Unsupervised Machine Learning of the Contact Process
Abstract
We investigate how unsupervised machine-learning methods can characterize the active-absorbing phase transition in the contact process in one and two spatial dimensions. Our analysis focuses on principal component analysis (PCA) and variational autoencoders (VAEs), and we show that direct applications of these methods encounter a central difficulty: contact process configurations are intrinsically positive-definite. In the PCA case, standard centering procedures fail to isolate the order parameter in the leading component. To address this, we augment the data with sign-reversed configurations, creating a balanced dataset with zero mean while preserving the density information required to describe the transition. This preprocessing allows the dominant principal component to recover the order parameter and reproduce the expected finite-size scaling near criticality. For VAEs, we find a related limitation: a fixed Gaussian prior does not provide sufficient latent regularization across the full range of control parameters and system sizes for positive-definite data. We overcome this by adopting a heteroscedastic Gaussian prior that adapts to the control parameter, leading to a substantial improvement in the finite-size scaling of reconstruction-based observables near the critical threshold. Taken together, these results highlight a key lesson: the success of machine learning in phase-transition problems does not rely on algorithmic complexity alone, but on tailoring the representation to respect the physical structure of the system, including its intrinsic constraints and symmetries.
Related
- Unsupervised and probabilistic learning with Contrastive Local Learning Networks: The Restricted Kirchhoff Machine
- Machine Learning β-decay Half-lives and Their Application to r-Process Observables
- Machine Learning Assisted Design of Complex and High Entropy Alloys by Hybrid HiPIMS/Pulsed-DC PVD Process for Low Carbon Energy Applications in Extreme Environments
- Entanglement-enabled Criticality in One-dimensional Quantum Contact Process
- Soft Contrastive Learning for Unsupervised Discovery of Phases of Matter
BibTeX
@misc{brito2026unsupervised,
title = {{Unsupervised Machine Learning of the Contact Process}},
author = {L. T. Brito and G. A. Alves and F. W. S. Lima and A. Macedo-Filho and R. S. Ferreira and T. F. A. Alves},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.16506},
eprint = {2609.16506},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.16506},
}
RIS
TY - UNPB
TI - Unsupervised Machine Learning of the Contact Process
AU - L. T. Brito
AU - G. A. Alves
AU - F. W. S. Lima
AU - A. Macedo-Filho
AU - R. S. Ferreira
AU - T. F. A. Alves
PY - 2026
DA - 2026/09/16
PB - arXiv
DO - 10.48550/arxiv.2609.16506
UR - https://arxiv.org/abs/2609.16506
AB - We investigate how unsupervised machine-learning methods can characterize the active-absorbing phase transition in the contact process in one and two spatial dimensions. Our analysis focuses on principal component analysis (PCA) and variational autoencoders (VAEs), and we show that direct applications of these methods encounter a central difficulty: contact process configurations are intrinsically positive-definite. In the PCA case, standard centering procedures fail to isolate the order parameter in the leading component. To address this, we augment the data with sign-reversed configurations, creating a balanced dataset with zero mean while preserving the density information required to describe the transition. This preprocessing allows the dominant principal component to recover the order parameter and reproduce the expected finite-size scaling near criticality. For VAEs, we find a related limitation: a fixed Gaussian prior does not provide sufficient latent regularization across the full range of control parameters and system sizes for positive-definite data. We overcome this by adopting a heteroscedastic Gaussian prior that adapts to the control parameter, leading to a substantial improvement in the finite-size scaling of reconstruction-based observables near the critical threshold. Taken together, these results highlight a key lesson: the success of machine learning in phase-transition problems does not rely on algorithmic complexity alone, but on tailoring the representation to respect the physical structure of the system, including its intrinsic constraints and symmetries.
ER -