arXiv:2606.14156v1 Announce Type: new 
Abstract: Learning systems based on IF-THEN rule representations readily offer interpretability, making them a crucial focus in contemporary AI research. A key objective for such rule sets is to achieve both high discriminative power and interpretability. While existing state-of-the-art algorithms implicitly prioritize predictive accuracy, they often fall short on one or more quality metrics that ensure interpretability, such as coverage and parsimony of rule sets. Motivated by this, this paper propose the development of CDPR, which aims to create highly accurate and interpretable rule sets for classification problems. To the best of our knowledge, this represents the first attempt to establish such an approach. In this study, we introduce two algorithms rooted in submodular maximization, which not only provide provable guarantees on coverage but also yield rule sets that are both discriminative and parsimonious. We empirically demonstrate that rule sets learned through our approaches achieve higher accuracy and interpretability and has more than a 2.5-fold improvement in average coverage rates when compared to the next best algorithm.

تقدم دراسة جديدة إطار عمل CDPR، المصمم لإنشاء مجموعات قواعد ذات تغطية عالية، تمييزية وبارسيمونية لمشكلات التصنيف، مما يعالج الحاجة إلى القابلية للتفسير في أنظمة الذكاء الاصطناعي. تستخدم هذه الطريقة خوارزميات تعتمد على تعظيم تحت المودولار لضمان الدقة والقابلية للتفسير في مجموعات القواعد.

Un nuevo estudio presenta CDPR, un marco diseñado para crear conjuntos de reglas de alta cobertura, discriminativos y parcimoniosos para problemas de clasificación, abordando la necesidad de interpretabilidad en los sistemas de IA. Este enfoque utiliza algoritmos basados en la maximización submodular para garantizar tanto la precisión como la interpretabilidad en los conjuntos de reglas.

Une nouvelle étude présente CDPR, un cadre conçu pour créer des ensembles de règles à haute couverture, discriminatoires et parcimonieux pour des problèmes de classification, répondant au besoin d'interprétabilité dans les systèmes d'IA. Cette approche utilise des algorithmes basés sur la maximisation sous-modulaire pour garantir à la fois précision et interprétabilité dans les ensembles de règles.

A new study introduces CDPR, a framework designed to create high coverage, discriminative, and parsimonious rulesets for classification problems, addressing the need for interpretability in AI systems. This approach utilizes algorithms based on submodular maximization to ensure both accuracy and interpretability in rule sets.