Exclusive report On May 28th, the 2019 Global Blockchain Technology Development Forum (GBF), sponsored by the Organizing Committee of the 2019 China International Big Data Industry Expo, undertaken by the China Academy of Information and Communications Technology, and co-organized by TelInfo and Odaily, was grandly opened in Guiyang .

Exclusive report On May 28th, the 2019 Global Blockchain Technology Development Forum (GBF), sponsored by the Organizing Committee of the 2019 China International Big Data Industry Expo, undertaken by the China Academy of Information and Communications Technology, and co-organized by TelInfo and Odaily, was grandly opened in Guiyang .

The theme of this forum is "Technology Innovation and Integration". Internet giants such as Baidu, Ant Financial, and JD.com, representatives of some blockchain companies registered with the Cyberspace Administration of China, and well-known universities at Peking University will discuss in depth the impact of blockchain technology on the digital society around the three levels of "construction, empowerment, and leadership". the value of.

Zhang Jiachen also introduced that the multi-party secure computing platform of Lighttree Technology can open up data islands of mutual distrust, and has gradually been implemented in banking, insurance, AI companies and other industries, such as Zhongyuan Bank and China Insurance Anti-Fraud Alliance.

secondary title

The following is the full text of Zhang Jiachen’s speech edited by Odaily:

Good afternoon, ladies and gentlemen.

Light Tree Technology is an innovative company established in 2017. I am very glad to have the opportunity to communicate with you today. We focus on a very new technology called "Blockchain-based Secure Multi-Party Computing Technology". This technology allows data owners to perform trusted computing without exposing the original data of the data, so as to achieve data availability. Invisible effects, so it can help leaders from all walks of life here to develop new business application scenarios.

First of all, let me introduce the principle of secure multi-party computing based on blockchain, as well as our early practice in this area. Our society is already in the moment of data explosion. From now to the next ten years, the data we produce will increase tenfold. IDC statistics show that only 1.5% of the data in today's world has ever been analyzed.

Everyone knows that data still exists in the form of data islands, so what is the reason for this? On the one hand, everyone understands that data-based multi-party computing or the value of data shared by multiple parties has huge potential.

But while understanding the value, whenever we need data, the owners of various data islands are often very worried. What if we take the data out to calculate information security and privacy leaks? Now there is trusted computing, also known as privacy computing or secure multi-party computing technology, which can be very well combined with blockchain to unlock a new way of data collaboration.

As I mentioned, we started engineering this technology in 2017. In fact, government leaders have also quickly noticed that this emerging technology can create great value. In March of this year, Fan Yifei, vice president of the central bank, clearly mentioned in this year’s financial technology plan after the two sessions that secure multi-party computing technology can play a very important role in data fusion, and it has become one of the five emerging technologies that the central bank is now vigorously promoting .

Guests who are more concerned about network security will also see that the Cyberspace Administration of China announced the first draft of national data security management in the early hours of this morning. Safe and secure technology to protect the orderly flow of data.

Having said so much, many people are curious, what does it mean that data is available but not visible? What is the principle of this black technology? In terms of technical implementation, in fact, secure multi-party computing or trusted computing technology generally has two major types of implementation methods, one is based entirely on cryptography, and is implemented entirely at the software level. The second is to build a software system based on an area called the trusted execution environment in the chip, each of which has its own advantages and disadvantages.

To put it simply, the route based on a pure software system will be more transparent and does not need to be applied to hardware, but its defect is that its versatility is relatively weak, and different methods must be adopted for different models used by various industries. In the field of machine learning, such methods are relatively far from usable. Then we mainly adopt a trusted computing method based on a trusted hardware environment. After 2015, chip manufacturers such as Intel, Huawei, and Nvidia proposed such a solution through SGX, and reserved this in the chip called trusted execution. area of ​​the environment.

Let's first understand what kind of area the trusted environment is. It has several characteristics:

First of all, the first model will be installed in this trusted environment, so it can only execute the models that you have installed, and will refuse to execute the models that have not been installed, so you don’t have to worry about the partner in the data collaboration saying that I want you The data is used to make a credit score. In fact, he uses customer data to do calculations that we don't know or are not allowed.

The second feature is that the original data is encrypted and entered into the trusted environment. After entering the trusted environment, it is decrypted for calculation. Even if you own this machine, you cannot see the data in the trusted environment. , when the calculation results are shared when they leave the trusted environment, both parties get the calculation results. During this process, the original data that we care about is not exposed. This is the realization of secure multi-party computing based on a trusted environment principle.

Then everyone has a question, saying that since this is mainly a chip technology, will I need to re-purchase a lot of machines? In fact, after 2015, many machines already have such an area, and we can apply it. However, if you want to directly develop and use this environment on the chip, there is currently a problem: if you use the native tools of the chip manufacturer, you need to rewrite all the models in EDL, which is very inconvenient and difficult. Therefore, our company has developed a set of software tools that can seamlessly render commonly used models in a trusted environment without rewriting the business model, which greatly improves efficiency. Therefore, secure multi-party computing technology can become a powerful and indispensable new module in the data center.

The topic of today's forum is "Blockchain". So how does the secure computing technology itself integrate with the blockchain? More and more people in the blockchain industry are beginning to realize that the smart economy based on blockchain needs to be integrated with other technologies in order to achieve a complete scenario. In the process of data collaboration, in addition to the fact that the data is available and invisible, a key point is to make a highly authentic witness to the process of data collaboration: the purpose of laws and regulations including GDPR is to require all parties to issue a true and complete statement about which Relevant evidence of when the data party participated in the operation. In addition, smart contracts can also perform authority management during the collaboration process.

Traditionally, many of the last batch of blockchain-based data collaboration projects were focused on data marketplaces. The secure multi-party computing technology based on the blockchain not only supports the data transaction itself, but also helps everyone exchange data usage rights and separate data usage rights from value.

As a general technology, blockchain-based secure multi-party computing technology has many potential application scenarios. At present, we have carried out practical applications in the fields of banking, insurance and AI, occupying a leading position in this segment.

In the field of banking, we help city commercial bank customers cooperate with real estate companies and Internet companies through secure computing technology, and select precise customer groups suitable for promoting high-end wealth management products among the millions of public wealth management users, which improves marketing efficiency and Avoid excessive interruption to customers. In the field of inclusive finance, we use secure multi-party computing technology and introduce new data partners to perform credit scoring on people in the sinking market who were previously unable to grant credit, so that they can enjoy corresponding credit services. We help members of the Anti-Fraud Alliance of the Insurance Association to provide risk warnings for large and long fraudulent insurance through the secure multi-party computing system without exposing their confidential customer information, and to pre-position risks in the underwriting process to avoid fraudulent transactions. 100 million accident insurance fraud and other risk events. In addition, we also help AI companies protect their models so that they can deploy locally without exposing their intellectual property rights on the models.

I am very glad to have such an opportunity to introduce the integration technology of secure multi-party computing and blockchain to everyone, and I am very much looking forward to discussing with you in this field and looking forward to more scenarios and practices. Internationally, this field has attracted more and more attention from giants such as Facebook, and they have entered the stage of prototype development. We have started to deploy research and development in this field since 2017, and there have been actual deployment cases. Therefore, we have also received support from many data owners, including the Ministry of Industry and Information Technology's Institute of Information Technology and China Chengxin Credit Investigation, etc. Currently, 2 billion data assets have been hosted on our network.