Rule Extraction Algorithm for Deep Neural Networks: A Review

By Tameru Hailesilassie. 2016

104 citations recorded by the source corpus.

Despite the highest classification accuracy in wide varieties of application areas, artificial neural network has one disadvantage. The way this Network comes to a decision is not easily comprehensible. The lack of explanation ability reduces the acceptability of neural network in data mining and decision system. This drawback is the reason why researchers have proposed many rule extraction algorithms to solve the problem. Recently, Deep Neural Network (DNN) is achieving a profound result over the standard neural network for classification and recognition problems. It is a hot machine learning area proven both useful and innovative. This paper has thoroughly reviewed various rule extraction algorithms, considering the classification scheme: decompositional, pedagogical, and eclectics. It also presents the evaluation of these algorithms based on the neural network structure with which the algorithm is intended to work. The main contribution of this review is to show that there is a limited study of rule extraction algorithm from DNN.

Subfields: Artificial Intelligence, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Safety, Robustness & Interpretability

Citations in the corpus

Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, neutral): “Automatic-Rule Extraction [217, 251, 252, 253, 254, 255, 256, 319, 320, 321]”
Explaining Explanations: An Overview of Interpretability of Machine Learning (2018, background): “Some of these sampling approaches only work on binary input [34] or use genetic algorithms to produce new training examples [35].”
Black-Box vs. White-Box: Understanding Their Advantages and Weaknesses From a Practical Point of View (2019, methodology): “At the end, all rules for one class are getting merged [156]–[158].”
Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI) (2018, background): “One effort to gain insight into highly complex models is the use of rule extraction [114]–[116].”
Trustworthy Artificial Intelligence: A Review (2022, methodology): “Hailesilassie [84] reviewed various rule-extraction techniques to give comprehensibility to artificial neural networks.”
Characterization of Symbolic Rules Embedded in Deep DIMLP Networks: A Challenge to Transparency of Deep Learning (2017, background): “However, very few works have investigated rule extraction from neural networks trained by deep learning, the main difficulty being the complexity of knowledge representation in the hidden layers [2].”
A comprehensive taxonomy for explainable artificial intelligence: a systematic survey of surveys on methods and concepts (2021, background): “While the previous surveys concentrate on directly training inherently interpretable models consisting of rules, Hailesilassie in 2016 shortly reviewed rule extraction methods [60].”
Explainable Artificial Intelligence: a Systematic Review (2020, methodology): “Scholars divided these methods into three classes [41, 132]: (I) decompositional methods work by extracting rules at the level of hidden and output neurons by analysing the values of their weights, (II) pedagogical methods treat an underlying neural network as a black-box and the rule extraction con”
Explainable Sentiment Analysis with Applications in Medicine (2018, background): “In the review of Hailesilassie [51] on Rule Extraction Algorithm for DNNs, three main approaches are highlighted, i.e. Decompositional, Pedagogical, and Eclectic.”
Learning Causal State Representations of Partially Observable Environments (2019, methodology): “The idea of extracting the implicit knowledge of a neural network [37, 11, 28, 15] is not novel and is rooted in early attempts to merge traditional rule-based methods with machine learning.”
Trends and Trajectories for Explainable, Accountable and Intelligible Systems: An HCI Research Agenda (2018, methodology): “Other works explored how to design for implicit interaction with sensing-based systems [14, 75, 86, 175], and help people predict what will happen through feedforward [12, 176].”
Explaining Explanations: An Approach to Evaluating Interpretability of Machine Learning (2018, background): “Some of these sampling approaches only work on binary input [34] or use genetic algorithms to produce new training examples [35].”