Sampling String Vacua Using Generative Models

2025/09/19 by Walden, Moritz, Larfors, Magdalena · 1 citation
#FOS: Physical sciences #High Energy Physics - Theory (hep-th)

paper · doi:10.48550/arxiv.2509.16029

Abstract

We apply generative models to a key problem in the string compactification program, namely construction of type IIB string vacua. To this end, we make use of a Bayesian Flow Network, a generative model capable of handling discrete data, to generate flux vectors that give rise to type IIB vacua. Furthermore, we sample flux vacua that have certain desirable properties by employing a Transformer as a conditional generative model. Both models demonstrate good performance in finding flux vacua and thus prove to be powerful tools in the exploration of the string landscape.

Citations

• Transforming Calabi-Yau Constructions: Generating New Calabi-Yau Manifolds with Transformers
• Solving inverse problems of Type IIB flux vacua with conditional generative models
• Learning Topological Invariance
• Int2Int: a framework for mathematics with transformers
• BPS spectroscopy with reinforcement learning
• Deep observations of the Type IIB flux landscape
• TASI Lectures on Physics for Machine Learning
• Candidate de Sitter Vacua
• Unifying Bayesian Flow Networks and Diffusion Models through Stochastic Differential Equations
• Lectures on Numerical and Machine Learning Methods for Approximating Ricci-flat Calabi-Yau Metrics
• Learning the greatest common divisor: explaining transformer predictions
• Bayesian Flow Networks
• JAXVacua -- A Framework for Sampling String Vacua
• Decoding Nature with Nature's Tools: Heterotic Line Bundle Models of Particle Physics with Genetic Algorithms and Quantum Annealing
• Machine Learning in Physics and Geometry
• CYTools: A Software Package for Analyzing Calabi-Yau Manifolds
• Towards structures in the flux landscape at large number of moduli
• Diffusion Models: A Comprehensive Survey of Methods and Applications
• Classifier-Free Diffusion Guidance
• Evolving Heterotic Gauge Backgrounds: Genetic Algorithms versus Reinforcement Learning
• Generative Adversarial Networks
• Revealing systematics in phenomenologically viable flux vacua with reinforcement learning
• A Survey of Transformers
• Statistical Predictions in String Theory and Deep Generative Models
• Attention Is All You Need
• Deep Unsupervised Learning using Nonequilibrium Thermodynamics
• de Sitter vacua in string theory

Cited by

• Machine Learning the Conformal Manifold of Holographic CFT₂s

Related

• Solving inverse problems of Type IIB flux vacua with conditional generative models
• JAXVacua -- A Framework for Sampling String Vacua
• Les Houches Lectures on Constructing String Vacua
• Probing the Structure of String Theory Vacua with Genetic Algorithms and Reinforcement Learning
• Statistics of String vacua