Navigator of Financial Risk Control: Exploring ERNIE's Application in Du Xiaoman's User Risk Control

‍

As technology progresses, borrowing money online has become more accessible for consumers, highlighting the convenience that modern advancements bring. However, this rise in consumer finance has also led to challenges, particularly in the credit system, which has become a major barrier for consumer finance companies. The prevalence of credit fraud further complicates the situation, posing significant challenges to the growth of the financial industry and leaving many lending activities inadequately addressed.

The Severity of the Problem

According to the People's Bank of China's "Overall Operation of the Payment System in the Third Quarter of 2018," the total amount of credit card debt overdue for six months in China reached 88.098 billion yuan, a quarter-on-quarter increase of 16.43%.

This indicates an urgent need for risk control in the financial industry. Faced with a more diverse customer base and increasingly complex user information, there is a need to ensure business security and compliance while balancing risk control measures and customer experience.

‍

Traditional Approaches to Financial Risk Control

In traditional financial institutions, financial risk controllers and auditors are employed to manually review lending qualifications. This job demands high standards from practitioners, requiring relevant background knowledge to fully understand customers' credit conditions and necessitating meticulousness and independent judgment capabilities. With the development of internet finance, thousands of lending activities occur daily on platforms, leading to massive human resource consumption and challenges in maintaining consistent and efficient review standards.
‍

Moreover, traditional risk control modeling techniques are based on small-sample supervised learning, heavily relying on feature extraction, consuming significant manpower and depending on individual experience. Small-sample text data processing often lacks context understanding, leading to misinterpretations of user information.

‍

The Solution: Baidu's ERNIE in Du Xiaoman's User Risk Control

Du Xiaoman currently offers services like education loans and educational cash loans, aiming to "make every dream comtrue with money," helping users further their education or learn new professional skills. This partnership with Baidu aims to apply Baidu's self-developed continuous learning semantic understanding framework ERNIE to Du Xiaoman's user risk control scenarios.

‍

‍