Seven challenges for harmonizing explainability requirements

Abstract

Regulators have signalled an interest in adopting explainable AI(XAI) techniques to handle the diverse needs for model governance, operational servicing, and compliance in the financial services industry. In this short overview, we review the recent technical literature in XAI and argue that based on our current understanding of the field, the use of XAI techniques in practice necessitate a highly contextualized approach considering the specific needs of stakeholders for particular business applications.

arXiv comment5 pages; Spotlight paper at the ACM SIGKDD Workshop on Machine Learning in Finance 2021

References in corpus (28)

• A Unified Approach to Interpreting Model Predictions
• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations
• Explaining Explanations in AI
• Consistent Individualized Feature Attribution for Tree Ensembles
• Expanding Explainability: Towards Social Transparency in AI systems
• Actionable Recourse in Linear Classification
• Anti-Money Laundering in Bitcoin: Experimenting with Graph Convolutional Networks for Financial Forensics
• Problems with Shapley-value-based explanations as feature importance measures
• Interpretable to Whom? A Role-based Model for Analyzing Interpretable Machine Learning Systems
• Algorithmic Recourse: from Counterfactual Explanations to Interventions
• Fast and Complete: Enabling Complete Neural Network Verification with Rapid and Massively Parallel Incomplete Verifiers
• Machine Learning Explainability for External Stakeholders
• True to the Model or True to the Data?
• Fairwashing Explanations with Off-Manifold Detergent
• Bias in Data-driven AI Systems -- An Introductory Survey
• Fair lending needs explainable models for responsible recommendation
• Model extraction from counterfactual explanations
• Explainable $k$-Means and $k$-Medians Clustering
• On the computation of counterfactual explanations -- A survey
• Towards Explainable AI Planning as a Service
• Who's responsible? Jointly quantifying the contribution of the learning algorithm and training data
• Distal Explanations for Model-free Explainable Reinforcement Learning
• Order in the Court: Explainable AI Methods Prone to Disagreement
• Towards Self-Regulating AI: Challenges and Opportunities of AI Model Governance in Financial Services
• Deep Argumentative Explanations
• Mathematical decisions and non-causal elements of explainable AI
• Interpretability of Epidemiological Models : The Curse of Non-Identifiability