Press Release
January 28, 2025
Aiming for improved user experience through more appropriate risk assessment, with potential to extend to other financial services.
GarageBank Corporation and Sophia University (Location: Chiyoda-ku, Tokyo; President: Yoshiaki Terumichi) have entered into a contract research agreement and have begun joint development of a new risk-evaluation model that leverages generative AI and unstructured data. By improving risk assessment, we aim both to enhance user convenience and to establish a method for detecting risks in advance, with the goal of extending the model to other financial-related services.
GarageBank, through cashari, provides a leaseback service centered on movable assets owned by individuals. Stable service operation requires sound risk control to support customer adoption. While we have already developed and operated machine-learning-based risk-evaluation models, we have started developing a new model that leverages generative AI and unstructured data to further strengthen our credit-evaluation capabilities.
The Momose Lab in the Applied Data Science Graduate Program at Sophia University, which is conducting the contract research, pursues new business outcomes and the commercialization of solutions through data science and has a track record of industry-academia collaboration in the application of generative AI and the development of new frameworks.
This contract research aims to improve credit-evaluation accuracy by using generative AI and unstructured data to analyze phenomena that conventional machine-learning approaches cannot fully capture. Specifically, we will analyze users' behavioral characteristics and external factors in addition to quantitative data such as age, gender, and occupation, integrating them to deliver improved credit-evaluation performance over our existing models.
By applying the resulting risk-evaluation model in the cashari service, we will improve user convenience through more appropriate risk assessment. Additionally, by establishing methods for detecting risks in advance through algorithm applications, we aim to extend the model to other financial-related services.
- Location: 7-1 Kioi-cho, Chiyoda-ku, Tokyo
- President: Yoshiaki Terumichi
- URL: https://www.sophia.ac.jp/eng/