Connecting Dualities and Machine Learning

Abstract

Dualities are widely used in quantum field theories and string theory to obtain correlation functions at high accuracy. Here we present examples where dual data representations are useful in supervised classification, linking machine learning and typical tasks in theoretical physics. We then discuss how such beneficial representations can be enforced in the latent dimension of neural networks. We find that additional contributions to the loss based on feature separation, feature matching with respect to desired representations, and a good performance on a `simple' correlation function can lead to known and unknown dual representations. This is the first proof of concept that computers can find dualities. We discuss how our examples, based on discrete Fourier transformation and Ising models, connect to other dualities in theoretical physics, for instance Seiberg duality.

arXiv comment35 pages, 19 figures

References in corpus (12)

• FaceNet: A Unified Embedding for Face Recognition and Clustering
• The Large N Limit of Superconformal Field Theories and Supergravity
• Large N Field Theories, String Theory and Gravity
• Viscosity in Strongly Interacting Quantum Field Theories from Black Hole Physics
• Electric-Magnetic Duality in Supersymmetric Non-Abelian Gauge Theories
• Shear viscosity of strongly coupled N=4 supersymmetric Yang-Mills plasma
• Lectures on supersymmetric gauge theories and electric-magnetic duality
• Mirror Symmetry, Mirror Map and Applications to Calabi-Yau Hypersurfaces
• Mirror Symmetry for Two Parameter Models -- I
• Lectures on Mirror Symmetry
• Duality and Global Symmetries
• One-dimensional Ising model with multispin interactions

Cited by in corpus (6)

• Machine Learning and LHC Event Generation
• Quiver Mutations, Seiberg Duality and Machine Learning
• Nonperturbative renormalization for the neural network-QFT correspondence
• Machine learning for complete intersection Calabi-Yau manifolds: a methodological study
• Nonlinear 3D Cosmic Web Simulation with Heavy-Tailed Generative Adversarial Networks
• Black holes and the loss landscape in machine learning