Replication and Information Extraction in a Minimal Agent-Environment Model

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Ariosto, Sebastiano, Garnier-Brun, Jerome, Saglietti, Luca, Straziota, Davide
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915519131549696
author Ariosto, Sebastiano
Garnier-Brun, Jerome
Saglietti, Luca
Straziota, Davide
author_facet Ariosto, Sebastiano
Garnier-Brun, Jerome
Saglietti, Luca
Straziota, Davide
contents We consider an unsupervised classifying agent that evolves by enforcing self-consistency of its labels under continual exposure to a data-generating environment. Because the agent's predictions feed back into its own regularized updates, the dynamics can stabilize into self-sustaining modes of operation, which we coin functional replicators. Remarkably, such replicators can spontaneously align with the latent structure of the environment, despite never being exposed to ground-truth labels or selected for adaptation. Using analytical tools from statistical mechanics and numerical experiments, we show that the onset of this regime corresponds to a transition driven by weak correlations between the agent's initial state and environmental structure. Extending the model to multiple agents, we find that their mutual influence can spontaneously break symmetry and produce consensus, illustrating a minimal setting for decentralized collective learning.
format Preprint
id arxiv_https___arxiv_org_abs_2509_23212
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Replication and Information Extraction in a Minimal Agent-Environment Model
Ariosto, Sebastiano
Garnier-Brun, Jerome
Saglietti, Luca
Straziota, Davide
Disordered Systems and Neural Networks
Statistical Mechanics
Adaptation and Self-Organizing Systems
We consider an unsupervised classifying agent that evolves by enforcing self-consistency of its labels under continual exposure to a data-generating environment. Because the agent's predictions feed back into its own regularized updates, the dynamics can stabilize into self-sustaining modes of operation, which we coin functional replicators. Remarkably, such replicators can spontaneously align with the latent structure of the environment, despite never being exposed to ground-truth labels or selected for adaptation. Using analytical tools from statistical mechanics and numerical experiments, we show that the onset of this regime corresponds to a transition driven by weak correlations between the agent's initial state and environmental structure. Extending the model to multiple agents, we find that their mutual influence can spontaneously break symmetry and produce consensus, illustrating a minimal setting for decentralized collective learning.
title Replication and Information Extraction in a Minimal Agent-Environment Model
topic Disordered Systems and Neural Networks
Statistical Mechanics
Adaptation and Self-Organizing Systems
url https://arxiv.org/abs/2509.23212