More and more people believe that blockchain can only achieve its greatest value in technology components. Whether it is big data, cloud computing, AI or blockchain, they are all maximizing the value of data, such as those that collect data, analyze data, and utilize data, while blockchain only provides credibility and consensus. , tamper-proof high-quality data, so the blockchain is only one of many technical components, and how to combine it with other technologies is a question that many people are currently discussing.

Forum guests:

Forum moderator:

BITAGORA CEO Song ting cai

Forum guests:

Development Director of Asia Union - Yi Huanhuan

Inspur Artificial Intelligence and High Performance Computing Director - Lu Fei,

Shuo Di from Tsinghua Arshan Blockchain Research Center,

Institute of Data Science, Tsinghua University-Liu Yunqu,

secondary title

Core question 1: What are the contradictions between artificial intelligence and blockchain technology? How to prevent and avoid conflicts?

Duan Kai, Technical Advisor of Sichuan Research and Development Center of Blockchain:When the blockchain serves AI, insufficient computing resources are the main contradiction.

Institute of Data Science, Tsinghua University - Yunqu Liu:AI is the processing of big data, and blockchain is to confirm the value of big data. Many people say that privacy is very important, but right confirmation and asset attributes are more important than privacy. We first know whose big data it is, and then process it through AI. I dwarfed AI as a tool for big data, and many people may not be happy. From my narrow perspective, this has been the case in the past. As a tool, it can bring out the value of big data, and under the protection of the blockchain, it can confirm rights and conduct transactions, thereby generating huge commercial value.

In my limited knowledge, there is no conflict between AI and blockchain, but they can complement each other very well, producing a drastic effect of 1+1 equals 11.

Shuo Di from Tsinghua Arshan Blockchain Research Center:AI is a tool to improve productivity, and blockchain is to improve production relations. How can the two be combined? We see that the AI ​​that everyone is talking about now mainly refers to simulated machine learning, mainly in terms of recognition, which is biased towards perception. The blockchain is a trust machine, and the data on it is valuable. Often represents a kind of business behavior. Including the value of the item and the ownership relationship of the asset. If you want to talk about AI, how does it relate to these data, I think the blockchain can help traditional AI obtain some data that it could not have obtained. For example, if the business behavior of enterprises is based on the blockchain, and the business data, financial data, and asset data of various enterprises can be put on it, we can create a new AI. What can this AI do? You can judge the value of this enterprise and whether it is a good investment target. For example, it looks for good investment targets in the blockchain field or certain fields. I think there is no contradiction between the two. These are two technologies, and they will be related only in certain application fields.

Inspur Artificial Intelligence and High Performance Computing Director - Lu Fei:From my personal point of view, consider solving this problem from these aspects.

First, consider computing power. From my personal point of view, how are you running and how are you developing? Can you objectively evaluate whether the current computing power is enough to support the entire system.

Second, data. In fact, big data has contributed to the development of AI, and I personally agree with it. In terms of data production, management, security, rule-making, and data legislation, these aspects are considered a little more.

Third, the optimization of AI algorithms. If there is no good algorithm to support it, it will also bring many problems.

Fourth, the accumulation of talent teams. Because this industry is developing too fast, the direction of talents is also very important. If we think more about these four aspects, it is possible to avoid solving the related contradictions in the process of combining AI and blockchain, and even combining with other applications.

Development Director of Asia Union - Yi Huanhuan:This question reminds me of a cross talk about Hou Baolin called "Qin Qiong Fighting Guan Gong". I think these two questions are not in the same category. To see what contradictions exist between them, I personally think that what contradictions exist in the blockchain itself, or what contradictions exist in artificial intelligence itself. How we will resolve these contradictions in the future will be of greater significance to everyone here.

First of all, the temple on the blockchain comes from the cover magazine of "The Economist", saying that the blockchain is a mechanism of trust. However, the current main contradiction of the blockchain is the impossible triangle of decentralization, performance and security.

The current contradiction of artificial intelligence is that it can only solve the problem of weak artificial intelligence. Like face recognition, image recognition, etc. It is also unable to solve industry-specific problems and cannot generate huge profits. Its contradiction is actually the contradiction between huge valuation and weak profit.

secondary title

Core question 2: Can the combination of blockchain and artificial intelligence find some commercial applications?

Duan Kai, Technical Advisor of Sichuan Research and Development Center of Blockchain:For artificial intelligence products, many people pay attention to the development of big data and machine learning. In fact, artificial intelligence products also need data during use. You will find a very special application mode, many common applications can be used in mobile phones. But if you use some artificial intelligence products, you will find that if you don't provide him with some information or data, this robot will be as indifferent as a fool. So when users face this kind of artificial intelligence product, I predict that in the future, you will put your own information in it first. So what's the way? How to ensure that your own information is not leaked, these things can be combined with the blockchain, and rely on it to solve these problems.

We have experienced the red envelope era of the Internet. Anyone who has experienced OFO or Mobike knows that red envelopes can open up the market. In the future, it will be reflected that for an individual, he does not need to do anything, and his own data can bring him value. This is the combination point we hope to see in the future.

Institute of Data Science, Tsinghua University - Yunqu Liu:When I just talked about the relationship between blockchain, AI and big data, I said that big data is the point of view of connecting blockchain and AI. In my opinion, it is a chain relationship and has not yet formed a triangular relationship. With big data, AI can process big data and discover the value of this big data. Equivalent to these values, through blockchain transactions, huge commercial value is generated. This is a visible joint, and its commercial value is enormous.

Technically, there is a challenge here, that is, the existing blockchain technology does not yet have the ability to handle large data sets.

Shuo Di from Tsinghua Arshan Blockchain Research Center:I quite agree that this is Guan Gong fighting Qin Qiong, what scene? It depends on the scene and application field. Blockchain allows data that could not be shared before to be shared. Let a lot of data that people were afraid of leaking before, and now put it on the chain, there is a way to safely hand it over to those who want to use it. Such openness of data brings about the new vertical direction of artificial intelligence. Now artificial intelligence can only do some recognition. I rarely see research on artificial intelligence in the economic field, business operation field, and other related things. The blockchain brings these data online and online, and perhaps these studies can be implemented.

In the future, if personal data is to be used in the future, each precise scenario must be confirmed by the owner of the data ownership. You can use it legally only if he expressly agrees to it. In the past, on a webpage or in an APP, users were generally asked to draw a tick, click to agree, and then use their data unscrupulously everywhere. This kind of thing will be illegal in the future, and the illegality may affect a lot of existing enterprise. Including some companies doing artificial intelligence research, it may be a blow to artificial intelligence.

Aershan Company and Tsinghua Research Center, the cornerstone chain system we built, there is exactly one module in it. It is for data management, data ownership, your data cannot be abused by others, and whoever owns the data is the master. This tool is called digital asset management or digital asset custody, and it can help you know, first of all, what data you have. Second, whoever wants to use your data, you will receive a message. For example, there is an APP on your mobile phone. When you go to a hospital to see a doctor, the APP tells you that the hospital wants to retrieve your previous health records. You can authorize him to use the prescriptions and diagnosis records prescribed by previous doctors through the APP. Under this legal framework, the data is opened up, and there should be opportunities to spawn some new artificial intelligence research fields.

Inspur Artificial Intelligence and High Performance Computing Director - Lu Fei:The integration of artificial intelligence and blockchain applications is what we expect. Artificial intelligence cannot solve many problems at the moment, such as quantitative trading. Why is artificial intelligence so popular now? Because it has already begun to get involved in all fields. The fields involved may not be imagined by us. Everyone is trying, which requires a process, and we need to face it rationally, or develop it selectively. So I personally think that this integration is our ultimate expectation. But it is necessary to face it rationally and make a rational choice.

Development Director of Asia Union - Yi Huanhuan:I understand this problem and its application in business scenarios. I would like to put it the other way around, and I would like to add to the host's question. What I just said is that weak artificial intelligence has no model yet, and the business model is not clear. However, the business model of the blockchain is very clear. It can be seen that the blockchain changes the production relationship because in the past, traditional companies depended on whether the boss held shares. In a joint-stock company, employees hold shares. When it comes to Internet companies, there is an AB share system, and the boss only accounts for 9%. It can trigger more people to participate, and the blockchain becomes a community. Become a community system, all for one, one for all. This is a huge change in its incentive mechanism, and this is the first one.

Second, I think that talking about blockchain and artificial intelligence is not about applying it to business, but because of the pain points and problems in the commercial field, which can be solved through blockchain and artificial intelligence at this time. In fact, it was because of the US government’s over-issued currency in 2008 that brought about a financial crisis, so Bitcoin appeared that did not trust the government.

When in the blockchain industry, there was only one payment system and no more problems were solved, smart contracts appeared and Ethereum was born. It is the same now, we found in the practice process that we use a smart contract system like Ethereum, which cannot solve the problem of massive data collection, and cannot solve the problem of point-to-point data confirmation and verification in the field of Internet of Things, so we need to do further work on it. Modifications, further repairs, but also stand on the original angle and basis. So I think, in essence, it is not how these technologies are applied, but what kind of pain points are found in the industry itself.

First, the pain points of data collection. Second, the pain point of granulation and fine dimensionality. Third, intelligence is very complex. In addition to meeting China's environmental protection standards, it must also meet the standards of overseas, the United States, and the European Union. Its calculation formulas are completely different in different scenarios. How do I automate the corresponding matching? After the matching is completed, there are more diverse organizations, how to identify, connect and compare with the existing system. This is in turn a further requirement for blockchain and artificial intelligence technology from realistic application scenarios, and then further iterations. And the development of the entire software industry is also like this. The company we used to serve was called Oracle. What problems did the Oracle database solve in those years? Solve the problem of US military data storage. Finally, this technology is extracted and becomes a general service. Including SAP's ERP, also starting from finance, a little bit of generalized plug-ins to further provide comprehensive services. So it is the same for blockchain and artificial intelligence. I think that in the future, in the bigger scene, the corresponding public chain of the refined blockchain and the smart contract system required by the blockchain will further form the link between the various chains. The corresponding infrastructure, I think this is the direction and core of future development.