Spectral Architecture Search for Neural Network Models

Gianluca Peri, Lorenzo Chicchi, Duccio Fanelli, Lorenzo Giambagli
4/1/2025
cs.LGcond-mat.dis-nncond-mat.stat-mech

No official code found. Best implementation: gianluca-peri/sparcs-test (1 star, last commit 2025-10-20).

Abstract

Architecture design and optimization are challenging problems in the field of artificial neural networks. Working in this context, we here present SPARCS (SPectral ARchiteCture Search), a novel architecture search protocol which exploits the spectral attributes of the inter-layer transfer matrices. SPARCS allows one to explore the space of possible architectures by spanning continuous and differentiable manifolds, thus enabling for gradient-based optimization algorithms to be eventually employed. With reference to simple benchmark models, we show that the newly proposed method yields a self-emerging architecture with a minimal degree of expressivity to handle the task under investigation and with a reduced parameter count as compared to other viable alternatives.

View on arXivView PDF

Code Implementations(1)

gianluca-peri/sparcs-testCommunity100%

This repo holds the code to reproduce the results present in the SPARCS paper.

10Shell, PythonFeb 20, 20251 years ago
neural-architecture-searchneural-networksspectral-methods

Similar Papers

OptiProxy-NAS: Optimization Proxy based End-to-End Neural Architecture Search

85% match1 repo22025

LayerNAS: Neural Architecture Search in Polynomial Complexity

84% match2023

Efficient Global Neural Architecture Search

83% match2025

A Survey on Neural Architecture Search Based on Reinforcement Learning

83% match2024

Structure of Artificial Neural Networks -- Empirical Investigations

82% match2024

A Neural Architecture Search Method using Auxiliary Evaluation Metric based on ResNet Architecture

82% match2025

Cite this paper

@article{peri2025spectral,
  title  = {Spectral Architecture Search for Neural Network Models},
  author = {Gianluca Peri and Lorenzo Chicchi and Duccio Fanelli and Lorenzo Giambagli},
  year   = {2025},
  eprint = {2504.00885},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2504.00885v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2504.00885.svg)](https://codeofpaper.com/paper/2504.00885)

Discussion