AI-Enabled Private and Public Data Integration Platform for Rapid Development of RegTech Applications

Professor Lam, Tak Wah   (Project Coordinator (PC))

Professor Cheung David Wai Lok   (Co-Investigator)

Professor Kao Chi Ming   (Co-Investigator)

30

2019-01-01

2021-06-30

4499950

AI-Enabled Private and Public Data Integration Platform for Rapid Development of RegTech Applications

AI-Enabled Private, Public Data Integration Platform, Rapid Development, RegTech Applications

Others - Computing Science and Information Technology

Engineering (E)

ITS/156/18FP

Innovation and Technology Support Programme (Tier 2)

2018

Completed

We propose to develop an intelligent data integration platform, codenamed RegBase, to streamline development of regulatory technology (RegTech) applications with artificial intelligence (AI). As one of the fastest growing financial technologies (FinTech), RegTech commonly leverages AI to detect financial malpractices and ensure regulatory compliance. To-date, different RegTech applications are developed from scratch separately with different primitive tools and unpolished datasets. These unconcerted development efforts incur repetitive data preprocessing and learning work and hinder AI data reuse across applications These issues would likely result in long application time-to-market as well as inconsistent data analysis. RegBase will support financial institutions to unify RegTech application development through automatic and continuous exploration of internal databases and public data sources to harvest regulatory knowledge with machine learning. Compared to developing RegTech applications from scratch, RegBase will provide reusable preprocessed data, machine training models, AI engines, and discovered knowledge for application development, cutting significant programming efforts and time. For example, business entities (e.g., listed companies, shareholders) and their relationships (e.g., stock holdings) can be mined from unstructured contents of annual reports and financial news, and reused across different applications. In addition, with the continuous feedback of training data from the applications, RegBase can automatically improve its AI accuracy over use, while providing users the total control over the data and knowledge. We will collaborate with the Securities and Futures Commission (SFC) to develop pilot projects to apply RegBase in daily financial regulatory activities to prove its commercial applicability. We will also solicit domain expertise from other industry partners to specialize the AI engines, training models, and other meta-data for the RegTech development in Hong Kong (e.g., an Hong Kong finance corpus) so as to facilitate future commercialization of RegBase.