Extracting symbolic rules from trained neural network ensembles

By Zhi-Hua Zhou, Yuan Jiang, Shifu Chen. 2003

216 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.

Citations in the corpus

Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, neutral): “Multi-Layer Neural Networks Explanation by simplification Rule-based learner: [82, 83, 147, 148, 246, 247, 248, 249, 250, 251]”
Logistic Model Tree Extraction From Artificial Neural Networks (2007, background): “A further advantage of the pedagogical approaches is that they can be applied to any “black-box” classifier, such as ensembles of neural networks [34].”
Learnware: on the future of machine learning (2016, background): “trying to improve the comprehensibility of black-box models [8], or even produce accurate and comprehensible models [9].”
Deep Class-Incremental Learning: A Survey (2023, methodology): “’ Assuming the old model is a good classifier for all the seen classes in Yb−1, how can we utilize it to resist forgetting in the new model? To enable the old model to assist the new model, an intuitive way utilizes the concept proposed in [144], [159], [160], i.”
Characterization of Symbolic Rules Embedded in Deep DIMLP Networks: A Challenge to Transparency of Deep Learning (2017, methodology): “Zhou et al. introduced the REFNE algorithm (Rule Extraction from Neural Network Ensemble) [12], which utilizes the trained ensembles to generate instances and then extracts symbolic rules from those instances.”
Seeing the Forest Through the Trees (2007, background): “Zhou et al. [19] utilize neural network ensembles to generate new instances and then extract symbolic rules from those instances.”
A Survey on the Explainability of Supervised Machine Learning (2020, methodology): “BRAINNE Feature Selection Classification Pedagogical Rule-based Sestito and Dilan (1992) VIA Sensitivity analysis Classification Pedagogical Rule-based Zilke et al. (2016) BIO-RE Feature Selection Classification Pedagogical Rule-based Taha and Ghosh (1996)”
Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence (2023, methodology): “The Rule Extraction from Neural Network Ensemble (REFNE), developed by Zhou et al. [100], was able to prevent pointless discretisations by using adaptive intervals.”
Explainable Artificial Intelligence: a Systematic Review (2020, background): “(REFNE) [247] extracts symbolic rules from instances generated by neural network ensembles.”
Questioning the AI: Informing Design Practices for Explainable AI User Experiences (2020, background): “, if-then rules [26, 93, 102] How, Why, Why not, What if”
Rule-Based Learning Systems for Support Vector Machines (2006, unknown): “With the aim of making the resulting classifier system interpretable, in the last 15 years rule extraction methods for trained neural networks have been developed [1, 5, 15, 22, 25].”
Rule Extraction from Support Vector Machines: An Overview of Issues and Application in Credit Scoring (2008, methodology): “5 (1993) [49] P C Decision tree induction TREPAN (1996) [18] P C Decision tree induction, M-of-N splits BIO-RE (1999) [56] P C Creates complete truth table, only applicable to toy problems ANN-DT (1999) [52] P C + R Decision tree induction, similar to TREPAN DecText (2000) [10] P C Decision tree ind”
A Model or 603 Exemplars: Towards Memory-Efficient Class-Incremental Learning (2022, methodology): “, 2017) builds knowledge distillation (Zhou et al., 2003; Zhou & Jiang, 2004; Hinton et al., 2015) regularization with exemplars to align the predictions of old and new models.”
Trustworthy AI: From Principles to Practices (2021, methodology): “been proposed to approximate ML models, such as random forests [317, 392] and neural networks [28, 86, 393].”