As machine learning becomes mainstream, the need to explain these models better to trust them enough to be implemented has increased. This field of model explainability is often referred as “explainable artificial intelligence” (XAI). Advances in XAI can open up new opportunities in anti-money laundering (AML) that may have previously been disregarded due to regulatory concerns.
Our new technical white paper - Explainable Artificial Intelligence for Anti-Money Laundering - describes how you can use today's XAI techniques for a transaction monitoring alert generation or alert prioritization/scoring/hibernation model for AML. Some techniques reviewed in this paper include: global feature importance, partial dependence (PD) plots, locally interpretable model-agnostic explanations (LIME), and SHapley Additive exPlanations (SHAP).
Read now.
