# Few-shot machine learning in the three-dimensional Ising model

## TL;DR

This paper demonstrates that few-shot machine learning strategies can accurately predict phase transitions in large 3D Ising models based on training on smaller lattices, offering an efficient approach for studying complex spin systems.

## Contribution

The paper introduces a novel few-shot machine learning approach for predicting phase transitions in 3D Ising models, combining supervised and unsupervised methods for high accuracy.

## Key findings

- Supervised models achieve ~99% accuracy in phase classification.
- Unsupervised models effectively reconstruct and classify spin configurations.
- Few-shot strategy enables efficient analysis of larger lattices from smaller training samples.

## Abstract

We investigate theoretically the phase transition in three dimensional cubic Ising model utilizing state-of-the-art machine learning algorithms. Supervised machine learning models show high accuracies (~99\%) in phase classification and very small relative errors ($< 10^{-4}$) of the energies in different spin configurations. Unsupervised machine learning models are introduced to study the spin configuration reconstructions and reductions, and the phases of reconstructed spin configurations can be accurately classified by a linear logistic algorithm. Based on the comparison between various machine learning models, we develop a few-shot strategy to predict phase transitions in larger lattices from trained sample in smaller lattices. The few-shot machine learning strategy for three dimensional(3D) Ising model enable us to study 3D ising model efficiently and provides a new integrated and highly accurate approach to other spin models.