Dialogue with IBM's Wang Yang: The third computing revolution behind IBM

The same place, similar date, Tencent WE Conference and APEC were held together. Compared with the political and economic significance of APEC, the second Tencent WE Conference invited representatives of cutting-edge science from around the world. Last year, someone commented that the WE Conference was "a high-profile public relations event of Tencent." Although this statement was a joke, it was not false. However, behind this technology thought summit, which may be the top in terms of scale and specifications in China, there are indeed many things to think about. Like last year, iFanr will also bring you guest interviews and on-site conference reports.

Wang Yang is IBM's global vice president and general manager of the China Development Center. He will share IBM's latest artificial intelligence Watson at the WE Conference. Compared with Google and Microsoft's artificial intelligence, Watson simulates human neurons from the hardware chip architecture. It can understand human natural language, generate hypotheses based on data, and has the ability to continuously self-learn. Watson can already provide services to smart terminal users through developers and has begun to serve some scientists, mainly in the medical field.

IBM has always played an important role in the history of human computing. In short, there are three computing revolutions behind IBM. The first phase is tabulation, inputting numbers, printing out tables, and classifying them. The second phase is programming, teaching computers our understanding and execution methods, and computers tirelessly repeat the data processing.

We are now in the early stages of the third phase, with the background of a data explosion. Data is not only structured or unstructured, but also fuzzy and has different possibilities, which is different from our previous computing methods. The requirements for computing in this stage are to teach computers how we understand the world, how humans learn, how humans understand the world, how they touch the world, how they master the world, how they improve their knowledge, and how they perceive and recognize the world. Once this is taught to the computer, all data generated will be left for the computer to learn on its own.

Watson, the IBM artificial intelligence computer that defeated top human players and won the brainpower competition show "Jeopardy!", is IBM's work in the field of artificial intelligence and is also what IBM calls its preparation for the third computing revolution.
Watson and Cloud Computing

Wang Yang recalled that the first version of Watson was a supercomputer, which has now been deployed on IBM's cloud. The idea is that IBM wants to open up a tool like Watson and provide this platform to everyone in the world, so that they can inject what he believes are the modules of human cognitive computing. It is a programmable open system deployed on cloud computing.

IBM, which ranks among the top in scientific research strength, has many Nobel Prize winners and Turing Award winners. Many of IBM's previous technologies were researched by its own R&D personnel. The change with Watson is that this platform is open and deployed on the cloud. Anyone who has good research in these areas can use this platform to process and then provide feedback.

Talking about the relationship between Watson and cloud computing, Wang Yang said that cloud is a virtualized concept. For ordinary people, there are three layers. The virtualized ones are computing resources, storage resources, and bandwidth resources. This is an infrastructure virtualization called IAS. It can be regarded as several data centers connected together, which is a very large computing power, equivalent to connecting supercomputers together.

He also believes that the entire Internet actually runs on a huge cloud, which is a virtual shared resource. When we talk about industry Internet, financial Internet or a certain industrial Internet, it is generally through a lot of SaaS (Software as a Service). For example, banks have some applications, such as core banking systems and counter systems. A company has ERP, CRM, and SCM, and these systems can also be virtualized. This is SaaS based on each industry. When application software provides services in such a virtualized form, the industry Internet is formed.

Watson is an application that is grafted onto the SaaS level. It can be said that SaaS and the underlying PaaS (Platform As A Service) are connected. It has the ability to analyze big data, conduct cognitive computing on the data generated by various industries, learn, summarize some rules, or provide some knowledge to assist industry experts in making decisions.
Watson and Neural Networks, Natural Language Processing

When it comes to Watson artificial intelligence, we have to mention neural networks and natural language processing. When introducing IBM's achievements in these two areas, Wang Yang mentioned that IBM has been researching natural language for a long time and has an internal voice project that can turn what people say into scripts.

Then there is the translation system, which translates from one language to another. After the translation is completed, it can be spoken. The machine can speak according to the text. In order to be more intuitive and the interface more user-friendly, IBM can make different voices according to different people. If you choose the voice of US President Obama, the sentence will be read in Obama's voice. As long as you say a paragraph, the pronunciation of the vowels is already known. After parsing it, you can use the machine to say anything, just like a real person speaking. This kind of technical application has already been used. AutoNavi uses Lin Chiling's voice navigation, which is a similar technology.

The next most important step is natural language recording and understanding. The first step is for the machine to record natural language, and then understand its semantics, as well as key words, and what they mean. Human language is very complex, with literal understanding and non-literal semantics, and some implicit meanings that need to be interpreted. In the early days, in order to help machines understand non-literal semantics, IBM researchers had to read Chinese and English repeatedly. In the Internet age, there is a large amount of data, and it can be summarized naturally in the form of big data analysis, making its translation more and more accurate, which is also a natural language learning process for it.

On the other hand, Watson uses artificial intelligence in the form of simulated neural networks. Speaking of this, Wang Yang said:

"Chips are all made of single crystal silicon. What are humans made of? Humans are made of carbon. The Si of single crystal silicon and the C of carbon are in the same vertical direction of the periodic table, just with one more circle. This also extends to how robots will be connected to humans in the future. There are similarities in physical and chemical structures."

IBM has launched a brain chip, which is equivalent to a single-crystal silicon simulated neuron. The intelligence of this chip is now about the same as that of a bee. Wang Yang expects that it can be stacked with several layers, and in the near future, it may even surpass the ability of the human brain. This is IBM's use of cognitive computing programming to change the way of learning, starting from the chip, from the bottom up.

Humans have extended the tentacles of IT to every corner and can sense everything. They can sense what is happening in every place, including people's sleep. The virtual world can completely simulate what happens in the real world. At this time, new innovations such as neural networks and cognitive computing are needed to further understand the world.

Wang Yang said:

"According to the theory of evolution, this is the best brain for understanding the world. Why don't we make a computer like this to help us understand the world? It does have feelings and can sense the changes in the world. It can also learn the language we have summarized. The future is very exciting. The future is a world where humans and machines merge."

Artificial intelligence is not evil

There have been many movies about artificial intelligence that show its devastating negative consequences, such as the "Terminator" series and the uncontrolled Ultron robot in "Avengers 2" to be released next year. In response, Tesla CEO Musk said that studying artificial intelligence is like summoning a demon, which will bring devastating consequences. IBM, which is at the forefront of artificial intelligence research, naturally has something to say. Wang Yang said:

"It's like a religious debate. When technology advances by leaps and bounds, it always makes people feel scared. The difference between humans and animals is that they can use tools. Can you run faster than a leopard? Yes, just drive a car. Can you catch up with a bird? Yes, our planes can catch it. Can you dive into the sea? Submarines can. All these things are tools, and humans can use tools in all aspects. The tool itself depends on who has it and what they do with it."

These are like gunpowder. Bullets can kill people, but explosives can also be used to cut through mountains, chisel rocks, dig tunnels and lay railways.

Moreover, from the perspective of justice, artificial intelligence must be studied. As mentioned above, Wang Yang believes that the development of technology is unstoppable. If righteous people do not do it, but evil people do, they will be more advanced than you, so humans must master this technology as soon as possible.

Although there are always some debates, just like the debate over stem cells, Wang Yang believes that scientific progress will always bring a better future for mankind.
IBM and Apple and other companies' cooperation

Although IBM and Apple had some conflicts in the past, their recent cooperation in the enterprise market has attracted much attention. Wang Yang said that IBM is now open-minded and wants to cooperate with the world's best technology companies, including Tencent, which just recently entered into a strategic cooperation.

How do IBM and Apple join forces? IBM is obviously very strong in the enterprise sector, while Apple is very strong in personal consumption and its terminals are very strong. IBM does not make terminals. During its transformation, IBM gave up on this aspect. IBM is relatively good at the other aspects, which are artificial intelligence, big data analysis, cloud computing, high-end servers, software services, etc. mentioned above. Enterprise customers obviously not only want large systems, but also want terminals, and they also need personal consumption to connect with each other, so this strong cooperation is very important.

For enterprises, how to put their applications on mobile devices is a problem, and IBM wants to provide solutions in this regard. Through mobile technology and social technology, enterprises can operate more efficiently. Currently, IBM and companies such as Apple have cooperated on 100 mobile applications for enterprises, and nearly 10 such applications have been released.

The interface between computers and people has changed, and mobile devices bring more data. Wang Yang gave an example, saying that mobile devices can take pictures and record voice. For example, in the insurance industry, if a car crashes, you can immediately take a photo and upload it. In the banking industry, why do you still need to deposit checks? You just need to take a photo and upload it. Don't you just want to prove a person's identity? Why do you have to use a piece of plastic to do this?

For example, in the cooperation with Tencent, Tencent's strength lies in social media, while IBM's strength lies in corporate users. The cooperation between the two can bring more promotion to enterprises. Twitter can understand what we are thinking in real time. When Twitter generates such a large amount of data, it is equivalent to what the human mind is thinking at this moment. It is distributed. Such data can be captured by text analysis. For corporate customers, they need to make decisions in real time. At this time, everyone is discussing something, or using such data, they can make correct business judgments.

Compared to the potential of Watson's artificial intelligence in enterprises, Apple's Siri has been used for many years. As for whether the cooperation between IBM and Apple can be extended to Watson and Siri in the future, Wang Yang said:

"I think it is entirely possible. The ultimate goal is to bring the strong together. If we believe that such cooperation will bring progress to society and humanity, we will do it. When others think it is meaningful, this is the market."