Dynamical Systems · arXiv:2611.16336 v1 · submitted 12 Nov 2026 · 14 references
Visibility graphs of irregularly sampled two-dimensional fields via Voronoi tessellations
The paper says
“Visibility graphs map a time series into a network by linking two data whenever the straight line between them is not obstructed by intermediate data.”
more of the abstract
Their extension to two-dimensional fields has so far been restricted to regular lattices, where visibility is checked along a finite number of lattice directions. We propose a definition for scalar data sampled at arbitrary, irregularly scattered points of the plane. Each point is assigned its Voronoi cell, so that the data become a landscape of prisms, and two points are linked when the line of sight between them clears the cells crossed by the connecting segment. This yields horizontal ($\HVGV$) and natural ($\NVGV$) Voronoi visibility graphs that reduce exactly to the classical horizontal and natural visibility graphs when the points are collinear, contain the Gabriel graph of the point set, and can be built by an exact walk through the Delaunay triangulation. For uncorrelated heights, the horizontal graph depends only on the ranks of the data, so its statistics are independent of the marginal distribution. The probability that two points see each other is 2/[(M + 1)(M + 2)], where M is the number of Voronoi cells crossed, which reduces the problem to the geometry of the tessellation. For Poisson points this gives a degree distribution with a scale-free tail P(k) ≃ (π3/4) k−2 and a mean degree that grows as (π3/8)lnN, in contrast with the finite value ⟨k⟩ = 4 of one-dimensional series. Two dimensions are therefore the marginal case in which isotropic horizontal visibility ceases to be local. Numerical simulations with up to N = 1.28 × 105 points support these predictions.
from the abstract
Pith says
Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded · not checked by a person
It can be wrong. Challenge this review or reply to it.
A clean extension of visibility graphs to irregularly sampled 2D fields, with honest mean-field predictions and a clear null model. The main result is plausible but the numerical evidence for the scale-free tail is still provisional. This line is wrong
the referee pass did not run
How often objections from reviews like this one are right: not yet measured
Not checked: figures, code, what the 14 references say.
What would settle it
Perform a high-statistics simulation for N = 10⁶ (or as large as feasible) using the exact height-averaged degree (Eq. 9) for central nodes, and fit ⟨k⟩ = a ln N + b over N = 10⁴, 10⁵, 10⁶. If the slope a is within 1% of π³/8 and the residuals are consistent with a pure logarithm, the mean-field prediction is strongly supported. If a deviates by more than 5% or a log-log fit of ⟨k⟩ vs N shows curvature, the asymptotic form may differ. Additionally, compute P(K ≥ k) for k up to 10³ and test whether the exponent is −1.00 ± 0.05 using a maximum-likelihood estimator with a lower cutoff k_min.
from the skeptic's reading
How the review went
3 readings by deepseek-v4-flash · 13 Nov 2026
- Reader holds up
- Skeptic agrees
- Editor wrote the verdict above
Not run on this paper: the referee, the simulated authors' defence.
Pith did not check this review's objections against the paper.
The review in full
The review in brief, machine-written
Hi [Colleague],
This paper does what it says: it defines visibility graphs for scattered 2D data using Voronoi cells, proves the reduction to 1D, shows the Gabriel graph is a backbone, and derives exact pair-visibility probabilities. The key new result is that for Poisson points with i.i.d. heights, the horizontal version (HVG-V) has a degree distribution with a k−2 tail and a mean degree that grows as (π³/8) ln N. That is qualitatively different from the 1D case (finite mean degree 4) and from lattice visibility graphs (exponential tails). The paper is well written, the definitions are natural, and the authors are upfront about the limitations.
What is actually new: the definition itself, the exact pair-visibility formula (Proposition 5), and the mean-field calculation for Poisson points. The paper also shows that HVG-V is distribution-free, which is a useful property for null-model testing. The numerical simulations are consistent with the predictions.
Where the soft spots are: the headline results for ⟨k⟩ and P(k) rest on a mean-field approximation that replaces the discrete sum over points by a continuum integral and neglects fluctuations of M at fixed r. The authors call this out explicitly (Section 4, 'Eqs. (12)–(14) neglect fluctuations of M at fixed r and are asymptotic; they should be read as mean-field predictions'). The numerical evidence is supportive but not yet decisive: the slope of ⟨k⟩ vs ln N is 3.66 ± 0.09, within 6% of π³/8 ≈ 3.88, but the fit uses N ≥ 2000 and the last four doublings give slopes between 3.1 and 4.0, a range that includes the prediction but also values 20% lower. The collapse of ⟨k|u⟩ onto (π³/4)/(1−u) is shown up to 1/(1−u) ≈ 10⁴, but the cutoff moves with N and the asymptotic regime may not be fully reached. The tail P(K ≥ k) is shown over about one decade of k (10–100) for the largest N, which is insufficient to confirm a power law with exponent −1. The concern is not that the prediction is wrong, but that the numerical evidence, while consistent, does not yet rule out a different asymptotic exponent or a logarithmic correction. Given the paper’s honesty about the mean-field nature, this is a limitation rather than an error, but it is the most load-bearing assumption because the headline result (scale-free tail, logarithmic mean degree) depends on it.
Who this is for: researchers working on visibility graphs, spatial network analysis, or statistical testing of spatial data. The paper provides a clean null model that could be useful for hypothesis testing.
Recommendation: this paper deserves a serious referee. The definitions are sound, the derivations are correct given the assumptions, and the limitations are honestly stated. The referee should focus on whether the numerical evidence is strong enough to support the claims as stated, or whether the authors should soften the language about the scale-free tail. I would accept peer review.
Best, [You]
The strongest objection
The central claim—that HVG-V on Poisson points has ⟨k⟩ ∼ (π³/8) ln N and P(k) ∼ (π³/4) k⁻²—rests on a mean-field calculation (Section 4) that replaces the discrete sum over points by a continuum integral and neglects fluctuations of M at fixed r. The authors explicitly call this a mean-field prediction and note that a rigorous derivation is open. The numerical tests (Section 5) are supportive but not yet conclusive: the slope of ⟨k⟩ vs ln N is 3.66 ± 0.09, within 6% of π³/8 ≈ 3.88, but the fit uses N ≥ 2000 and the last four doublings give slopes between 3.1 and 4.0, a range that includes the prediction but also values 20% lower. The collapse of ⟨k|u⟩ onto (π³/4)/(1−u) is shown up to 1/(1−u) ≈ 10⁴, but the cutoff moves with N and the asymptotic regime may not be fully reached. The tail P(K ≥ k) is shown over about one decade of k (10–100) for the largest N, which is insufficient to confirm a power law with exponent −1. The concern is not that the prediction is wrong, but that the numerical evidence, while consistent, does not yet rule out a different asymptotic exponent or a logarithmic correction. Given the paper’s honesty about the mean-field nature, this is a limitation rather than an error, but it is the most load-bearing assumption because the headline result (scale-free tail, logarithmic mean degree) depends on it.
The reader's weakest_assumption correctly identifies the mean-field approximation and the neglect of fluctuations as the soft spot. My analysis agrees that this is the most load-bearing concern. However, the paper is transparent about this limitation, lists it as an open problem, and provides numerical evidence that is consistent with the predictions. The concern does not undermine the paper's core contribution—the definition of HVG-V and NVG-V, the structural properties (Propositions 1–5), and the reduction of the problem to tessellation geometry—which are rigorous and independent of the Poisson-point asymptotics. The mean-field predictions are presented as plausible and tested, not as proven. Therefore the verdict should remain ACCEPT (UNCHANGED). The concrete test I propose would strengthen the evidence but is not required for the paper's validity as a scientific contribution. I agree with the reader's assessment that the paper is sound and the limitations are acknowledged.
Pith's reading of the paper (a machine's summary, not the authors')
The paper extends the concept of visibility graphs, originally used for time series, to data points scattered irregularly in the plane. Each point is assigned a Voronoi cell (the region of the plane closer to it than to any other point), and the whole cell is given the height of the point, creating a landscape of prisms. Two points are connected if the straight line between them stays above all the prisms it crosses. This yields two graph variants: horizontal (HVG-V) and natural (NVG-V). The construction reduces exactly to the classical 1D visibility graphs when points are collinear, and always contains the Gabriel graph (a standard proximity graph). For uncorrelated random heights, the probability that two points see each other depends only on the number of cells crossed, not on the height distribution. Using this, the authors derive that for points placed uniformly at random (Poisson process), the mean degree grows logarithmically with the number of points, and the degree distribution has a power-law tail P(k) ~ k^{-2}. This is qualitatively different from the 1D case, where the mean degree is constant (4) and the distribution decays exponentially. The paper provides numerical simulations supporting these predictions.
Core claim. For uncorrelated heights on Poisson points, the horizontal Voronoi visibility graph (HVG-V) has a degree distribution with a scale-free tail P(k) ~ (π^3/4) k^{-2} and a mean degree that grows as (π^3/8) ln N, in contrast to the finite mean degree of 4 in one dimension.
The premise it rests on. The derivation of the degree distribution and mean degree for Poisson points relies on a mean-field approximation that neglects fluctuations of the number of crossed cells M at fixed distance r, and replaces the discrete sum over points by a continuum integral. The authors state this explicitly (Section 4, 'Eqs. (12)–(14) neglect fluctuations of M at fixed r and are asymptotic; they should be read as mean-field predictions'). The numerical evidence supports the predictions, but a rigorous derivation is left as an open problem.
Check this review
Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.
How often reviews like this one are right
Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.
How often objections from reviews like this one are right: not yet measured
Reviews by people
This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.
The paper
Abstract
Visibility graphs map a time series into a network by linking two data whenever the straight line between them is not obstructed by intermediate data. Their extension to two-dimensional fields has so far been restricted to regular lattices, where visibility is checked along a finite number of lattice directions. We propose a definition for scalar data sampled at arbitrary, irregularly scattered points of the plane. Each point is assigned its Voronoi cell, so that the data become a landscape of prisms, and two points are linked when the line of sight between them clears the cells crossed by the connecting segment. This yields horizontal ($\HVGV$) and natural ($\NVGV$) Voronoi visibility graphs that reduce exactly to the classical horizontal and natural visibility graphs when the points are collinear, contain the Gabriel graph of the point set, and can be built by an exact walk through the Delaunay triangulation. For uncorrelated heights, the horizontal graph depends only on the ranks of the data, so its statistics are independent of the marginal distribution. The probability that two points see each other is 2/[(M + 1)(M + 2)], where M is the number of Voronoi cells crossed, which reduces the problem to the geometry of the tessellation. For Poisson points this gives a degree distribution with a scale-free tail P(k) ≃ (π3/4) k−2 and a mean degree that grows as (π3/8)lnN, in contrast with the finite value ⟨k⟩ = 4 of one-dimensional series. Two dimensions are therefore the marginal case in which isotropic horizontal visibility ceases to be local. Numerical simulations with up to N = 1.28 × 105 points support these predictions.
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.
abstract compiled · 2026-11-13
The authors' main claim, as the classifier picked it out 0.77
“Two dimensions are therefore the marginal case in which isotropic horizontal visibility ceases to be local.”
How the main claim is stated: asserted 1.00
What the abstract itself states, as probabilities: claims first 0.27 · states limitation 0.29 · human participants 0.01 · controlled comparison 0.11 · randomized assignment 0.01 · states quantitative result 0.99 · compares with prior results 0.95 · mentions code or data availability 0.02
The abstract, sentence by sentence
This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.
| Sentence | Role | The authors' own work | States the main conclusion | Strength |
|---|---|---|---|---|
| Visibility graphs map a time series into a network by linking two data whenever the straight line between them is not obstructed by intermediate data. | background 1.00 | 0.05 | 0.02 | asserted 1.00 |
| Their extension to two-dimensional fields has so far been restricted to regular lattices, where visibility is checked along a finite number of lattice directions. | gap 0.84 | 0.03 | 0.02 | asserted 1.00 |
| We propose a definition for scalar data sampled at arbitrary, irregularly scattered points of the plane. | contribution 0.98 | 0.86 | 0.10 | asserted 0.98 |
| Each point is assigned its Voronoi cell, so that the data become a landscape of prisms, and two points are linked when the line of sight between them clears the cells crossed by the connecting segment. | method 0.93 | 0.87 | 0.14 | asserted 1.00 |
| This yields horizontal ($\HVGV$) and natural ($\NVGV$) Voronoi visibility graphs that reduce exactly to the classical horizontal and natural visibility graphs when the points are collinear, contain the Gabriel graph of the point set, and can be built by an exact walk through the Delaunay triangulation. | method 0.48 | 0.85 | 0.21 | asserted 1.00 |
| For uncorrelated heights, the horizontal graph depends only on the ranks of the data, so its statistics are independent of the marginal distribution. | result 0.98 | 0.79 | 0.06 | asserted 1.00 |
| The probability that two points see each other is $2/[(M+1)(M+2)]$, where $M$ is the number of Voronoi cells crossed, which reduces the problem to the geometry of the tessellation. | result 0.96 | 0.88 | 0.11 | asserted 1.00 |
| For Poisson points this gives a degree distribution with a scale-free tail $P(k)\simeq(\pi^3/4)\,k^{-2}$ and a mean degree that grows as $(\pi^3/8)\ln N$, in contrast with the finite value $\langle k\rangle=4$ of one-dimensional series. | result 1.00 | 0.88 | 0.25 | asserted 1.00 |
| Two dimensions are therefore the marginal case in which isotropic horizontal visibility ceases to be local. | interpretation 1.00 | 0.63 | 0.34 | asserted 1.00 |
| Numerical simulations with up to $N=1.28\times10^5$ points support these predictions. | result 0.61 | 0.96 | 0.06 | asserted 0.97 |
What kind of paper the classifier reports this is: method or tool 0.95 · evidence from simulation or computation 0.54
What was examined, and what was not
Sources read
- arXiv listing abstract v1
Not examined
- full text
- figures and tables
- references and citing sentences
- relations to other papers
Levels run
- P abstract
Not run
- S sections
- T body sentences
- C citations
- R cross-paper relations
How the labels were made
Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-11-13: 1 request, 7,825 input tokens.
Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON
Figures (5)
Figures are the authors', reproduced from the paper.
References (14)
- [1] Lucas Lacasa, Bartolo Luque, Fernando Ballesteros, Jordi Luque, and Juan Carlos Nuño. From time series to complex networks: The visibility graph.Proceedings of the National Academy of Sciences, 105(13):4972–4975, 2008. doi: 10.1073/pnas.0709247105 on Pithdoi:10.1073/pnas.0709247105
- [2] Bartolo Luque, Lucas Lacasa, Fernando Ballesteros, and Jordi Luque. Horizontal visibility graphs: Exact results for random time series.Physical Review E, 80:046103, 2009. doi: 10.1103/PhysRevE.80.046103 on Pithdoi:10.1103/physreve.80.046103
- [3] Lucas Lacasa, Vincenzo Nicosia, and Vito Latora. Network structure of multivariate time series.Scientific Reports, 5:15508, 2015. doi: 10.1038/srep15508 on Pithdoi:10.1038/srep15508
- [4] Juan Carlos Nuño and Francisco J. Muñoz. The partial visibility curve of the Feigenbaum cascade to chaos.Chaos, Solitons & Fractals, 131:109537, 2020. ISSN 0960-0779. doi: 10.1016/j.chaos.2019.109537. URL https://www.sciencedirect.com/science/article/ pii/S0960077919304886 on Pithdoi:10.1016/j.chaos.2019.109537
- [5] Lucas Lacasa and Jacopo Iacovacci. Visibility graphs of random scalar fields and spatial data.Physical Review E, 96:012318, 2017. doi: 10.1103/PhysRevE.96.012318 on Pithdoi:10.1103/physreve.96.012318
- [6] Jacopo Iacovacci and Lucas Lacasa. Visibility graphs for image processing.IEEE Transac- tions on Pattern Analysis and Machine Intelligence, 42(4), 2020. doi: 10.1109/TPAMI.2019. 2891742 on Pithdoi:10.1109/tpami.2019
- [7] Leila De Floriani and Paola Magillo. Intervisibility on terrains. In Paul A. Longley, Michael F. Goodchild, David J. Maguire, and David W. Rhind, editors,Geographical Information Systems: Principles, Techniques, Management and Applications, pages 543–556. Wiley, 1999. 9 on Pith
- [8] Ferran Hurtado, Maarten Löffler, Inês Matos, Vera Sacristán, Maria Saumell, Rodrigo I. Silveira, and Frank Staals. Terrain visibility with multiple viewpoints.arXiv preprint arXiv:1309.4323, 2013 on PithPith's reviewarXiv
- [9] Vincent Froese and Malte Renken. Advancing through terrains.arXiv preprint arXiv:1904.08746, 2019 on PithPith's reviewarXiv
- [10] Atsuyuki Okabe, Barry Boots, Kokichi Sugihara, and Sung Nok Chiu.Spatial Tessellations: Concepts and Applications of Voronoi Diagrams. Wiley, Chichester, 2nd edition, 2000 on Pith
- [11] Olivier Devillers, Sylvain Pion, and Monique Teillaud. Walking in a triangulation.Interna- tional Journal of Foundations of Computer Science, 13(2):181–199, 2002 on Pith
- [12] K. Ruben Gabriel and Robert R. Sokal. A new statistical approach to geographic variation analysis.Systematic Zoology, 18(3):259–278, 1969 on Pith
- [13] David W. Matula and Robert R. Sokal. Properties of Gabriel graphs relevant to geographic variation research and the clustering of points in the plane.Geographical Analysis, 12(3): 205–222, 1980 on Pith
- [14] Jesper Møller.Lectures on Random Voronoi Tessellations, volume 87 ofLecture Notes in Statistics. Springer, New York, 1994. 10 on Pith
Challenges and discussion
Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.
Come back by email
A short list of the week's newly reviewed papers. Nothing is sent until you confirm from your inbox, and every email has a link that stops it. No account.
No comments yet.
Sign in to comment