Uninorm-like parametric activation functions for human-understandable neural models

We present a deep learning model for finding human-understandable connections between input features. Our approach uses parameterized, differentiable activation function, based on the theoretical background of nilpotent fuzzy logic and multi-criteria decision-making (MCDM). The learnable parameter has semantic meaning indicating level compensation neural network determines parameters using gradient descent to find relationships demonstrate utility effectiveness by successfully applying it classification problems from UCI Machine Learning Repository.