Weight-Sharing Regularization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shakerinava, Mehran, Sohrabi, Motahareh, Ravanbakhsh, Siamak, Lacoste-Julien, Simon
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913259646353408
author Shakerinava, Mehran
Sohrabi, Motahareh
Ravanbakhsh, Siamak
Lacoste-Julien, Simon
author_facet Shakerinava, Mehran
Sohrabi, Motahareh
Ravanbakhsh, Siamak
Lacoste-Julien, Simon
contents Weight-sharing is ubiquitous in deep learning. Motivated by this, we propose a "weight-sharing regularization" penalty on the weights $w \in \mathbb{R}^d$ of a neural network, defined as $\mathcal{R}(w) = \frac{1}{d - 1}\sum_{i > j}^d |w_i - w_j|$. We study the proximal mapping of $\mathcal{R}$ and provide an intuitive interpretation of it in terms of a physical system of interacting particles. We also parallelize existing algorithms for $\operatorname{prox}_\mathcal{R}$ (to run on GPU) and find that one of them is fast in practice but slow ($O(d)$) for worst-case inputs. Using the physical interpretation, we design a novel parallel algorithm which runs in $O(\log^3 d)$ when sufficient processors are available, thus guaranteeing fast training. Our experiments reveal that weight-sharing regularization enables fully connected networks to learn convolution-like filters even when pixels have been shuffled while convolutional neural networks fail in this setting. Our code is available on github.
format Preprint
id arxiv_https___arxiv_org_abs_2311_03096
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Weight-Sharing Regularization
Shakerinava, Mehran
Sohrabi, Motahareh
Ravanbakhsh, Siamak
Lacoste-Julien, Simon
Machine Learning
Weight-sharing is ubiquitous in deep learning. Motivated by this, we propose a "weight-sharing regularization" penalty on the weights $w \in \mathbb{R}^d$ of a neural network, defined as $\mathcal{R}(w) = \frac{1}{d - 1}\sum_{i > j}^d |w_i - w_j|$. We study the proximal mapping of $\mathcal{R}$ and provide an intuitive interpretation of it in terms of a physical system of interacting particles. We also parallelize existing algorithms for $\operatorname{prox}_\mathcal{R}$ (to run on GPU) and find that one of them is fast in practice but slow ($O(d)$) for worst-case inputs. Using the physical interpretation, we design a novel parallel algorithm which runs in $O(\log^3 d)$ when sufficient processors are available, thus guaranteeing fast training. Our experiments reveal that weight-sharing regularization enables fully connected networks to learn convolution-like filters even when pixels have been shuffled while convolutional neural networks fail in this setting. Our code is available on github.
title Weight-Sharing Regularization
topic Machine Learning
url https://arxiv.org/abs/2311.03096