Stably unactivated neurons in ReLU neural networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Brownlowe, Natalie, Cornwell, Christopher R., Montes, Ethan, Quijano, Gabriel, Stulman, Grace, Zhang, Na
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909431300620288
author Brownlowe, Natalie
Cornwell, Christopher R.
Montes, Ethan
Quijano, Gabriel
Stulman, Grace
Zhang, Na
author_facet Brownlowe, Natalie
Cornwell, Christopher R.
Montes, Ethan
Quijano, Gabriel
Stulman, Grace
Zhang, Na
contents The choice of architecture of a neural network influences which functions will be realizable by that neural network and, as a result, studying the expressiveness of a chosen architecture has received much attention. In ReLU neural networks, the presence of stably unactivated neurons can reduce the network's expressiveness. In this work, we investigate the probability of a neuron in the second hidden layer of such neural networks being stably unactivated when the weights and biases are initialized from symmetric probability distributions. For networks with input dimension $n_0$, we prove that if the first hidden layer has $n_0+1$ neurons then this probability is exactly $\frac{2^{n_0}+1}{4^{n_0+1}}$, and if the first hidden layer has $n_1$ neurons, $n_1 \le n_0$, then the probability is $\frac{1}{2^{n_1+1}}$. Finally, for the case when the first hidden layer has more neurons than $n_0+1$, a conjecture is proposed along with the rationale. Computational evidence is presented to support the conjecture.
format Preprint
id arxiv_https___arxiv_org_abs_2412_06829
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Stably unactivated neurons in ReLU neural networks
Brownlowe, Natalie
Cornwell, Christopher R.
Montes, Ethan
Quijano, Gabriel
Stulman, Grace
Zhang, Na
Machine Learning
Probability
The choice of architecture of a neural network influences which functions will be realizable by that neural network and, as a result, studying the expressiveness of a chosen architecture has received much attention. In ReLU neural networks, the presence of stably unactivated neurons can reduce the network's expressiveness. In this work, we investigate the probability of a neuron in the second hidden layer of such neural networks being stably unactivated when the weights and biases are initialized from symmetric probability distributions. For networks with input dimension $n_0$, we prove that if the first hidden layer has $n_0+1$ neurons then this probability is exactly $\frac{2^{n_0}+1}{4^{n_0+1}}$, and if the first hidden layer has $n_1$ neurons, $n_1 \le n_0$, then the probability is $\frac{1}{2^{n_1+1}}$. Finally, for the case when the first hidden layer has more neurons than $n_0+1$, a conjecture is proposed along with the rationale. Computational evidence is presented to support the conjecture.
title Stably unactivated neurons in ReLU neural networks
topic Machine Learning
Probability
url https://arxiv.org/abs/2412.06829