＂The value of scientific research lies in benefiting the people＂ —— A record of Jiang Changjun, a newly elected academician of China Academy of Engineering and professor of Tongji University.

For more than 30 years, he devoted himself to the research of network financial security, overcame the major technical problems of instantaneous accurate identification of system risk prevention and control, presided over the establishment of China’s first network transaction risk prevention and control system, system and standard, and made pioneering contributions for China to become an international "leader" in this field, providing effective support for China’s Internet finance volume to jump to the first place in the world. The achievements serve the national anti-financial fraud, escort the financial, trade, e-commerce, banking, insurance and other industries, and ensure the national financial security, which has been highly praised by the national regulatory agencies.

He is Jiang Changjun, a newly elected academician of China Academy of Engineering and professor of School of Electronics and Information Engineering of Tongji University.

China’s first network transaction risk prevention and control system

168.2 billion yuan, 213.5 billion yuan, 268.4 billion yuan, 498.2 billion yuan … This is the turnover of Tmall’s "Double Eleven" in the past four years. In the "Double Eleven" that just passed this year, Tmall’s turnover reached a new high of 540.3 billion yuan. It is proud that, as the representative of digital economy, the risk prevention and control technology developed by Jiang Changjun and his team is behind Alibaba’s vigorous development.

With the rise of e-commerce and the rapid development of Internet payment, how to prevent and control the risks involved? How can the electronic payment platform not only deal with the flood peak-like transactions, but also deal with the security problems of transactions, and instantly identify the "black hand" hiding behind the transactions and trying to steal other people’s accounts and passwords through Trojan horses?

In 2006, after encountering this technical problem, Alipay found Jiang Changjun, hoping that he could put forward new methods for anti-fraud in online transactions.

"Financial security concerns national security, and we are duty-bound." Jiang Changjun said. "The difficulty is that, on the one hand, the previous risk control is based on identity authentication, and it is difficult to identify identity theft and transaction fraud; On the other hand, transaction fraud shows a trend of high concealment and variability, and the risk control technology for specific scenarios (such as payment) is difficult to cope with all kinds of online transaction cross-industry portfolio fraud, lacking universality and systematicness. "

In the absence of successful experience at home and abroad, Jiang Changjun led the team to build the first risk prevention and control system of online transactions from scratch with "behavior authentication" as the core-based on behavior identification, it can accurately determine and instantly identify transaction fraud and stop fraudulent transactions in real time.

"In real life, everyone’s behavior habits will be reflected in words and deeds. The same is true on the Internet. Any small behavior will leave a’ trace’, and the equipment used, the duration and the frequency … will form a habit over time. " Jiang Changjun said, "This habit is as difficult to copy as fingerprints, and it is more complicated than passwords. Through in-depth analysis of these’ traces’, online transaction payment users can be instantly and accurately identified.". By collecting and analyzing the fragmented data left by users in the system, Jiang Changjun team constructed a representation model reflecting the user’s behavior texture. This representation model is different from the existing behavior portraits, and its emphasis is on representing the user’s internal behavior mechanism. Although fraudulent means are changeable, it is difficult to match the invariable behavior texture of legitimate users, so as to realize the authenticity identification of transactions.

The breakthrough of behavior authentication technology comes from basic research.

"To solve the security problem of online transactions and build a new real-time identification system to prevent identity theft and transaction fraud involves very basic theoretical problems." At the beginning, Jiang Changjun really thought about this tough nut to crack for some time.

"In fact, we have to solve not only the transaction fraud, but also the flood of online transactions." For example, in 2020, the order creation peak of Tmall’s "Double Eleven" reached 583,000/s, making it the largest traffic peak in the world. Jiang Changjun told reporters that the explosive growth of Internet application business, instant massive access easily leads to system paralysis, and such incidents have occurred in Amazon, Google, 12306, etc., resulting in huge economic and social impact.

"To solve the problem, it is urgent to create adaptive, real-time and efficient concurrent computing technology." Thus, Jiang Changjun thought of his doctoral thesis and postdoctoral report in Chinese Academy of Sciences. In his doctoral thesis, he established the PN machine behavior theory of network concurrent system, and later, during the postdoctoral research, he proposed a concurrent program verification method based on the PN machine behavior theory. Referring to his vision many years ago, he laughed and said, "Maybe this is my mission."

Having found the breakthrough and direction, Jiang Changjun devoted himself to tackling key scientific problems and constantly overcame technical difficulties.

Combined with the characteristics of online trading, he proposed the behavior authentication method of concurrent decoupling and risk control of trading system, and led the team to develop a large-scale distributed real-time prevention and control platform and special equipment for transaction payment risk, innovatively proposed the hierarchical diagnostic cloud technology system and risk intelligent analysis method for fingerprint verification of trading system, intelligent general inspection-special diagnosis-consultation of abnormal trading, changed special technology into common technology, and established a set of common method base for risk control. A number of risk control tools have been developed, and a cloud platform for online financial risk control has been established, which effectively solves the problems of consistent anti-fraud game and universality of risk control technology, makes China’s technology in this field reach the international leading level, and significantly enhances China’s international position in the Internet industry.

Rich original achievements have won unanimous recognition in the industry. The evaluation results of Asoke Nandi, an academician of the Royal Academy of Engineering, "put forward the original theory and technology of behavior authentication for risk prevention and control of online transactions". Professor R. Prasad, a fellow of IEEE (Institute of Electrical and Electronics Engineers), commented that "spatial correlation is used to study the optimal trade-off between aggregation throughput and aggregation efficiency". The achievement was shortlisted in the 2019 world internet conference Achievement Case Collection in Wuzhen, and successively won the second prize of the National Technology Invention Award, the second prize of the National Science and Technology Progress Award, the first prize of the Wu Wenjun Artificial Intelligence Technology Invention Award, and many international awards, such as HO PAN CHING YI Award, a representative award in the field of international discrete event dynamic systems.

Casting China Shield for Network Finance Security

Today, Alipay’s intelligent real-time risk control system has been upgraded to the fifth generation, which has helped Tmall to successfully survive many trading peaks such as "Double Eleven". The Alipay system has been tested and won praises from the industry. So far, Alipay has never experienced large-scale downtime and collapse, which is admired by the industry for a payment platform with an annual transaction volume of trillions.

Jiang Changjun showed such a set of data: PayPal, the largest online payment platform in the United States, has a transaction loss rate of 1.9 per thousand, while Alipay, which provides online payment services to more than 900 million real-name users in more than 200 countries and regions around the world, has reduced its transaction loss rate to 5 per million after applying their achievements, and the direct release rate of trusted transactions by the risk control system has increased to 96.91%, with an average response time of only 44.14 milliseconds.

Many teachers and students in the R&D team frankly said that their greatest sense of accomplishment is to see the skyrocketing transaction data, but the payment platform is still running smoothly and safely, and the technology they participated in can stand the test and provide convenience for consumers.

"The value of scientific research lies in benefiting the people." Jiang Changjun has always advocated and adhered to the pursuit of scientific research. In his view, it is not only the paper that can best prove the value of his research, but more fields can directly apply relevant results.

Indeed, it is not only Alipay and the online shopping army that benefit.

Jiang Changjun introduced that up to now, the achievements have been successfully applied in more than 500 enterprises including China Industrial and Commercial Bank, Sunshine Insurance, Shanghai International Pilot Free Trade Zone and Shanghai National E-commerce Comprehensive Innovation Practice Zone.

The application field is also expanding. In recent years, the achievements have served the national anti-financial fraud, successfully applied in the national anti-fraud center, effectively supported the "Yun Jian 2020" action of the Ministry of Public Security, and provided a strong guarantee for the healthy development of China’s Internet finance, which was fully affirmed by the national regulatory authorities. The achievements have also been successfully applied in transportation, security and other fields in 20 provinces, and played an important role in major events such as Shanghai World Expo and Beijing Olympic Games. The achievements also serve cross-domain aerial reconnaissance, surveillance, tracking, etc., realizing the dynamic perception and behavior monitoring of moving targets, ensuring the airspace safety, and playing an important role in the work of national day celebration cooperation and stability monitoring.

At present, he is leading the Shanghai Network Finance Security Collaborative Innovation Center to strengthen in-depth cooperative research with China People’s Bank digital currency Research Institute, Bank of China, non-bank payment institution online payment and clearing platform (Netlink) and other related financial institutions and enterprises, so as to provide security guarantee for the healthy and orderly promotion of digital RMB, thus laying an important foundation for the development of digital economy and constantly making it stronger and better. (Wei Dan)