pith. sign in

arxiv: 2508.05406 · v2 · submitted 2025-08-07 · ❄️ cond-mat.str-el · cond-mat.stat-mech

Global Tensor Network Renormalization for 2D Quantum systems: A new window to probe universal data from thermal transitions

Atsushi Ueda , Sander De Meyer , Adwait Naravane , Victor Vanthilt , Frank Verstraete This is my paper

Pith reviewed 2026-05-19 00:40 UTC · model grok-4.3

classification ❄️ cond-mat.str-el cond-mat.stat-mech
keywords tensor network renormalizationglobal optimizationfinite-temperature density matrixconformal field theorythermal transitions2D quantum systemsphase transitionsuniversal data
0
0 comments X

The pith

Combining global optimization TNR with a new finite-temperature density matrix construction yields highly accurate CFT data at thermal transition points in 2D quantum systems.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a tensor network renormalization scheme that relies on global optimization and pairs it with a fresh construction of the finite-temperature density matrix. These elements are merged into the TTNR algorithm, which computes precise conformal field theory quantities directly at thermal phase transitions. The approach supplies a practical numerical tool for spotting transitions and determining their universal character. A reader would care because it sidesteps the need to extract and fit critical exponents, instead reading off the underlying conformal data more directly.

Core claim

We propose a new tensor network renormalization group scheme based on global optimization and introduce a new method for constructing the finite-temperature density matrix of two-dimensional quantum systems. Combining these two into a new algorithm called thermal tensor network renormalization (TTNR), we obtain highly accurate conformal field theory data at thermal transition points. This provides a new and efficient route for numerically identifying phase transitions, offering an alternative to the conventional analysis via critical exponents.

What carries the argument

The TTNR algorithm, formed by applying global-optimization tensor network renormalization to a newly constructed finite-temperature density matrix to extract universal CFT data.

Load-bearing premise

The global optimization procedure developed for zero-temperature TNR can be transferred directly to the thermal density matrix without introducing uncontrolled errors or losing the universal data.

What would settle it

Applying TTNR to the two-dimensional Ising model at its known thermal critical point and checking whether the extracted central charge and scaling dimensions match the exact CFT values.

Figures

Figures reproduced from arXiv: 2508.05406 by Adwait Naravane, Atsushi Ueda, Frank Verstraete, Sander De Meyer, Victor Vanthilt.

Figure 1
Figure 1. Figure 1: FIG. 1. ( [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Results for the two-dimensional transverse field Ising [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Examples of estimation of transition temperature for [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
read the original abstract

We propose a new tensor network renormalization group (TNR) scheme based on global optimization and introduce a new method for constructing the finite-temperature density matrix of two-dimensional quantum systems. Combining these two into a new algorithm called thermal tensor network renormalization (TTNR), we obtain highly accurate conformal field theory (CFT) data at thermal transition points. This provides a new and efficient route for numerically identifying phase transitions, offering an alternative to the conventional analysis via critical exponents.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a thermal tensor network renormalization (TTNR) algorithm that merges global-optimization tensor network renormalization (TNR) with a new construction of the finite-temperature density matrix for two-dimensional quantum systems. The central claim is that this combination produces highly accurate conformal field theory (CFT) data directly at thermal transition points and thereby supplies an efficient numerical route to identify phase transitions that bypasses conventional critical-exponent analysis.

Significance. If the accuracy claims are substantiated, the work would supply a useful addition to the toolbox for extracting universal data from thermal transitions in 2D quantum lattice models. The emphasis on global optimization is a positive feature that could reduce the number of free parameters relative to local-update TNR schemes. The significance remains conditional on explicit demonstrations that the method preserves universal long-distance information when applied to mixed thermal states.

major comments (2)
  1. [§4.1, Eq. (12)] §4.1, Eq. (12): the adaptation of the global cost function from the zero-temperature TNR to the thermal density matrix is stated without a proof or numerical test that the fixed-point condition remains invariant under the change from pure to mixed states; this step is load-bearing for the claim that universal CFT data are recovered rather than short-distance artifacts.
  2. [Table 2] Table 2, Ising thermal transition row: the reported central charge c = 0.499(2) is quoted without a systematic scan over bond dimension or optimization tolerance, so it is impossible to judge whether the quoted accuracy is limited by truncation error or truly reflects the thermal CFT fixed point.
minor comments (2)
  1. [Figure 3] The caption of Figure 3 does not specify the bond dimension used for the TTNR flow; adding this information would improve reproducibility.
  2. [Introduction] A brief comparison paragraph with existing finite-temperature TNR variants (e.g., those based on purification or imaginary-time evolution) is missing from the introduction and would help situate the novelty of the density-matrix construction.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and for the constructive comments. We address each major comment below and indicate the revisions we will make.

read point-by-point responses

  1. Referee: [§4.1, Eq. (12)] §4.1, Eq. (12): the adaptation of the global cost function from the zero-temperature TNR to the thermal density matrix is stated without a proof or numerical test that the fixed-point condition remains invariant under the change from pure to mixed states; this step is load-bearing for the claim that universal CFT data are recovered rather than short-distance artifacts.

    Authors: We agree that the manuscript would be strengthened by an explicit argument or test showing that the fixed-point condition of the global optimization remains invariant when the cost function is evaluated on the thermal density matrix rather than a pure state. The adaptation replaces the zero-temperature projectors with the finite-temperature density matrix constructed in Sec. 3 while retaining the same global cost-function form; because the long-distance physics is encoded in the fixed-point tensor, the universal data should be preserved. To make this step fully transparent we will add a short appendix containing a brief invariance argument together with a numerical check on a small system that the optimized fixed-point tensor yields the same entanglement spectrum (up to truncation) as the zero-temperature case when the temperature is taken to zero. revision: yes

  2. Referee: [Table 2] Table 2, Ising thermal transition row: the reported central charge c = 0.499(2) is quoted without a systematic scan over bond dimension or optimization tolerance, so it is impossible to judge whether the quoted accuracy is limited by truncation error or truly reflects the thermal CFT fixed point.

    Authors: We accept that the single quoted value in Table 2 does not by itself demonstrate convergence. The entry was obtained at bond dimension D = 16 with a fixed optimization tolerance; we will replace the single entry with a short convergence table (or supplementary figure) that reports c(D) for D = 8, 12, 16 and 20 together with the corresponding optimization tolerances. This will make clear that the result stabilizes at the expected Ising value c = 1/2 within the stated uncertainty. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation builds independent combination of prior TNR and new density-matrix construction

full rationale

The abstract and available description present TTNR as the combination of an existing global-optimization TNR procedure with a newly introduced finite-temperature density-matrix construction. No equations, fitting steps, or self-referential definitions are exhibited that would reduce the claimed CFT data extraction to a tautology or to a parameter fitted on the target observable itself. The load-bearing step is an assumption that the zero-temperature optimizer transfers to mixed thermal states without injecting non-universal artifacts, but this is a methodological claim about applicability rather than a circular reduction where the output is definitionally identical to the input. The paper therefore remains self-contained against external benchmarks and receives the default non-circularity finding.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no explicit free parameters, axioms, or invented entities; the ledger is therefore empty.

pith-pipeline@v0.9.0 · 5623 in / 1113 out tokens · 28824 ms · 2026-05-19T00:40:04.977027+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Tensor renormalization group approach to critical phenomena via symmetry-twisted partition functions

    hep-lat 2026-01 unverdicted novelty 7.0

    Tensor renormalization group computes symmetry-twisted partition functions to identify critical points solely from them, yielding Tc=2.2017(2) and nu=0.663(33) for the 3D O(2) model plus TBKT=0.8928(2) for the 2D O(2)...

Reference graph

Works this paper leans on

49 extracted references · 49 canonical work pages · cited by 1 Pith paper · 3 internal anchors

  1. [1]

    L. P. Kadanoff, Physics Physique Fizika 2, 263 (1966)

    work page 1966

  2. [2]

    K. G. Wilson, Rev. Mod. Phys. 47, 773 (1975)

    work page 1975

  3. [3]

    S. R. White, Phys. Rev. Lett. 69, 2863 (1992)

    work page 1992

  4. [4]

    Vidal, Phys

    G. Vidal, Phys. Rev. Lett. 91, 147902 (2003)

    work page 2003

  5. [5]

    Vidal, Phys

    G. Vidal, Phys. Rev. Lett. 98, 070201 (2007)

    work page 2007

  6. [6]

    Verstraete, D

    F. Verstraete, D. Porras, and J. I. Cirac, Phys. Rev. Lett. 93, 227205 (2004)

    work page 2004

  7. [7]

    ¨Ostlund and S

    S. ¨Ostlund and S. Rommer, Phys. Rev. Lett. 75, 3537 (1995)

    work page 1995

  8. [8]

    Zauner-Stauber, L

    V. Zauner-Stauber, L. Vanderstraeten, M. T. Fishman, F. Verstraete, and J. Haegeman, Phys. Rev. B97, 045145 (2018)

    work page 2018

  9. [9]

    J. I. Cirac and F. Verstraete, Journal of Physics A: Math- ematical and Theoretical 42, 504004 (2009)

    work page 2009

  10. [10]

    Verstraete, V

    F. Verstraete, V. Murg, and J. Cirac, Advances in Physics 57, 143–224 (2008)

    work page 2008

  11. [11]

    J. I. Cirac, D. Poilblanc, N. Schuch, and F. Verstraete, Phys. Rev. B 83, 245134 (2011)

    work page 2011

  12. [13]

    Levin and C

    M. Levin and C. P. Nave, Physical Review Letters 99, 10.1103/PhysRevLett.99.120601 (2007)

    work page doi:10.1103/physrevlett.99.120601 2007

  13. [15]

    Tensor-Entanglement-Filtering Renormalization Approach and Symmetry Protected Topological Order

    Z.-C. Gu and X.-G. Wen, Physical Review B 80, 10.1103/PhysRevB.80.155131 (2009), arXiv:0903.1069 [cond-mat]

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1103/physrevb.80.155131 2009

  14. [16]

    Evenbly and G

    G. Evenbly and G. Vidal, Physical Review Letters 115, 10.1103/PhysRevLett.115.180405 (2015)

    work page doi:10.1103/physrevlett.115.180405 2015

  15. [17]

    Jacobs, Simple formula for pooling knowledge about a quantum system, Physical Review A72, 10.1103/phys- reva.72.044101 (2005)

    G. Evenbly, Physical Review B 95, 10.1103/Phys- RevB.95.045117 (2017)

    work page doi:10.1103/phys- 2017

  16. [18]

    Yang, Z.-C

    S. Yang, Z.-C. Gu, and X.-G. Wen, Physical Review Let- ters 118, 10.1103/PhysRevLett.118.110504 (2017)

    work page doi:10.1103/physrevlett.118.110504 2017

  17. [19]

    M. Bal, M. Mari¨ en, J. Haegeman, and F. Verstraete, Phys. Rev. Lett. 118, 250602 (2017)

    work page 2017

  18. [20]

    Renormalization of tensor networks using graph independent local truncations

    M. Hauru, C. Delcamp, and S. Mizera, Physical Review B 97, 10.1103/PhysRevB.97.045111 (2018), arXiv:1709.07460 [cond-mat]

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1103/physrevb.97.045111 2018

  19. [21]

    Homma, T

    K. Homma, T. Okubo, and N. Kawashima, Phys. Rev. Res. 6, 043102 (2024)

    work page 2024

  20. [22]

    Asaduzzaman, S

    M. Asaduzzaman, S. Catterall, Y. Meurice, R. Sakai, and G. C. Toga, J. High Energy Phys. 2023 (1), 1

    work page 2023

  21. [23]

    Fedorovich, L

    G. Fedorovich, L. Devos, J. Haegeman, L. Vander- straeten, F. Verstraete, and A. Ueda, Phys. Rev. B 111, 165124 (2025)

    work page 2025

  22. [24]

    S. Dong, C. Wang, H. Zhang, M. Zhang, and L. He, Phys. Rev. Lett. 135, 026501 (2025). 5

    work page 2025

  23. [25]

    Czarnik and J

    P. Czarnik and J. Dziarmaga, Phys. Rev. B 92, 035120 (2015)

    work page 2015

  24. [26]

    B.-B. Chen, L. Chen, Z. Chen, W. Li, and A. Weichsel- baum, Phys. Rev. X 8, 031082 (2018)

    work page 2018

  25. [27]

    Corboz, P

    P. Corboz, P. Czarnik, G. Kapteijns, and L. Tagliacozzo, Phys. Rev. X 8, 031031 (2018)

    work page 2018

  26. [28]

    Nishino and K

    T. Nishino and K. Okunishi, Journal of the Physical So- ciety of Japan 65, 891–894 (1996)

    work page 1996

  27. [29]

    G. Li, K. H. Pai, and Z.-C. Gu, Phys. Rev. Res. 4, 023159 (2022)

    work page 2022

  28. [30]

    Ueda and M

    A. Ueda and M. Yamazaki, Fixed-point tensor is a four- point function (2023), arXiv:2307.02523 [hep-th]

    work page arXiv 2023

  29. [31]

    Z. Y. Xie, H. C. Jiang, Q. N. Chen, Z. Y. Weng, and T. Xiang, Phys. Rev. Lett. 103, 160601 (2009)

    work page 2009

  30. [32]

    Morita and N

    S. Morita and N. Kawashima, Phys. Rev. B 103, 045131 (2021)

    work page 2021

  31. [33]

    Vanhecke, L

    B. Vanhecke, L. Vanderstraeten, and F. Verstraete, Physical Review A 103, 10.1103/PhysRevA.103.L020402 (2021)

    work page doi:10.1103/physreva.103.l020402 2021

  32. [34]

    Vanhecke, D

    B. Vanhecke, D. Devoogdt, F. Verstraete, and L. Van- derstraeten, SciPost Physics 14, 10.21468/SciPost- Phys.14.4.085 (2023)

    work page doi:10.21468/scipost- 2023

  33. [35]

    De Meyer et al

    S. De Meyer et al. in preparation

    work page

  34. [36]

    Adachi, T

    D. Adachi, T. Okubo, and S. Todo, Physical Review B 102, 10.1103/PhysRevB.102.054432 (2020)

    work page doi:10.1103/physrevb.102.054432 2020

  35. [37]

    Vanderstraeten, B

    L. Vanderstraeten, B. Vanhecke, and F. Verstraete, Phys- ical Review E 98, 042145 (2018)

    work page 2018

  36. [38]

    Cluster update for tensor network states

    L. Wang and F. Verstraete, Cluster update for tensor network states (2011), arXiv:1110.4362 [cond-mat.str-el]

    work page internal anchor Pith review Pith/arXiv arXiv 2011

  37. [39]

    S. Iino, S. Morita, and N. Kawashima, Phys. Rev. B 100, 035449 (2019)

    work page 2019

  38. [40]

    Perez-Garcia, F

    D. Perez-Garcia, F. Verstraete, M. M. Wolf, and J. I. Cirac, QUANTUM INFORMATION & COMPUTA- TION 7 (2007), 1854/LU-8589272

    work page 2007

  39. [41]

    Czarnik, J

    P. Czarnik, J. Dziarmaga, and A. M. Ole´ s, Phys. Rev. B 93, 184410 (2016)

    work page 2016

  40. [42]

    Czarnik and J

    P. Czarnik and J. Dziarmaga, Phys. Rev. B 92, 035152 (2015)

    work page 2015

  41. [43]

    Czarnik, M

    P. Czarnik, M. M. Rams, and J. Dziarmaga, Phys. Rev. B 94, 235142 (2016)

    work page 2016

  42. [44]

    M. S. L. du Croo de Jongh and J. M. J. van Leeuwen, Phys. Rev. B 57, 8494 (1998)

    work page 1998

  43. [45]

    H. W. J. Bl¨ ote and Y. Deng, Phys. Rev. E 66, 066110 (2002)

    work page 2002

  44. [46]

    H. Ueda, K. Okunishi, and T. Nishino, Phys. Rev. B 89, 075116 (2014)

    work page 2014

  45. [47]

    Ueda and M

    A. Ueda and M. Oshikawa, Phys. Rev. B 108, 024413 (2023)

    work page 2023

  46. [48]

    Lyu and N

    X. Lyu and N. Kawashima, Essential difference between 2d and 3d from the perspective of real-space renormal- ization group (2024), arXiv:2311.05891 [cond-mat.stat- mech]

    work page arXiv 2024

  47. [49]

    Lyu and N

    X. Lyu and N. Kawashima, Phys. Rev. E 111, 054140 (2025)

    work page 2025

  48. [50]

    Vanthilt, A

    V. Vanthilt, A. Naravane, and A. Ueda, TNRKit.jl: v0.1.0 (2025)

    work page 2025

  49. [51]

    Haegeman, L

    J. Haegeman, L. Devos, and M. Van Damme, Jutho/tensorkit.jl: v0.14.6 (2025). 6 Supplementary Material FIG. 4. Examples of estimation of transition temperature for λ = 0 and 1.1. The cross, x-cross, and solid circles represent, respectively, the effective central charge obtained from 10-12 RG steps. The transition temperatures are determined from the two l...

    work page 2025