Chaindigg CEO Ye Mao: Mining the flow of illegal funds from blockchain data analysis can detect and prevent risks in time | Blockchain POD Conference
袁辉腾
2018-09-07 09:53
本文约2332字,阅读全文需要约9分钟
Early warning at the source of risk and remediation after the risk has passed.

On September 5th, the POD conference hosted by Odaily and strategically co-organized by 36Kr Group was held in Beijing. In the security forum of the conference, Ye Mao, CEO of Chaindigg, a blockchain big data service provider, discussed the security issues of blockchain with the guests from the perspective of data. Ye Mao introduced that Chaindigg refers to some ideas of artificial intelligence, uses data to escort the blockchain, and briefly introduces the general context.

secondary title

The following is the transcript of Chaindigg CEO Ye Mao’s speech:

Good afternoon everyone, just now many guests talked about blockchain security, including from the perspective of wallets and contracts. We talk about blockchain security from the perspective of data.

First of all, blockchain technology has many similar jobs to artificial intelligence in terms of data. Our team has the genes of the postdoctoral artificial intelligence team of Peking University, so we refer to some practices in artificial intelligence when processing data, including information labeling, information filtering methods, etc. Apply it to the blockchain data, so that the data can be well mined and correlated. Therefore, parsing the blockchain data chain is the lowest-level work, and on this basis, the data is mined, cleaned, correlated and marked. With these annotation information, you can use this information for analysis, and at the same time, you can perform on-chain data statistics.

Currently, the above data services are mainly applied to the following three types of users. The first is the supervision department. At present, we have more contacts with the public security department; the second is institutional users, including some exchange users; the third is individual users.

As the popularity of the blockchain is getting higher and higher, there are now a large amount of illegal funds evading supervision through various channels, and the number of blockchain-related crimes is also increasing year by year. Only then does the country’s “one size fits all” for some blockchain-related projects appear. Condition. We are also connecting with regulatory agencies to help the country regulate this aspect.

There will be a lot of specific technical details, and I will mainly talk about the context inside. First of all, it is necessary to conduct mining analysis on the transaction mode on the chain. After mining and analysis, we can understand the flow of funds in it. By tracking these flows, we can sort out the entire network of capital flows, and use our core labeling capabilities to label some centralized institutions, which are connected to the physical world. At the same time, if some important core addresses are found in the capital flow network, they can be monitored, and when the assets in these core addresses change, users including regulatory authorities will be notified as soon as possible.

Regarding the above-mentioned transaction pattern recognition, we draw on artificial intelligence technology technically, including pattern classification, recognition, and data cleaning. As we all know, a large part of artificial intelligence is to deal with very detailed data annotation and data training work. The same is true for the blockchain. It is necessary to repeatedly clean the many small and dirty data on the chain before they can be labeled and learned.

With these annotation results, based on this chain transfer mode, we can know how the funds in the specified address are transferred to other addresses step by step. For example, after a wallet address or an actual address is given, through this model, it can be understood that the funds are transferred to a centralized exchange, and later may be transferred from the exchange to other institutions, and this thread can come out .

By merging the node information in different contexts, a large knowledge graph can be finally formed, that is, the interaction graph between addresses in the network. The graph can display all transactions on the blockchain, usually a graph of hundreds of millions of nodes. In addition, the process of processing will also involve how to effectively improve performance and how to effectively mine.

Based on the above technology and work results, we have established cooperation with some provincial and municipal public security departments.

This is a letter of thanks, proof materials, etc. sent by the public security and other regulatory authorities. One of our goals is to embrace regulation through our data analysis, so that regulators can use these tools to control the risks on the blockchain. When the public security and other government departments get involved, they will promote the compliance of blockchain technology and the stable development of the blockchain industry.

There are similar applications in foreign countries, such as CHAINALYSIS to assist the government in tracking work.

The same goes for individual users who are similar to it. Individuals can use this service to monitor whether the assets in personal addresses or wallet addresses have abnormal changes. If there is an abnormal change, you can use this tracking service to know which platform or institution the asset flows to. Therefore, users of such stolen digital assets can also use this service for security protection.

Finally, let me talk about institutional users. We have found that there have been a large number of stolen incidents in the entire digital assets in the past, and the amount of loss was very large. In what way can we better improve security and reduce losses? One of the methods is to build a high-risk address library and collect different high-risk addresses. Use our core labeling capabilities to expand and collect relevant addresses and expand the richness of the library. At present, our high-risk address library covers hacker addresses, stolen coins addresses, phishing addresses, and other addresses related to high-risk addresses that we have mined from the system. When an individual user has a relationship with other addresses, you can check whether the address in the library is related to this address. If there is a correlation, remind in advance of this risk that a transfer to an abnormal address may occur.

The same is true for institutions. Institutions can use this address library as an on-chain risk control system, conduct daily inspections of transaction scenarios, determine whether addresses in institutions interact with high-risk addresses, and generate reports regularly to evaluate whether It's risky. Therefore, using the high-risk address library and on-chain risk control can do KYC, allowing organizations to better understand users, reminding of exceptions, and regularly form reports.

Finally, an example is given. In response to the theft of Bithumb some time ago, we recovered the theft of Bithumb through on-chain address tracking and analysis, and made a report. After docking with the Bithumb exchange, it believed that more than 90% of the report was accurate, which fully complied with the theft process at that time. After tracking, it was found that these digital assets finally flowed to another exchange on August 2. With this information, we can better help him avoid risks and recover assets in this area.

To sum up, Chaindigg provides some data-based services to escort the blockchain. On the one hand, through data analysis, it can better supervise the application of blockchain for the government and other departments, so that the industry can develop better; on the other hand, through the results of data analysis, it can provide security solutions for institutions and individuals. Through the joint efforts of everyone, the blockchain can develop steadily in a normal and healthy direction. thank you all.

袁辉腾
作者文库