interpretable-learning

iFuzz-Meta is an interpretable fuzzy learning framework that combines top-down knowledge integration with bottom-up data-driven adaptation to enhance transparency and generalization in neural models.

The paper introduces Top-NMF, a framework that incorporates topological regularization via persistent homology into non-negative matrix factorization to learn interpretable and structurally coherent bases, applicable to spatially coherent image components, periodic time-series structures, and clique-like graph signals.