Guardado en:
Detalles Bibliográficos
Autor principal: Nakajima, Shuta
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2512.23195
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914223385214976
author Nakajima, Shuta
author_facet Nakajima, Shuta
contents We study the replica-symmetric saddle point equations for the Ising perceptron with Gaussian disorder and margin $κ\ge 0$. We prove that for each $κ\ge 0$ there is a critical capacity $α_c(κ)=\frac{2}{π\,\mathbb E[(κ-Z)_+^2]}$, where $Z$ is a standard normal and $(x)_+=\max\{x,0\}$, such that the saddle point equation has a unique solution for $α\in(0,α_c(κ))$ and has no solution when $α\ge α_c(κ)$. When $α\uparrow α_c(κ)$ and $κ>0$, the replica-symmetric free energy at this solution diverges to $-\infty$. In the zero-margin case $κ=0$, Ding and Sun obtained a conditional uniqueness result, with one step verified numerically. Our argument gives a fully analytic proof without computer assistance. We used GPT-5 to help develop intermediate proof steps and to perform sanity-check computations.
format Preprint
id arxiv_https___arxiv_org_abs_2512_23195
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Uniqueness of Replica-symmetric Saddle Point for Ising Perceptron
Nakajima, Shuta
Probability
Statistical Mechanics
We study the replica-symmetric saddle point equations for the Ising perceptron with Gaussian disorder and margin $κ\ge 0$. We prove that for each $κ\ge 0$ there is a critical capacity $α_c(κ)=\frac{2}{π\,\mathbb E[(κ-Z)_+^2]}$, where $Z$ is a standard normal and $(x)_+=\max\{x,0\}$, such that the saddle point equation has a unique solution for $α\in(0,α_c(κ))$ and has no solution when $α\ge α_c(κ)$. When $α\uparrow α_c(κ)$ and $κ>0$, the replica-symmetric free energy at this solution diverges to $-\infty$. In the zero-margin case $κ=0$, Ding and Sun obtained a conditional uniqueness result, with one step verified numerically. Our argument gives a fully analytic proof without computer assistance. We used GPT-5 to help develop intermediate proof steps and to perform sanity-check computations.
title Uniqueness of Replica-symmetric Saddle Point for Ising Perceptron
topic Probability
Statistical Mechanics
url https://arxiv.org/abs/2512.23195