News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
The era of artificial intelligence
The era of artificial intelligence
Win in the artificial intelligence Era
2017-08-31 classification: industry information reading (1) review (0)
Introduction
The potential for artificial intelligence is becoming reality, and machines are beginning to possess the capabilities that belong to humanity. It's time to ask: how do business leaders take advantage of artificial intelligence to give full play to the unique advantages of people and machines?
straight
Until recently, artificial intelligence has made us feel closer and closer to reality. Although artificial intelligence technology has developed for some time, it has not been able to reach the very beginning of its birth. Today, the potential for artificial intelligence is becoming reality, and machines are beginning to possess the capabilities that belong to humanity. So it's time to ask: how do business leaders take advantage of artificial intelligence to give full play to the unique advantages of people and machines?
Artificial intelligence is rapidly becoming the basic technology in many industries, with its impact ranging from automated cars to financial transactions. Autonomous learning algorithm is now widely embedded into mobile and online services; digital devices and network sensor data processing and data transmission capacity has been greatly improved, continuously improve the performance of artificial intelligence machine; basically able to identify specific voice and image, can understand human communication. This trend is of extraordinary significance:
Because the machine can speak, read, accept and store encyclopedic knowledge, it can discuss a wide range of topics naturally with human beings;
Because the machine recognizes objects and optical images, it can move out of the virtual world.
Artificial intelligence, once a disappointment to its supporters, is now well established, and machines begin to engage in activities that humans have been able to do in the past (see Figure 1 and appendix "how machines think and operate"). For example, now that artificial intelligence can diagnose some cancers more accurately than radiologists, it is no wonder that the traditional financial, retail, medical and other industries have invested billions of dollars in artificial intelligence.
How do machines think and operate?
The three milestone made the general public aware of artificial intelligence, and each event presented some key elements of AI technology.
In 1997, the deep blue computer beat world chess champion Garry Kasparov. Chess was once regarded as a game of the core strategic elements of human intelligence, and thus became the touchstone of new AI algorithms. For decades, programmers have tried to beat artificial humans with artificial intelligence, but have made no headway. Finally, in 1997, IBM developed the deep blue computer, beating the world chess champion. But many people still feel disappointed because playing chess doesn't equate to having universal artificial intelligence. The deep blue computer relies on brute force algorithms and memory. It does not learn or handle any task other than chess.
This incident reveals two reasons. First, the way machines solve problems is different from human beings; second, many intelligence tasks have a narrow coverage and can be solved by special procedures.
As AlphaGo defeated Li Shishi in 2016, the computer dominated the board game. AlphaGo, developed by DeepMind Technologies, relies on deep learning - a neural network (also known as computer brain) - to beat the world go champion. There was an interesting story about the preparation of the game: in the last few months before the start, AlphaGo learned all the human races and played chess with him for the rest of the day.
In 2011, Watson beat the champion in variety show "dangerous edge". After winning the challenge, IBM's Watson went through the artificial intelligence Turing test. Watson in the game full of speech recognition, Natural Language Processing search and the most cutting-edge technology, but this victory is on a different skill: Watson in the "double BET method (Daily Doubles)" won the other contestants in the double - BET method, game player can bet on their own all or part to win money, in order to obtain the absolute leader. To make the best choice, players need quick sequence reasoning, game theory, and the ability to compute probabilities and results correctly. Nobel prize winner Daniel Kahneman noted in his famous book "thinking fast and slow" that humans are extremely hard at these areas, and that machines can make decisions that are informative.
In 2012, Google demonstrated the self - driving car. Google is not a pioneer in the field of self driving automobiles. The honor belonged to a German computer vision expert named Ernst Dickmanns, who used the autopilot model on the German motorway in 1995 to drive 1785 kilometers at an average speed of 170 kilometers per hour.
Dickmanns does not turn left during the automatic driving. The Frank Levy and Richard Murnane in their 2004 book co authored the "new labor division", "in the oncoming traffic in turn left will involve a lot of factors, it is difficult to imagine that a set of rules to copy the behavior of the driver." However, Google's self driving cars can successfully do this, it is the integration of computer, computer vision and real-time data processing, the intelligent agent can not only explore the real world, also can gain experience from the real world.
Artificial intelligence systems, which have the ability to think and interact, are often inevitably compared with others. Although humans can do fast parallel processing (pattern recognition), but the order processing (logic) speed is very slow, and in a few areas in computer has completely mastered the parallel processing, and can process the rapid sequence. Just as a submarine can dive without swimming, the machine has its own way of solving problems and completing tasks.
If the computer processing capacity can not make a huge breakthrough, the machine will not be able to achieve universal artificial intelligence. General artificial intelligence refers to the ability to solve many different problems simultaneously, and is the exclusive characteristic of human intelligence. For example, today's robot cars do not show the intuitive wisdom we often speak of. They do not stop on the way to help a child who falls off his bicycle. But if applied properly, artificial intelligence will be able to deal with many business activities quickly, excellently, intelligently and comprehensively.
Artificial intelligence is no longer an elective course". It is important for enterprises to know how to connect people with machines, to complement each other and to create competitive advantages.
Evolution of competitive advantage
Once upon a time, a simple technology tool could be an excellent source of business. WAL-MART's logistics tracking system in the 80s of the last century is an example. But today's artificial intelligence is different, the algorithm itself cannot bring the competitive advantage for the enterprise. Pure algorithms are spread over public platforms, and companies can easily use these open-source software platforms, such as Google's TensorFlow. OpenAI, a nonprofit organization created by Elon Mask (Elon, Musk, Tesla founder), is working to promote the widespread use of artificial intelligence tools and research. Many prominent AI researchers continue to retain the right to publish research results while joining Baidu, Facebook, and Google.
Artificial intelligence does not erase the traditional sources of competitive advantage such as market position and key capabilities, but only re defines these advantages (see Figure 2). As a result, companies need to look at their strengths in a dynamic light. For example, firms usually gain market share for a relatively solid advantage, taking advantage of the firm's unique assets, distribution networks, customer contacts, and size. But in the era of artificial intelligence, we need to redefine competitive advantage.
As to how artificial intelligence can change traditional competitive advantages, let's look at three examples:
Data. The powerful application of artificial intelligence can not be separated from data. Facebook, Google and other artificial intelligence pioneer step to build their own "advantage", is by far more than the traditional ways of data collection, access to the current and future data users and others, as a raw material for artificial intelligence applications. These companies, because of their large size, are able to deliver more training data to the algorithms and to improve their performance. For example, in the fully automated racing drive, the advantage of excellence is that you can collect 100 million miles of data per day from drivers, which can eventually be used to improve travel services. Facebook and Google can use their size and depth to enhance advertising positioning.
However, not all companies can become Facebook, Google, or best, but that doesn't matter. Companies can build, capture, and utilize shared, leased or complementary data sets, even if it means working with competitors, but can help companies replenish their assets and build their own niche. Sharing is not a bad thing, the key is to create an impeccable, unique open and closed data portfolio.
Customer contact. Artificial intelligence has changed the way customers come in contact. Whether the location of the entity store or high flow shop, have to submit to artificial intelligence generated insight. For example, large retailers can use artificial intelligence engines to analyze loyalty, sales points, weather and location data, thereby designing personalized marketing campaigns and promotions. Businesses can even predict customers' paths and preferences without customers' awareness, so that they can easily provide familiar, complementary, or new buying options. Using the implied effect of these preferential activities, firms can increase revenue while keeping costs almost unchanged.
Ability. Capabilities are traditionally divided into multiple independent sources of excellence, including knowledge, skills, and processes. The automation based on artificial intelligence integrates these fields into a continuous cycle, including three parts: execution, exploration and learning. As the algorithm absorbs more data, its output quality will be improved. Similarly, for humans, and the rapid prototype design and perfect fast feedback cross department team according to the customers and end users, the agile way of working the boundary between the traditional ability to become even more blurred.
Both artificial intelligence and agile are iterative in nature, and both products and processes are continuous cycles. The algorithm can absorb lessons from experience, and enable enterprises to combine fast and extensive exploration of unknown fields with the development of known fields, and flourish in highly uncertain and rapidly changing environment.
In addition to redefining some sources of competitive advantage, artificial intelligence can also improve decision-making speed and quality. In certain tasks, the number of input and processing speed of the machine beyond the millions of times of mankind; forecast analysis and objective data to replace the human intuition and experience, become the core of many decision factors driving. Stock trading, online advertising, supply chain management and retail pricing are moving in that direction.
Of course, even if there is a revolution similar to the Industrial Revolution (but this time the rate of subversion is certainly faster), people will not be eliminated. First of all, the development of the system still needs to be done by people. For example, the company excellent step to recruit hundreds of automatic driving vehicle experts, about 50 of them from the Carnegie Mellon University Robotics Institute; artificial intelligence experts are now the largest recruitment needs of Wall Street. Second, humans possess common sense, judgment, social skills, and intuition that the machine does not possess. Even if routine tasks are given to robots, human involvement will be needed for a long time to ensure quality.
In the new era inspired by artificial intelligence, the sources of enterprise's advantages change, and strategic problems change with organizational, technical and knowledge issues. As a result, flexible organizational structures and agility become the key to dealing with large-scale, rapid change, both for people and machines.
Scalable hardware and adaptive software are the basis for artificial intelligence systems to take full advantage of scale and flexibility. The common practice is to build a central intelligent engine and a distributed semi automation agent. For example, Tesla's auto - driven vehicle can deliver data to a central unit, periodically updating the distributed software by the central unit.
The company's winning strategy is to focus on agility, flexibility, hiring, and continuing training and education. Companies focused on artificial intelligence rarely employ large numbers of traditional full-time staff, and forms of open innovation and contractual collaboration are gaining ground. The chief operating officer of an innovative mobile bank once admitted that the biggest problem was to turn the company's leadership team members into managers who were good at managing people and machines.