A Survey on the Explainability of Supervised Machine Learning

Published in

Journal of Artificial Intelligence Research(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

This survey paper provides essential definitions, an overview of the different principles and methodologies of explainable Supervised Machine Learning, and a state-of-the-art survey that reviews past and recent explainable SML approaches and classifies them according to the introduced definitions.

Abstract

Predictions obtained by, e.g., artificial neural networks have a high accuracy but humans often perceive the models as black boxes. Insights about the decision making are mostly opaque for humans. Particularly understanding the decision making in highly sensitive areas such as healthcare or finance, is of paramount importance. The decision-making behind the black boxes requires it to be more transparent, accountable, and understandable for humans. This survey paper provides essential definitions, an overview of the different principles and methodologies of explainable Supervised Machine Learning (SML). We conduct a state-of-the-art survey that reviews past and recent explainable SML approaches and classifies them according to the introduced definitions. Finally, we illustrate principles by means of an explanatory case study and discuss important future directions.

Authors

Nadia Burkart

1 papers

Marco F. Huber

2 papers

References272 items

Explainable Artificial Intelligence (XAI)

An Interpretable Machine Learning Framework for Accurate Severe vs Non-severe COVID-19 Clinical Type Classification

Triaging moderate COVID-19 and other viral pneumonias from routine blood tests

An Investigation of COVID-19 Spreading Factors with Explainable AI Techniques

DeepCOVIDExplainer: Explainable COVID-19 Predictions Based on Chest X-ray Images

A Survey on the Explainability of Supervised Machine Learning

Published in

Journal of Artificial Intelligence Research(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

Abstract

Authors

Nadia Burkart

1 papers

Marco F. Huber

2 papers

References272 items

Explainable Artificial Intelligence (XAI)

An Interpretable Machine Learning Framework for Accurate Severe vs Non-severe COVID-19 Clinical Type Classification

Triaging moderate COVID-19 and other viral pneumonias from routine blood tests

An Investigation of COVID-19 Spreading Factors with Explainable AI Techniques

DeepCOVIDExplainer: Explainable COVID-19 Predictions Based on Chest X-ray Images

Doctor XAI: an ontology-based approach to black-box sequential data classification explanations

Preserving Causal Constraints in Counterfactual Explanations for Machine Learning Classifiers

Towards Explainable Deep Neural Networks (xDNN)

Forcing Interpretability for Deep Neural Networks through Rule-Based Regularization

FACE: Feasible and Actionable Counterfactual Explanations

KANDINSKY Patterns as IQ-Test for Machine Learning

Attention is not not Explanation

A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI

Interpretable Counterfactual Explanations Guided by Prototypes

An Ontology-based Approach to Explaining Artificial Neural Networks

A Study on Trust in Black Box Models and Post-hoc Explanations

Understanding the Effect of Accuracy on Trust in Machine Learning Models

Enhancing Decision Tree Based Interpretation of Deep Neural Networks through L1-Orthogonal Regularization

Attention is not Explanation

Faithful and Customizable Explanations of Black Box Models

Quantifying Interpretability and Trust in Machine Learning Systems

Human-centric Transfer Learning Explanation via Knowledge Graph [Extended Abstract]

Towards Aggregating Weighted Feature Attributions

Definitions, methods, and applications in interpretable machine learning

Efficient Search for Diverse Coherent Explanations

iForest: Interpreting Random Forests via Visual Analytics

Metrics for Explainable AI: Challenges and Prospects

Interactive machine learning: experimental evidence for the human in the algorithmic loop

Please Stop Explaining Black Box Models for High Stakes Decisions

Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)

Knowledge-based Transfer Learning Explanation

Model Agnostic Supervised Local Explanations

Towards Robust Interpretability with Self-Explaining Neural Networks

Defining Locality for Surrogates in Post-hoc Interpretablity

Deep Neural Decision Trees

ML-Schema: Exposing the Semantics of Machine Learning with Schemas and Ontologies

Explaining Explanations: An Overview of Interpretability of Machine Learning

Local Rule-Based Explanations of Black Box Decision Systems

Explainable artificial intelligence: A survey

Anchors: High-Precision Model-Agnostic Explanations

Explainable Recommendation: A Survey and New Perspectives

Manipulating and Measuring Model Interpretability

Global Model Interpretation Via Recursive Partitioning

A Survey of Methods for Explaining Black Box Models

How do Humans Understand Explanations from Machine Learning Systems? An Evaluation of the Human-Interpretability of Explanation

Model Class Reliance: Variable Importance Measures for any Machine Learning Model Class, from the "Rashomon" Perspective

Inverse Classification for Comparison-based Interpretability in Machine Learning

Distilling a Neural Network Into a Soft Decision Tree

The Promise and Peril of Human Evaluation for Model Interpretability

The Doctor Just Won't Accept That!

Rule Extraction from Decision Trees Ensembles: New Algorithms Based on Heuristic Search and Sparse Group Lasso Methods

Beyond Sparsity: Tree Regularization of Deep Models for Interpretability

Explaining Trained Neural Networks with Semantic Web Technologies: First Steps

Counterfactual Explanations Without Opening the Black Box: Automated Decisions and the GDPR

What Does Explainable AI Really Mean? A New Conceptualization of Perspectives

Explainable Artificial Intelligence: Understanding, Visualizing and Interpreting Deep Learning Models

Using Explainability for Constrained Matrix Factorization

Human-Centric Justification of Machine Learning Predictions

Optimized Risk Scores

A glass-box interactive machine learning approach for solving NP-hard problems with the human-in-the-loop

How AI detectives are cracking open the black box of deep learning

Efficient Data Representation by Selecting Prototypes with Importance Weights

ProtoDash: Fast Interpretable Prototype Selection

Interpretable & Explorable Approximations of Black Box Models

Methods for interpreting and understanding deep neural networks

Explanation in Artificial Intelligence: Insights from the Social Sciences

Interpretable Predictions of Tree-based Ensembles via Actionable Feature Tweaking

Interpretable Active Learning

Challenges for Transparency

Interpreting Classifiers through Attribute Interactions in Datasets

SmoothGrad: removing noise by adding noise

Rule Extraction from Training Data Using Neural Network

Interpreting Blackbox Models via Model Extraction

A Unified Approach to Interpreting Model Predictions

Explaining machine learning models in sales predictions

Understanding Black-box Predictions via Influence Functions

Towards A Rigorous Science of Interpretable Machine Learning

Neural Decision Trees

Context-Aware Captions from Context-Agnostic Supervision

Attentive Explanations: Justifying Decisions and Pointing to the Evidence

Bayesian Rule Sets for Interpretable Classification

Gradients of Counterfactuals

DeepRED - Rule Extraction from Deep Neural Networks

IBM Watson Analytics: Automating Visualization, Descriptive, and Predictive Statistics

Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization

A model explanation system

Learning sparse two-level boolean rules

Interpretable Decision Sets: A Joint Framework for Description and Prediction

European Union Regulations on Algorithmic Decision-Making and a "Right to Explanation"

EU regulations on algorithmic decision-making and a "right to explanation"

Enhancing Transparency and Control When Drawing Data-Driven Inferences About Individuals

Explainable Restricted Boltzmann Machines for Collaborative Filtering

Making Tree Ensembles Interpretable

Model-Agnostic Interpretability of Machine Learning

Modeling Missing Data in Clinical Time Series with RNNs

Rationalizing Neural Predictions

The mythos of model interpretability

A Budget-Constrained Inverse Classification Framework for Smooth Classifiers

Algorithmic Transparency via Quantitative Input Influence: Theory and Experiments with Learning Systems

100

Interacting with Predictions: Visual Inspection of Black-box Machine Learning Models

101

Not Just a Black Box: Learning Important Features Through Propagating Activation Differences

102

Natural Language Generation enhances human decision-making with uncertain information

103

Distribution-Free Predictive Inference for Regression

104

Generating Visual Explanations

105

Scalable Bayesian Rule Lists

106

Auditing black-box models for indirect influence

107

“Why Should I Trust You?”: Explaining the Predictions of Any Classifier

108

Mind the Gap: A Generative Approach to Interpretable Feature Selection and Extraction

109

Interpretable Two-level Boolean Rule Learning for Classification

110

A Neural Attention Model for Abstractive Sentence Summarization

111

Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model

112

Effective Approaches to Attention-based Neural Machine Translation

113

Trading Interpretability for Accuracy: Oblique Treed Sparse Additive Models

114

Optimal Action Extraction for Random Forests and Boosted Trees

115

Intelligible Models for HealthCare: Predicting Pneumonia Risk and Hospital 30-day Readmission

116

Ontological Random Forests for Image Classification

117

Understanding classifier errors by examining influential neighbors

118

Or's of And's for Interpretable Classification, with Application to Context-Aware Recommender Systems

119

Supersparse linear integer models for optimized medical scoring systems

120

Show, Attend and Tell: Neural Image Caption Generation with Visual Attention

121

The Bayesian Case Model: A Generative Approach for Case-Based Reasoning and Prototype Classification

122

Explaining prediction models and individual predictions with feature contributions

123

Falling Rule Lists

124

Neural Machine Translation by Jointly Learning to Align and Translate

125

A peek into the black box: exploring classifiers by randomization

126

MofN rule extraction from neural networks trained with augmented discretized input

127

Principal Component Analysis (PCA)

128

Methods and Models for Interpretable Linear Classification

129

Comprehensible classification models: a position paper

130

Neural Network Rule Extraction by a New Ensemble Concept and its Theoretical and Historical Background: a Review

131

Peeking Inside the Black Box: Visualizing Statistical Learning With Plots of Individual Conditional Expectation

132

Learning AND-OR Templates for Object Recognition and Detection

133

Accurate intelligible models with pairwise interactions

134

Improving the interpretability of classification rules discovered by an ant colony algorithm

135

Explaining Data-Driven Document Classifications

136

Interpretable models from distributed data via merging of decision trees

137

Building a diseases symptoms ontology for medical diagnosis: An integrative approach

138

ORC: Ordered Rules for ClassificationA Discrete Optimization Approach to Associative Classification

139

Intelligible models for classification and regression

140

Ontology learning from text: A look back and into the future

141

Extended Taxonomy of Rule Extraction Techniques and Assessment of KDRuleEx

142

Reverse Engineering the Neural Networks for Rule Extraction in Classification Problems

143

Prototype selection for interpretable classification

144

Performance of classification models from a user perspective

145

Opening black box Data Mining models using Sensitivity Analysis

146

An empirical evaluation of the comprehensibility of decision table, tree and rule based predictive models

147

Building comprehensible customer churn prediction models with advanced rule induction techniques

148

REx: An Efficient Rule Generator

149

Explanation and reliability of prediction models: the case of breast cancer recurrence

150

What is an upper level ontology

151

How to Explain Individual Classification Decisions

152

Node harvest: simple and interpretable regression and classication

153

The Cambridge Dictionary of Psychology

154

Classification by Set Cover: The Prototype Vector Machine

155

Decompositional Rule Extraction from Support Vector Machines by Active Learning

156

PREDICTIVE LEARNING VIA RULE ENSEMBLES

157

Explaining Classifications For Individual Instances

158

Recursive Neural Network Rule Extraction for Data With Mixed Attributes

159

Comprehensible Credit Scoring Models Using Rule Extraction from Support Vector Machines

160

Understanding Discrete Classifiers with a Case Study in Gene Prediction

161

Classification With Ant Colony Optimization

162

Ontology - Supported Machine Learning and Decision Support in Biomedicine

163

Real-time beat-synchronous audio effects

164

Rule Extraction from Support Vector Machines: A Sequential Covering Approach

165

Confident Interpretation of Bayesian Decision Tree Ensembles for Clinical Applications

166

Greedy rule generation from discrete data and its use in neural network rule extraction

167

Learning interpretable models

168

ITER: An Algorithm for Predictive Regression Rule Extraction

169

Orthogonal search-based rule extraction (OSRE) for trained neural networks: a practical and efficient approach

170

Defense Advanced Research Projects Agency (DARPA) Agent Markup Language Computer Aided Knowledge Acquisition

172

Discovering the Mysteries of Neural Networks

173

An Explainable Artificial Intelligence System for Small-unit Tactical Behavior

174

Learning Domain Ontologies from Document Warehouses and Dedicated Web Sites

175

Learning with Local Models

176

Extracting symbolic rules from trained neural network ensembles

177

Extracting decision trees from trained neural networks

178

Ontology Learning for the Semantic Web

179

A statistics based approach for extracting priority rules from trained neural networks

180

Bayesian Network Classifiers

181

NeuroLinear: From neural networks to oblique decision rules

182

Three techniques for extracting rules from feedforward networks

183

Beyond Accuracy: What Data Quality Means to Data Consumers

184

Survey and critique of techniques for extracting rules from trained artificial neural networks

185

NeuroRule: A Connectionist Approach to Data Mining

186

Fast Effective Rule Induction

187

Rule Generation from Neural Networks

188

A translation approach to portable ontology specifications

189

Very Simple Classification Rules Perform Well on Most Commonly Used Datasets

190

C4.5: Programs for Machine Learning

191

Review of the Development of Multidimensional Scaling Methods

192

Explanations in knowledge systems: design for explainable expert systems

193

Nonlinear principal component analysis using autoassociative neural networks

194

Minimalist explanations in knowledge-based systems

195

The CN2 Induction Algorithm

196

The need for user models in generating expert system explanation

197

The framing of decisions and the psychology of choice.

198

BLAH, A System Which Explains its Reasoning

199

Judgment under Uncertainty: Heuristics and Biases

200

A Statistically Precise and Relatively Simple Method of Estimating the Bio-Assay with Quantal Response, Based on the Logistic Function

201

Association for Computing Machinery. ISBN 9781450369367

202