One Explanation Does Not Fit All: A Toolkit and Taxonomy of AI Explainability Techniques

Abstract

As artificial intelligence and machine learning algorithms make further inroads into society, calls are increasing from multiple stakeholders for these algorithms to explain their outputs. At the same time, these stakeholders, whether they be affected citizens, government regulators, domain experts, or system developers, present different requirements for explanations. Toward addressing these needs, we introduce AI Explainability 360 (http://aix360.mybluemix.net/), an open-source software toolkit featuring eight diverse and state-of-the-art explainability methods and two evaluation metrics. Equally important, we provide a taxonomy to help entities requiring explanations to navigate the space of explanation methods, not only those in the toolkit but also in the broader literature on explainability. For data scientists and other users of the toolkit, we have implemented an extensible software architecture that organizes methods according to their place in the AI modeling pipeline. We also discuss enhancements to bring research innovations closer to consumers of explanations, ranging from simplified, more accessible versions of algorithms, to tutorials and an interactive web demo to introduce AI explainability to different audiences and application domains. Together, our toolkit and taxonomy can help identify gaps where more explainability methods are needed and provide a platform to incorporate them as they are developed.

References in corpus (8)

• Distilling the Knowledge in a Neural Network
• Towards A Rigorous Science of Interpretable Machine Learning
• Understanding Black-box Predictions via Influence Functions
• Towards Robust Interpretability with Self-Explaining Neural Networks
• Multifaceted Feature Visualization: Uncovering the Different Types of Features Learned By Each Neuron in Deep Neural Networks
• Boolean Decision Rules via Column Generation
• Improving Simple Models with Confidence Profiles
• Generalized Linear Rule Models

Cited by in corpus (50)

• What Do We Want From Explainable Artificial Intelligence (XAI)? -- A Stakeholder Perspective on XAI and a Conceptual Model Guiding Interdisciplinary XAI Research
• Explainable Artificial Intelligence (XAI) 2.0: A Manifesto of Open Challenges and Interdisciplinary Research Directions
• Expanding Explainability: Towards Social Transparency in AI systems
• Explainable Artificial Intelligence Applications in Cyber Security: State-of-the-Art in Research
• The Metacognitive Demands and Opportunities of Generative AI
• A Survey of the State of Explainable AI for Natural Language Processing
• To trust or not to trust an explanation: using LEAF to evaluate local linear XAI methods
• Counterfactual Explanations for Machine Learning on Multivariate Time Series Data
• Charting the Sociotechnical Gap in Explainable AI: A Framework to Address the Gap in XAI
• On the Explanation of Machine Learning Predictions in Clinical Gait Analysis
• Rule Extraction in Unsupervised Anomaly Detection for Model Explainability: Application to OneClass SVM
• XAIR: A Framework of Explainable AI in Augmented Reality
• Responsible and Regulatory Conform Machine Learning for Medicine: A Survey of Challenges and Solutions
• Explainable Machine Learning for Public Policy: Use Cases, Gaps, and Research Directions
• In AI We Trust? Factors That Influence Trustworthiness of AI-infused Decision-Making Processes
• The Grammar of Interactive Explanatory Model Analysis
• Machine Learning Explainability for External Stakeholders
• Informing clinical assessment by contextualizing post-hoc explanations of risk prediction models in type-2 diabetes
• How Much Automation Does a Data Scientist Want?
• Can Explainable AI Explain Unfairness? A Framework for Evaluating Explainable AI
• Question-Driven Design Process for Explainable AI User Experiences
• Model Agnostic Multilevel Explanations
• Explainable Active Learning (XAL): An Empirical Study of How Local Explanations Impact Annotator Experience
• Why Should I Choose You? AutoXAI: A Framework for Selecting and Tuning eXplainable AI Solutions
• Synthetic Benchmarks for Scientific Research in Explainable Machine Learning
• Transparency, Auditability and eXplainability of Machine Learning Models in Credit Scoring
• Calibrating Healthcare AI: Towards Reliable and Interpretable Deep Predictive Models
• Foundations of Explainable Knowledge-Enabled Systems
• Understanding Person Identification through Gait
• Machine Learning in Biomechanics: Key Applications and Limitations in Walking, Running, and Sports Movements
• Interpretable Machine Learning: Moving From Mythos to Diagnostics
• An Open API Architecture to Discover the Trustworthy Explanation of Cloud AI Services
• Learning Global Transparent Models Consistent with Local Contrastive Explanations
• Interpreting Deep Neural Networks Through Variable Importance
• TextFlint: Unified Multilingual Robustness Evaluation Toolkit for Natural Language Processing
• Thoughtful Adoption of NLP for Civic Participation: Understanding Differences Among Policymakers
• Interpretable Machine Learning for COVID-19: An Empirical Study on Severity Prediction Task
• Resisting Out-of-Distribution Data Problem in Perturbation of XAI
• Advancing Attribution-Based Neural Network Explainability through Relative Absolute Magnitude Layer-Wise Relevance Propagation and Multi-Component Evaluation
• SurvLIME-Inf: A simplified modification of SurvLIME for explanation of machine learning survival models
• On Explainability in AI-Solutions: A Cross-Domain Survey
• Stakeholder-in-the-Loop Fair Decisions: A Framework to Design Decision Support Systems in Public and Private Organizations
• What Would You Ask the Machine Learning Model? Identification of User Needs for Model Explanations Based on Human-Model Conversations
• Controlling Neural Networks with Rule Representations
• Explaining classifiers to understand coarse-grained models
• Pilot Investigation for a Comprehensive Taxonomy of Autonomous Entities
• Deep Learning Based Decision Support for Medicine -- A Case Study on Skin Cancer Diagnosis
• Explanation Ontology in Action: A Clinical Use-Case
• An Explainable Probabilistic Classifier for Categorical Data Inspired to Quantum Physics
• Do's and Don'ts for Human and Digital Worker Integration