AI Explainability 360 - Demo

  • Data
  • Consumer
  • Explanation

A Bank Customer wants to understand:

Why was my application rejected?
What can I improve to increase the likelihood my application is accepted?

Providing Contrastive Explanations for Insight into Loan Application Outcomes

The Bank Customer wants to know how and why the decision was made to accept or reject their loan application. The explanation given will help them understand if they’ve been treated fairly, and also provide insight into what – if their application was rejected – they can improve in order to increase the likelihood it will be accepted in the future. To help provide that insight and suggest avenues for improvement, we will use the Contrastive Explanations Method (CEM) algorithm available in AI Explainability 360. This algorithm sits on top of an existing predictive model and helps detect both the features that a bank customer could improve (e.g., amount of time since last credit inquiry, average age of accounts), and also further detects the features that will increase the likelihood of approval and those that are within reach for the customer. See examples below.

Select a customer asking for explanations