third layer: Counterfactual, which is the reflection and abduction of what happened in the past, allowing machines to have “imagination ability”. For example: Did aspirin really cure my headache?
Source: “Why: The New Science of Cause and Effect”
Most problems in the field of artificial intelligence are decision-making problems, and the transition from correlation to causality will lead the way of thinking in AI to be advanced.
As the only AI company that explores and innovates based on the causal AI theory, Zero Rhino believes in “using technology to warm people’s hearts”, not as a tool for efficiency, but to serve the effect, so the company’s AI technology also continues to “focus on users” , under the concept of “focusing on results”.
2. Based on causal AI, explore the commercialization path
Founded in 2018, Zero Rhino was born out of Baidu’s cognitive intelligence technology team, and has successfully completed hundreds of millions of C rounds of financing.
Based on the causal AI theory, Lingxi Technology has innovatively built a causal AI service architecture that combines business practice and technological innovation, created a “user demand inquiry” system, and took the lead in cutting into sales scenarios, in the service sales process. Constantly “understand and stimulate user needs”, deeply analyze the complex interference factors of users’ purchasing decisions, and iterate the “demand model” and “effect model”.
Specifically, AI optimization is realized from three levels:
-
The first layer: Based on the depiction of past objective data based on big data, the user’s subjective data is inquired and digitized;
-
The second layer: replace correlation with causality, use correlation to “guess what you like” on the Internet, and use causal inquiry to “dig out your needs”;
-
The third layer: continue to explore “strong” artificial intelligence, starting from the user, through the user’s “open expression” to explore their real needs to accurately meet.
For example, as an excellent representative of going deep into the insurance value chain vertically and helping the insurance industry to transform and innovate, Zero Rhino’s insurance sales transformation model based on causal AI can stabilize output capacity and provide direct-effect services for many partners.
According to the introduction of Zero Rhino, its insurance sales transformation model compared with the pure human model, the performance of the human-machine model can reach 6 times; its cognitive pipeline BPO has an average increase of 36% in the insurance telesales transaction rate compared with the traditional method; and Lingxi also performed well in various process data. The daily outbound call data volume increased by 200% compared with the traditional, the daily call volume increased by 150% compared with the traditional, and the daily return visit call volume increased by 500% compared with the traditional.
Zerorhino Technology won the “Excellent Insurance Digital Marketing Pilot Award”
In the cooperation with JD Digital, Lingxi, relying on its self-developed enhanced intelligent platform + distributed operations, realized remote control and operation of outbound call pipelines during the epidemic, and completed various requirements of JD Digital with high quality. Customers work together to overcome the difficulties. Anhui Xinchen, a wholly-owned subsidiary of Zero Rhino, has won the Excellent Cooperation Award from many brand companies such as JD Digits and Meituan.”
Up to now, Zero Rhino has more than 10,000 people and machine fusion service armies, which are distributed all over the country. Its “user demand inquiry” system, cognitive pipeline, enhanced intelligence platform and other products have provided direct sales conversion services for more than 100 well-known enterprises, covering dozens of fields including insurance, finance, knowledge payment, education, etc. This field has more than 440 million users.
With its strong technical strength and meticulous service, Zero Rhino has continued to expand its influence, and has gained support and recognition from capital, customers, and authoritative media organizations.
In the past year, Zero Rhino has won the “Leifeng.com 2021 Best AI Digital Intelligence Annual List-Causal AI Best Practice Innovation Award”, “The Most Influential Innovative Enterprise in Lieyun.com’s 2021 AI Scenario Landing” Award”, and was also selected as the “WISE 2021 New Economy King Annual Hard-core Enterprise” by the authoritative media in the field of venture capital and new economy, and the “2021 Synced Machine Intelligence AwardsAI” by the Heart of Machine “The most commercially valuable solution in the field of technology and practice” “.
It can be said that based on the practical practice in the field of cognitive intelligence and the in-depth thinking of causal science, Zero Rhino Technology has successfully combined theory and practice, not only has the unique advantage of “paying for results”, but also in the commercial society. Continuously verify the true value of technology.
3. Bless the service industry and be a quality booster
Nowadays, “digital intelligence” is the general trend of all walks of life. However, compared with the “hot” transformation and upgrading of technology and manufacturing, the service industry appears to be “insufficient”.
But in fact, the service industry is the most important support for the rise of my country’s economic scale.
In 1978, the GDP of my country’s primary, secondary and tertiary industries accounted for 28.2%, 47.9% and 23.9% respectively. By 2021, the proportion of GDP of the primary, secondary and tertiary industries will become 7.3%, 39.4% and 53.3%, and the contribution rates to GDP growth will be 6.7%, 38.4% and 54.9% respectively.
In other words, over the past 40 years, the primary and secondary industries have dropped by about 20 and 10 percentage points respectively, while the tertiary industry has increased by nearly 30 percentage points. The tertiary industry is basically a service industry. Whether it is the overall scale proportion or the growth contribution, the service industry far exceeds the manufacturing industry, and the sustainable development of China’s economy must also rely on the continuous expansion of the service industry.
With the rise of causal AI, AI systems are no longer just simple statistical fitting, but instead actively understand the laws and causal connections behind the development of things. The evolution from perceptual intelligence to cognitive intelligence is not only about whether the next generation of artificial intelligence can break through the existing limitations, but also is of great significance for further enriching and expanding the service industry.
Compared with traditional AI, causal AI has more application advantages in the face of more interactive, contextual and random service industries, such as:
-
Decision-making ability: make better, more human-logical decisions based on causal models, rather than predictions;
-
Strong adaptability: The service industry scenarios are complex and changeable, and causal AI has better robustness;
-
Human-machine collaboration: not just a simple “you ask me an answer”, there is more in-depth common thinking and active cooperation;
-
Counterfactual: Reflect on past behavior and imagine possible scenarios, which can undoubtedly save huge communication and trial-and-error costs for the service industry;
-
…
It can be said that the causal Al is the universal Al technical base for the intelligent upgrade of the service industry, which determines the development process and ceiling of the service industry to a large extent. As a cognitive intelligence company that integrates “user understanding theory, causal thinking and causal AI” earlier in China and serves as an Al commercialization path in the service industry, Zero Rhino has not only completed multi-dimensional innovation, but also established a service industry. A new paradigm for smart upgrades.
Based on considerable practical experience, Zero Rhino believes that in the service industry, intelligent upgrade requires the combination of technical and non-technical solutions, as well as the integration and collaboration of humans and machines.
Specifically, technical and non-technical solutions include AI technology implementation paths, incentive models, and operating rules and compliance frameworks. The human-machine integration requires that in these solutions, matching people or matching machines are embedded in the required nodes, so that the complex, multi-objective system can be operated, and the decision-making and operation of the global optimal solution can be completed.
Therefore, Zero Rhino not only conducts in-depth inquiry and demand restoration on the user side, but also deeply deconstructs service providers, service processes and service strategies. Relying on the zero-rhino transaction causality model, through the communication and understanding of target users, users are accurately classified, and then matched with different sales strategies and service capabilities, and the atomic-level human-machine model reconstruction is carried out in the service scene, so as to achieve “this time”. The optimal solution under the current situation and this situation” helps enterprises to complete the transformation results with controllable costs and better efficiency.
At present, compared with developed countries where the service industry supports at least 60% of GDP, my country’s service industry still has at least 15 to 20 percentage points of space, which corresponds to a scale of one trillion yuan, and an increment of tens of trillions of yuan. . Zerorhino Technology has rich experience in the business practice of causal AI, and has been selected as the first batch of “specialized, special and new” enterprises in Beijing in 2022. While its core strength has been recognized, it has also been entrusted by the country and society. More to look forward to.
It is believed that on the road of intelligent upgrading of the service industry in the future, Zero Rhino will usher in a broader world of development, and is expected to become a powerful quality booster for the domestic service industry.
Leifeng Network
This article is reproduced from: https://www.leiphone.com/category/industrynews/k2wBZ96Ez3LkrFvd.html
This site is for inclusion only, and the copyright belongs to the original author.