Explainable AI in Fintech Risk Management

Abstract

The paper proposes an explainable AI model that can be used in fintech risk management and, in particular, in measuring the risks that arise when credit is borrowed employing peer to peer lending platforms. The model employs Shapley values, so that AI predictions are interpreted according to the underlying explanatory variables. The empirical analysis of 15,000 small and medium companies asking for peer to peer lending credit reveals that both risky and not risky borrowers can be grouped according to a set of similar financial characteristics, which can be employed to explain and understand their credit score and, therefore, to predict their future behavior.

Shapley-Lorenz eXplainable Artificial Intelligence

“…This idea has recently led some scholars to promote XAI methods aimed at making both the financial technology risk measurement models interpretable and transparent, and the risks of financial innovations, enabled by the application of AI, sustainable (see, e.g. Bracke, Datta, Jung, & Shayak, 2019;Bussmann et al, 2020). In particular, in Bussmann et al (2020) an explainable AI model based on similarity networks (Mantegna & Stanley, 1999) and Shapley values is proposed to measure the credit risks associated to the use of AI based credit scoring platforms.…”

“…Bracke, Datta, Jung, & Shayak, 2019;Bussmann et al, 2020). In particular, in Bussmann et al (2020) an explainable AI model based on similarity networks (Mantegna & Stanley, 1999) and Shapley values is proposed to measure the credit risks associated to the use of AI based credit scoring platforms.…”

“…The Shapley value approach appears as more intuitive than further typically used AI methods, as shown by the use case about explainability of risk management models described and developed by Bussmann et al (2020) within the European FINTECH Project (https://www.fintech-h o2020.eu). The use case was indeed presented, in a "human-centric" study, to the regulators of most European countries, and one of the main feedback was that the approach is nice and promising but should be compared to what obtained with "classical" model assessment methods, which is what the Shapley-Lorenz approach provides.…”

“…As discussed by Bussmann, Giudici, Marinelli, and Papenbrock (2020), the Shapley value based approach can be developed by using the SHAP (SHapley Additive exPlanations) computational framework (see, e.g. Lundberg & Lee, 2017).…”

Failure Prognostic of Turbofan Engines with Uncertainty Quantification and Explainable AI (XIA)

“…Moreover, this obscurity could result in ethical issues where black box applications risk offending race or gender of the users. To date, the only legal directive affecting the use of AI is the European General Data Protection Regulation (GDPR) whose interpretation regarding the obligation of logical explanation in automated processing, is currently being debated between AI experts and practitioners (Hacker et al, 2020;Chazette& Schneider, 2020;Bussmann et al, 2020). In term of ethical guideline, the European Commission's High-Level Expert Group on AI presented the key requirements in Ethics Guidelines for Trustworthy Artificial Intelligence in 2019, whose key requirements correspond directly or indirectly on the use of XAI (Bussmann et al, 2020).…”