“The Future Of Federated Learning On Alternative Credit Scoring” Webinar
20/04/2021
EVENTS
MSMEs make a crucial contribution to Hong Kong and its economic success, accounting for more than 98% of the business establishments and employing about 46% of the workforce in the private sector. Yet many SMEs face significant challenges in securing a bank loan unless they pay high interest rates or offer collateral to banks.
HKMA has engaged ASTRI to explore the use of fintech to support alternative credit scoring. This work capitalizes on the advantages of AI and machine learning to explore a wide variety of data from many sources to assess the creditworthiness of SMEs, including cash flow, point-of-sale transaction records, utility bill payments and even information from online accounting software programs.
The potential benefits of sharing enterprises’ data to support alternative credit scoring that rely on machine learning are tremendous. However, the issue of data privacy is restricting enterprises to share their data. Federated Learning is seen as the last mile of machine learning because the recent advancements in Federated Learning allow the consolidation of the outputs of machine learning models running on individual enterprises without violating the requirements of data privacy.
MSMEs make a crucial contribution to Hong Kong and its economic success, accounting for more than 98% of the business establishments and employing about 46% of the workforce in the private sector. Yet many SMEs face significant challenges in securing a bank loan unless they pay high interest rates or offer collateral to banks.
HKMA has engaged ASTRI to explore the use of fintech to support alternative credit scoring. This work capitalizes on the advantages of AI and machine learning to explore a wide variety of data from many sources to assess the creditworthiness of SMEs, including cash flow, point-of-sale transaction records, utility bill payments and even information from online accounting software programs.
The potential benefits of sharing enterprises’ data to support alternative credit scoring that rely on machine learning are tremendous. However, the issue of data privacy is restricting enterprises to share their data. Federated Learning is seen as the last mile of machine learning because the recent advancements in Federated Learning allow the consolidation of the outputs of machine learning models running on individual enterprises without violating the requirements of data privacy.