English

Smooth Min-Max Monotonic Networks

Machine Learning 2024-05-28 v3 Artificial Intelligence

Abstract

Monotonicity constraints are powerful regularizers in statistical modelling. They can support fairness in computer-aided decision making and increase plausibility in data-driven scientific models. The seminal min-max (MM) neural network architecture ensures monotonicity, but often gets stuck in undesired local optima during training because of partial derivatives of the MM nonlinearities being zero. We propose a simple modification of the MM network using strictly-increasing smooth minimum and maximum functions that alleviates this problem. The resulting smooth min-max (SMM) network module inherits the asymptotic approximation properties from the MM architecture. It can be used within larger deep learning systems trained end-to-end. The SMM module is conceptually simple and computationally less demanding than state-of-the-art neural networks for monotonic modelling. Our experiments show that this does not come with a loss in generalization performance compared to alternative neural and non-neural approaches.

Keywords

Cite

@article{arxiv.2306.01147,
  title  = {Smooth Min-Max Monotonic Networks},
  author = {Christian Igel},
  journal= {arXiv preprint arXiv:2306.01147},
  year   = {2024}
}
R2 v1 2026-06-28T10:54:02.202Z