3D Tensor Renormalisation Group at High Temperatures

Abstract

Building upon previous $2D$ studies, this research focuses on describing $3D$ tensor renormalisation group (RG) flows for lattice spin systems, such as the Ising model. We present a novel RG map, which operates on tensors with infinite-dimensional legs and does not involve truncations, in contrast to numerical tensor RG maps. To construct this map, we developed new techniques for analysing tensor networks. Our analysis shows that the constructed RG map contracts the region around the tensor $A_{\ast }$, corresponding to the high-temperature phase of the $3D$ Ising model. This leads to the iterated RG map convergence in the Hilbert-Schmidt norm to $A_{\ast }$ when initialised in the vicinity of $A_{\ast }$. This work provides the first steps towards the rigorous understanding of tensor RG maps in $3D$.

Annales Henri Poincaré (2024)

References in corpus (3)

• Tensor renormalization group approach to 2D classical lattice models
• Tensor RG approach to high-temperature fixed point
• Tensor Renormalization Group at Low Temperatures: Discontinuity Fixed Point

Cited by in corpus (4)

• Three-dimensional real space renormalization group with well-controlled approximations
• Rotations, Negative Eigenvalues, and Newton Method in Tensor Network Renormalization Group
• A Practical Introduction to Tensor Network Renormalization with TNRKit.jl
• Lattice and PT symmetries in tensor-network renormalization group: Case study of a hard-square lattice gas model