Understanding counterfactual generation using maximum mean discrepancy
A quantitative approach using Maximum Mean Discrepancy (MMD) to evaluate and compare counterfactual explanation methods.
With the dramatic development of deep learning in the past decade, interpretability has been one of the most important challenges that often prevents neural networks from being applied to fields such as finance. Among many existing explainable analyses, counterfactual generation has become widely used for understanding neural networks and making tailored recommendations. However, few studies are devoted to providing quantitative measures for evaluating counterfactuals. In this paper, we propose a quantitative approach based on maximum mean discrepancy (MMD). We employ several existing counterfactual methods to demonstrate this proposed method on the MNIST image data set and two tabular financial data sets, Lending Club (LCD) and Give Me Some Credit (GMC). The results demonstrate the potential usefulness as well as the simplicity of the proposed method.
Latest publications
Mixed membership recurrent neural networks for modeling customer purchases
A mixed membership RNN approach for modeling customer purchase behavior with varying time intervals between purchases.
ICAIFInterpretable deep classifier for counterfactual generation
Framework using a supervised VAE with Normalizing Flow to generate counterfactual explanations in deep learning models. (ICAIF)
ICAIFALARM: human-in-the-loop anomaly detection
An end-to-end framework for anomaly detection, explanation and management with human-in-the-loop processes for effective action.
ICAIF