ASTRI Leverages Privacy-preserving “federated Learning” Technology to Facilitate Credit Scoring for MSME Financing in Collaboration With Standard Chartered, PAOB, Openrice and Freightamigo

  • Artificial Intelligence , Financial , Technical
  • 19.10.2021 01:00 pm

The Hong Kong Applied Science and Technology Research Institute (ASTRI) joins forces with tech-embracing companies to leverage a privacy-preserving technology, called “Federated Learning”, to develop artificial intelligence (AI) models and output in the form of encrypted parameters that serve as a reference for financial institutions to conduct comprehensive credit analyses for micro, small and medium-sized enterprises (MSMEs) to help them get access to financing. ASTRI’s partners are Standard Chartered Bank (Hong Kong) Limited; Ping An OneConnect Bank (Hong Kong) Limited (Ping An OneConnect Bank or PAOB), the first virtual bank serving MSMEs in Hong Kong; OpenRice, a Hong Kong restaurant guide and review platform; and FrieghtAmigo, a logistics and freight pricing platform. 

Unlike traditional machine-learning methods, Federated Learning does not require data to be transferred directly to a central database, thus protecting privacy and mitigating the risk of data security breaches. Data partners and financial institutions can establish common credit evaluation models by combining their encrypted parameters. During the process, the collaborators do not have access to any consumer personal data, nor are the identities of the enterprises identified. Only when an enterprise applies for financing and is undergoing authorisation can the designated financial institution obtain the relevant parameters and conduct a credit evaluation. 

Dr Denis Yip, Chief Executive Officer of ASTRI, said: “ASTRI develops and leverages innovative technologies. We leverage the Federated Learning technology to provide alternative data for credit assessment, while protecting privacy and data security to help financial institutions reduce the cost of vetting and approving loans for MSMEs and help enterprises get financing. We look forward to collaborating with more organisations to promote the implementation of open data. The Federated Learning technology will also effectively promote the development of other Fintech applications and support the government’s efforts to drive smart city transformation.”

Ms Winnie Tung, Head of Business Banking, Standard Chartered Bank (Hong Kong) Limited, said, “Standard Chartered has always been actively driving innovations and digitisation to enhance our capability to serve our clients. The Federated Learning technology enables us to evaluate the financing needs and the health of SMEs from a new perspective, while ensuring our clients’ data privacy is well protected. Our business clients no longer need to go through the time-consuming process of providing many financial reports to get banking facilities. The streamlined process makes financing solutions more accessible to our clients. We will continue our efforts to make innovation and client-centricity part of our business ethos.”

Ms Mabel Chu, Deputy Chief Executive of PAOB, said, “As a pioneer in the virtual banking sector servicing SMEs, PAOB is the first virtual bank in town adopting alternative credit scoring as the foundation for credit assessment, and leveraging financial technology and strategic partnerships with various business partners to close the service gaps in SME banking. We look forward to our strategic collaboration with ASTRI in this Federal Learning exercise. With the AI models, we can access our customers’ comprehensive operational data in real time to help realise financial inclusion by expediting loan approvals and meeting the financing needs of SMEs.”

Mr Joe Yau, CEO and Acting CTO of OpenRice Limited, said, “OpenRice is the most popular food dining platform in Hong Kong, accumulating 10 trillion data units every year, including restaurant distribution and categories, business performance, and social media data sharing. With the strong level of personal data protection, we expect to use Big Data and Federated Learning to train AI models and develop professional industry parameters to make credit applications from small and medium-sized restaurant partners faster and easier to quickly meet their operational needs.”

Ms Ivy Tse, Co-founder and Director of FreightAmigo, said, "As a pioneer of one-stop supply-chain finance eMarketPlace in Asia, FreightAmigo integrates complex logistics industry data in real time through innovative technologies to simplify operations in the supply chain industry. FreightAmigo is collaborating with ASTRI, using the Federated Learning technology to strictly protect privacy. The massive data on the platform is used to train AI models to help to build industry parameters, which helps SMEs apply for trade financing and credit with ease, and effectively addresses their financing needs. It coincides with the mission of FreightAmigo — ‘Dedicating to the advancement of the logistics and financial industry with technologies’. Our ultimate goal is to "make trade easier".

As data partners, OpenRice and FreightAmigo will leverage big data, including various restaurant popularity metrics, the transaction status of consignment merchants, and business operation status, to identify the elements affecting the credit risks of enterprises from alternative data by using AI and Federated Learning. Through this, a model will be trained to derive parameters to assist credit scoring. No information about the enterprises will be transferred from the data partners to other institutions.

With the authorisation of the enterprises applying for loans, Standard Chartered and PAOB can refer to data providers’ assessments of an enterprise’s competitiveness in its industry and its credits status, which is determined using the enterprise’s operation parameters through the model developed to process an MSME loan application. 

During the loan-approval process, financial institutions will be able to make more reliable credit assessments based on the projections of their own credit evaluation models and the assessments of their data partners. The first phase of the models developed using Federated Learning is expected to be in use within 12 months.

At the end of 2020, commissioned by Hong Kong Monetary Authority (HKMA), ASTRI published a white paper titled “Alternative Credit Scoring of Micro-, Small and Medium-sized Enterprises”, which outlines how FinTech can be adopted to utilise alternative data to evaluate borrowers’ creditworthiness. ASTRI collaborated with financial institutions and data partners to implement the white paper and adopt Federated Learning to protect privacy.

The Office of the Privacy Commissioner for Personal Data issued “Guidance on the Ethical Development and Use of Artificial Intelligence” a month ago. It also recommends that organisations adopt Federated Learning to train AI models to avoid unnecessary sharing of training data from different sources.

“The Protecting Privacy in Practice” report, published by the Royal Society in 2019, points out that Federated Learning is an emerging practice for training machine-learning models to protect privacy. 

 

Related News