Rule extraction from neural networks — A comparative study

55 citations recorded by the source corpus.

How researchers see this paper

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

Rule Extraction Algorithm for Deep Neural Networks: A Review (2016, background): “Regarding computational limitation and execution time, pedagogical approach is better than decompositional [29].”
Extended Taxonomy of Rule Extraction Techniques and Assessment of KDRuleEx (2012, background): “Many researchers have discussed rule extraction techniques, their evaluation criterion and taxonomy [11], [12], [13], [14].”
RuleMatrix: Visualizing and Understanding Classifiers with Rules (2018, methodology): “There are mainly three types of methods to derive approximate models (often as rule sets) as summarized in related surveys [3, 4], namely, decompositional, pedagogical and eclectic.”
Interpretable Machine Learning - A Brief History, State-of-the-Art and Challenges (2020, methodology): “ng strategy, or the interpretable model that is used. There are also methods for extracting, e.g., decision rules from specic models based on their internal components such as neural network weights [5,9]. LIME [95] is an example of a local surrogate method that explains individual predictions by le”
A comprehensive taxonomy for explainable artificial intelligence: a systematic survey of surveys on methods and concepts (2021, methodology): “For further rule extraction examples, we refer the reader to the comprehensive surveys Hailesilassie (2016); Wang et al (2018a); Augasta and Kathirvalavakumar (2012) on the topic and the survey by Wang et al (2018b) for recurrent DNNs.”