Mapping the Phase Diagram of the Vicsek Model with Machine Learning

Forum topic · 小凯 · 2026-05-02

Summary

Researchers Jonathan Bongolan, Guillermo Romer, and Christian Alis applied machine learning to map the phase structure of the Vicsek flocking model across its three-dimensional parameter space (eta, rho, v_0). They built a dataset of simulated parameter points characterized by long-time dynamical observables, then used K-Means clustering to assign each point to a disorder, order, or coexistence phase. These clustered labels served as training data for a neural-network classifier that learned the mapping from model parameters to phase behavior, achieving 0.92 classification accuracy. The resulting phase diagram resolves the narrow coexistence region separating the ordered and disordered phases and extends the inferred phase boundaries beyond the originally sampled simulation points. Published as arXiv preprint 2604.28167, the study demonstrates a systematic pipeline for converting sparse simulation data into global phase diagrams of collective motion models, combining unsupervised clustering with supervised classification.

Paper Overview

Field: ML/Physics
Authors: Jonathan Bongolan, Guillermo Romer, Christian Alis
Published: 2026-04-30
arXiv: 2604.28167

Summary

In this study, the authors use machine learning to classify and interpolate the phase structure of the Vicsek flocking model across the three-dimensional parameter space (eta, rho, v_0).

The methodology proceeds in stages:

1. A dataset of simulated parameter points is constructed, with each point characterized using long-time dynamical observables. 2. These observables serve as inputs to a K-Means clustering procedure, which assigns each point to a disorder, order, or coexistence phase. 3. The clustered labels are then used to train a neural-network classifier that learns the mapping from model parameters to phase behavior, reaching a classification accuracy of 0.92.

The resulting phase diagram resolves the narrow coexistence region separating the ordered and disordered phases, and the inferred phase boundaries extend beyond the originally sampled simulation points.

More broadly, the authors present this approach as a systematic pipeline for transforming sparse simulation data into global phase diagrams of collective motion models.

Original Abstract (excerpt)

> In this study, we use machine learning to classify and interpolate the phase structure of the Vicsek flocking model across the three-dimensional parameter space (eta, rho, v_0). We construct a dataset of simulated parameter points and characterize each point using long-time dynamical observables. These observables are then used as inputs to a K-Means clustering procedure, which assigns each point to a disorder, order, or coexistence phase. Using these clustered labels, we train a neural-network...

---

*Auto-collected on 2026-05-02*

Mapping the Phase Diagram of the Vicsek Model with Machine Learning

Summary

Paper Overview

Summary

Original Abstract (excerpt)

Tags