# Explicit neural network classifiers for non-separable data

arXiv: https://arxiv.org/abs/2504.18710 · HTML: https://mlsanity.com/paper/2504.18710/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 6 of its 8 references that are in the map.
- First posted on arXiv 25 Apr 2025.

## Builds on (most cited first, 6 of 6)

- [Approximating Continuous Functions by ReLU Nets of Minimal Width](https://mlsanity.com/paper/1710.11278/) (2017)
- [Universal Function Approximation by Deep Neural Nets with Bounded Width and ReLU Activations](https://mlsanity.com/paper/1708.02691/) (2017)
- [Deep, Skinny Neural Networks are not Universal Approximators](https://mlsanity.com/paper/1810.00393/) (2018)
- [Geometric structure of Deep Learning networks and construction of global mathcal L2 minimizers](https://mlsanity.com/paper/2309.10639/) (2023)
- [The Geometric Structure of Fully-Connected ReLU Layers](https://mlsanity.com/paper/2310.03482/) (2023)
- [Interpretable global minima of deep ReLU neural networks on sequentially separable data](https://mlsanity.com/paper/2405.07098/) (2024)

## Abstract (arXiv)

We fully characterize a large class of feedforward neural networks in terms of truncation maps. As an application, we show how a ReLU neural network can implement a feature map which separates concentric data.