Entanglement Entropy in Quantum Networks with Tunable Geometry
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.
Kai: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We're looking at the paper: "Entanglement Entropy in Quantum Networks with Tunable Geometry".
Mira: Here's the gist of the paper. Entanglement entropy in quantum networks with tunable geometry is investigated by studying how graph geometry affects transport and entanglement growth The investigation identifies a new localization regime robust in both single-particle and finite-density cases by tuning the effective range parameter of a random graph model This provides a model of Anderson localization induced by structural disorder with implications for amorphous materials and tunable-range analog quantum simulators.
Kai: Model Description The study considers a random graph model defined by two parameters, c in
0, 1: and α in
0, +∞), where the probability that nodes j and l are linked is given by c/j - lα The Hamiltonian studied consists only of hopping terms along the edges of the graphs with nodes corresponding to qubits This Hamiltonian is rewritten as a random hopping Hamiltonian for hard-core bosons H = J X j l A j l b† j bl. [Mira: Numerical Methods The research employs Monte Carlo sampling to sample over the graph distribution and uses Exact Diagonalization (ED) in the single-particle case to determine the dynamics For the half-filling case, a replacement of hard-core boson operators with spinless fermions is used, leading to a quadratic Hamiltonian that is non-interacting.
Kai: Entanglement entropy S = −tr
ρA log ρA: is computed for the subsystem A =
1, N/2: ∩ Z.
Paper summary: Kai: So, we've touched on the setup; now let’s look at what the paper actually claims about their findings regarding the entanglement entropy in Quantum Networks with Tunable Geometry.
Mira: The core thesis seems to be that by tuning those parameters c and alpha, they can interpolate smoothly between different graph structures—like nearest-neighbor connections and fully connected graphs—and this allows them to pinpoint a specific region where the system exhibits localization behavior.
Lev: So, when you tune it, you're essentially mapping out a phase diagram in the space of c and alpha that shows where things localize versus where they delocalize.
Kai: Right, and they specifically identify an unexpected localized regime when they look at the case where c equals one but alpha is at some intermediate value, while recovering known behaviors when you go to the nearest-neighbor or all-to-all limits.
Mira: That minimum in the asymptotic entanglement entropy, which they call S infinity, seems to be directly linked to how nodes are distributed based on their graph distance from a central point N over N.
Kai: They connect the statistical property of those nodes as a function of that distance, f of delta, with the probability p of finding a particle at that specific graph distance delta.
Lev: That’s interesting because it means we aren't just looking at the total entanglement; we’re looking at how the spatial arrangement dictates that entanglement saturation point.
Mira: And for the half-filling case, they found that this asymptotic entanglement entropy reaches a maximum value of S infinity over N log two which is about zero point twenty-eight.
Kai: That number is compatible with what we expect in a random Gaussian state in the thermodynamic limit, but the main point is that the localized regime for those intermediate values of alpha seems robust and doesn't depend on where the particles start.
Lev: The robustness part is important because it suggests this structural disorder localization isn't just a fluke based on initial conditions; it’s built into the geometry itself.
Conclusion: Kai: So, wrapping up the Entanglement Entropy in Quantum Networks with Tunable Geometry paper by Azzali and Kuriyattil and Daley and Chiofalo, they've shown that tuning graph parameters c and alpha lets you create a new kind of localization regime.
Mira: What that means in plain terms is that the geometry of the connections can be tuned to create specific disorder effects, like Anderson localization, which is robust across different filling levels.
Lev: It suggests that in systems where you can engineer the connectivity—like an analog quantum simulator—the structure itself dictates whether information stays localized or spreads out.
Kai: It gives us a model for how structural disorder in amorphous materials could affect quantum transport, and it shows that this localization doesn't depend on the system size N at all for the typical length scale of the wavefunction.
Mira: The paper points toward future work looking at other moments of the probability density beyond just the inverse participation ratio to get a better picture of multifractal features in that non-ergodic regime.
Lev: And also, they suggest using this simple model to probe how things might change if we combine this tunable geometry with noise engineering in systems that aren't strictly Markovian.
Andrea Azzali, Sridevi Kuriyattil, Andrew J. Daley, Marilûlu Chiofalo
Technical University of Munich School of Natural Sciences Physics Department · Munich Center for Quantum Science and Technology Clarendon Laboratory University of Oxford INFN Sezione di Pisa
quant-ph
Submitted: 2026-07-17
Updated: 2026-11-09
Comments: 5 pages, 4 figures + Supplement; Improved main text and extended Supplement
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 87/100
The gist: Entanglement entropy in quantum networks with tunable geometry is investigated by studying how graph geometry affects transport and entanglement growth The investigation identifies a new localization…
Key concepts
- Random Graph Model
- This mathematical structure defines the quantum network where nodes (qubits) are connected randomly based on two parameters, c and alpha. The connection probability between any two nodes depends on their distance and these tuning parameters, creating a specific type of structural disorder in the system.
- Entanglement Entropy (S)
- This measures the amount of quantum correlation or entanglement present in a specific part of the network (subsystem A). It quantifies how complex the quantum state is within that region. The study tracks how this value changes as particles spread across the graph.
- Localization Regime
- This is a state where quantum particles become trapped in specific areas of the network due to structural disorder, preventing them from spreading freely. The research identifies a new, robust localization regime for intermediate values of alpha, which is independent of where the particles start.
- Geometric Interpretation
- The analysis links statistical properties of nodes (like their distribution as a function of graph distance) to the physical spreading length-scale of the quantum wavefunction. This connects abstract graph theory to how quantum information propagates in physical space.
Terminology
Summary
Entanglement entropy in quantum networks with tunable geometry is investigated by studying how graph geometry affects transport and entanglement growth The investigation identifies a new localization regime robust in both single-particle and finite-density cases by tuning the effective range parameter of a random graph model This provides a model of Anderson localization induced by structural disorder with implications for amorphous materials and tunable-range analog quantum simulators. p. 0
Model Description
The study considers a random graph model defined by two parameters, c in [0, 1] and α in [0, +∞), where the probability that nodes j and l are linked is given by c/j - lα The Hamiltonian studied consists only of hopping terms along the edges of the graphs with nodes corresponding to qubits This Hamiltonian is rewritten as a random hopping Hamiltonian for hard-core bosons H = J X j l A j l b† j bl.
Numerical Methods
The research employs Monte Carlo sampling to sample over the graph distribution and uses Exact Diagonalization (ED) in the single-particle case to determine the dynamics For the half-filling case, a replacement of hard-core boson operators with spinless fermions is used, leading to a quadratic Hamiltonian that is non-interacting. Entanglement entropy S = −tr[ρA log ρA] is computed for the subsystem A = [1, N/2] ∩ Z. p. 1
Key Results on Asymptotic Behavior
The asymptotic entanglement entropy displays an increase from the initial value of 0, eventually saturating to a value denoted as S∞ The behavior at c = 1 shows an unexpected localized regime at intermediate values of α, while recovering known behaviors at nearest-neighbors and all-to-all limits. This minimum in S∞ is connected to the graph geometry by analyzing the distribution of nodes as a function of the graph distance dG. p. 2
Geometric Interpretation
The analysis connects the statistical property f(δ) of nodes as a function of graph distance δ from node N/2 with the probability p(δ) for the particle to lie at a graph distance δ. The spreading length-scale in Euclidean geometry interpolates between regimes, and the behavior is also observed in the asymptotic spreading of the wavefunction. p. 3 In the large-α limit at c = 1, an approximate model treats r = 2 links as impurities to derive a prediction for S∞(α) that matches numerical data. p. 3
Finite Density Observations
For half-filling, the asymptotic entanglement entropy reaches a maximum value of S∞/(N log 2) ≈ 0.28 which is compatible with the expected value in the thermodynamic limit for a random Gaussian state. p. 3 The results suggest that the localized regime for intermediate values of α seems to be robust and independent of the initial positions of the particles. p. 8
Outlook
Further insight can be gained by computing other moments of the probability density beyond the Inverse Participation Ratio to characterize multifractal features in the non-ergodic regime Another direction involves replacing the minimal model with random graphs tailored to describe specific amorphous materials to probe the robustness of this unexpected localization regime experimentally. The simplicity of the model also paves the way to investigate how outcomes change when tunable geometry is combined with noise-engineering in Markovian and non-Markovian regimes. p. 4 Alternatively, one could consider nonreciprocal hopping which may support the presence of topological invariants. p. 4 The study also addresses the N dependence by showing that the typical length-scale of the localized wavefunction only depends on the density of impurities, hence on α, and is independent of N. p. 3 The results are left untouched by Kac normalization for α > 1. p. 9 The study shows that the delocalized regime for small values of α may be reached within times sub-extensive in the system size. p. 8 The unexpected localization and subsequent minimum in the entanglement entropy for intermediate values of α are left untouched by Kac normalization. p. 9
Acknowledgments
A.A. and M.L.C. acknowledge useful discussion with Giuseppe La Rocca A. p. 4A. acknowledges support by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy – EXC-2111 – 390814868 M. p. 4L.C. acknowledges support from the National Centre on HPC, Big Data and Quantum Computing—SPOKE 10 (Quantum Computing) and received funding from the European Union Next-GenerationEU—National Recovery and Resilience Plan (NRRP)—MISSION 4 COMPONENT 2, INVESTMENT N. p. 4 1.4—CUP N. I53C22000690001 This research has received funding from the European Union’s Digital Europe Programme DIGIQ under Grant Agreement No. p. 4 101084035. This work was supported by the EPSRC through the QQQS programme grant (EP/Y01510X/1) and by the QCi3 hub (EP/Z53318X/1). p. 4 The paper includes numerous references, such as [4] A. p. 1 S. Buyskikh, M. p. 4 Fagotti, J. p. 4 Schachenmayer, F. p. 4 Essler, and A. p. 4 J. Daley and [55] M. p. 0 Kac, G. E. Uhlenbeck, and P. p. 5 C Hemmer. The paper also references work on Anderson localization [40] P. p. 0 W Anderson, Physical Review 109, 1492 (1958). p. 5 The paper also references work on quantum percolation [30] S. p. 3 Kirkpatrick and T. p. 4 P. Eggarter, Physical Review B 6, 3598 (1972). p. 4 The paper also references work on many-body localization [22] K. p. 0 Tikhonov and A. p. 4 Mirlin, Annals of Physics 435, 168525 (2021). p. 4 The paper also references work on glassy dynamics [23] A. p. 0 Lunkin et al. p. 4, Hilbert space signatures of non-ergodic glassy dynamics (2026) arXiv:2601. p. 401309. The paper also references work on amorphous materials [24] C. p. 0-T. Toh, H. Zhang, J. Lin, A. S. Mayorov, Y. p. 4-P Wang, C M Orofeo. p. 5 The paper also references work on Quantum Reservoir Computing [26] K Fujii and K Nakajima, Physical Review Applied 8, 024030 (2607. p. 416394>. The paper also references work on entanglement dynamics after quantum quenches in generic integrable systems [54] V. p. 5 Alba and P. p. 5 Calabrese, Entanglement dynamics after quantum quenches in generic integrable systems (2018) arXiv:1712. p. 507529. The paper also references work on entanglement entropy in the thermodynamic limit [43] E. p. 8 Bianchi, L. p. 5
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Incorporate graph geometry into localization modeling to predict quantum dynamics on tunable-range systems. The improved AI can model quantum hopping on random graphs with power-law scaling of link probabilities to identify new localization regimes robust in both single-particle and finite-density cases, providing a model for Anderson localization induced by structural disorder.
-
Develop a predictive tool for entanglement entropy saturation based on graph parameters. The system can predict the asymptotic entanglement entropy S infinity as a function of the power-law exponent alpha and connectivity parameter c, distinguishing between localized and delocalized regimes in tunable quantum networks.
-
Enhance simulation capabilities by using graph distance metrics to characterize wavefunction spreading. The AI can analyze statistical properties like f(delta) and p(delta) to relate the graph geometry (shortest path length) directly to the observed localization behavior of the asymptotic wavefunction, moving beyond Euclidean geometry approximations.
-
Create a robust classification system for quantum phase transitions in random graphs. The improved system can determine whether a given graph configuration, defined by parameters c and alpha, falls into a percolating or non-percolating regime using classical phase diagrams as a guide for quantum effects.
Sources
- Entanglement growth and correlation spreading with variable-range interactions in spin and fermionic tunnelling models
- Fast scrambling on sparse graphs
- Complex Quantum Networks: a Topical Review
- Multifractal and Glassy Signatures of Non-Ergodic 2D Quantum Dynamics
- The crossover region between long-range and short-range interactions for the critical exponents
- Complex networks with tuneable dimensions as a universality playground
- Evolution of Entanglement Entropy in One-Dimensional Systems
- Entanglement dynamics after quantum quenches in generic integrable systems
Related papers
- Computational regimes in matrix-product-state-based quantum trajectory simulations
- Toward chemically accurate quantum simulations on quantum hardware: A quantum-HPC hybrid approach
- Transversal surface-code game with reconfigurable qubits
- Fulqrum: a generalized framework for quantum subspace diagonalization
- Flux-Activated Resonant Control of a Bosonic Quantum Memory
- Machine learning the effects of many quantum measurements