Nanowire networks' interconnection graphs from their photomicrographs

Cynthia P. Quinteros; Eduardo D. Mart\'inez; Javier Tau Anzo\'ategui; Juan Ignacio Diaz Schneider; Oscar Filevich; Pablo Levy

REVIEW 3 major objections 4 minor 24 references

Nanowire networks' interconnection graphs from their photomicrographs

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A post-processing pipeline on dark-field photomicrographs extracts the interconnection graph of silver nanowire networks, whose degree distributions and community structure resemble small-world and modular topologies.

desk verdict An honest and useful methods paper whose central topology claim is currently underdetermined by an unspecified crossing-to-edge rule; worth refereeing with serious revision. read the letter →

arxiv 2607.16445 v1 pith:5K4SEB6R submitted 2026-07-17 cond-mat.dis-nn cond-mat.mes-hallmath-phmath.MP

Javier Tau Anzoátegui , Juan Ignacio Diaz Schneider , Eduardo D. Martínez , Pablo Levy , Oscar Filevich , Cynthia P. Quinteros This is my paper · ORCID

classification cond-mat.dis-nncond-mat.mes-hallmath-phmath.MP

keywords nanowirenetworksgraphextractionimageprocessingskeletonizationsmall-worldmodularclusteringcoefficientpathlength

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that standard dark-field photomicrographs of silver nanowire networks contain enough information to reconstruct the network's interconnection graph: after identifying electrodes, binarizing, cleaning, and skeletonizing the image, wire segments become nodes and their intersections become edges. Applying this pipeline to three real samples yields graphs with a moderate mean degree (3.2–3.65), degree distributions similar to those of small-world networks, and adjacency matrices showing modular community structure. The authors also test the effect of false junctions — crossings that are not true contacts — by randomly removing edges, and show that clustering coefficient drops and path length rises until the graph disconnects. The motivation is to turn experimental devices into simulation-ready graphs, so that simulated electrical responses can be compared with real measurements rather than with synthetic network models.

What carries the argument

The load-bearing mechanism is the image-processing pipeline: after identifying the macroscopic electrodes, the RGB image is binarized, cleaned of isolated noise and holes, and skeletonized to one-pixel-wide lines. The intersections of those lines are read as graph edges and the intervening segments as nodes, converting the physical sample into an abstract graph suitable for network analysis. The paper then computes the degree distribution, adjacency matrix, community sizes, clustering coefficient (average fraction of triangles through each node), and path length (mean shortest-path distance among a sampled subset of the largest connected component). The false-junction problem — a top-down im

What would settle it

Re-extract the graph from one sample using two extreme crossing rules: pairing the two collinear segments of the same wire (pass-through, no junction) versus pairing one perpendicular pair (junction). If the degree distribution and clustering coefficient shift enough to change the small-world or modular classification, then the paper's central topological claim is an artifact of the chosen (unstated) rule rather than a property of the physical network.

Watch

Extended reading notes

Core claim

The central claim is that a pipeline of electrode detection, binarization, cleaning, and skeletonization can convert a zenithal-view optical image of a silver nanowire network into a graph whose nodes are the nanowire segments and whose edges are the crossings. On graphs extracted from three samples (A, B, C), the average degree is moderate (3.65, 3.60, 3.20), the degree distribution resembles that of the small-world network model, and the adjacency matrices show well-defined communities with sparse inter-community links, a modular arrangement. The paper further shows that randomly removing edges with probability p_er lowers the clustering coefficient and raises the path length, and that def

Load-bearing premise

The pipeline assumes an unambiguous rule for turning each skeletonized crossing — where four wire ends meet — into graph edges, but the rule is never stated; the degree distribution that drives the small-world classification is set entirely by that choice.

Editorial extensions

If this is right

Graphs extracted from real devices can replace synthetic topologies as inputs to network simulators, making simulated responses directly comparable to measured electrical behavior.
The moderate mean degree and modular adjacency structure suggest silver nanowire assemblies sit in a regime favorable to reservoir computing — dense local clusters with sparse long-range links.
The monotonic response of clustering and path length to edge removal links the unknown true junction density to observable graph metrics; once true densities are known (via higher-resolution imaging), the metrics' realistic values can be read off the curves.
The electrode-dependent path length implies that device-level metrics depend on the contact geometry, so comparing different experimental devices requires reporting electrode shape and placement alongside network density.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The unstated crossing-to-edge rule is a pivotal implementation choice: pairing collinear segments (pass-through) versus perpendicular segments (junction) changes every downstream metric, so the small-world classification should be re-run under both pairings.
An immediate extension is to correlate the effective-electrode path length with two-terminal resistance measurements on the same samples, testing whether the graph metric predicts transport.
The same pipeline could be applied to nanotube and nanoparticle percolation networks; if those also show small-world-plus-modular signatures, the topology may be generic to self-assembled conductors rather than specific to silver nanowires.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know. First, this is a straightforward, honest methods paper: the pipeline is clearly described at the image-processing level, the p_er sweep is a sensible way to probe the false-junction problem, and the electrode-constrained path-length analysis is a real idea that could matter for simulation. Second, the graph-construction step is the weak load-bearing point. The authors define segments as nodes and intersections as edges, but never say how a detected crossing becomes an edge. A crossing has four incident segment-ends. The reported mean degree 2E/N ≈ 3.65 implies each crossing contributes roughly 3.65 edges on average under the natural segment-count interpretation, or else the node definition is something other than what the text says. Without that rule, Fig. 4's degree distributions, adjacency matrices, and community sizes are not reproducible, and the small-world/modular classification is underdetermined.

The paper does some things well. It openly acknowledges the zenithal-view overcounting and does not pretend the removal probability is a fix. The distinction among base graph, effective graph, and effective electrode graph in the path-length analysis is thoughtful, and the result that electrode geometry shifts measured path length is likely useful to the AgNWN simulation community. The writing is clear, and the limitations are stated plainly.

The soft spots beyond the crossing rule are real but mostly addressable. There is no null-model comparison: the small-world-like and modular-like labels come from visual comparison with hand-picked reference networks, not from quantitative baselines. Extraction is not validated against known ground truth, and no code or data are provided, so the pipeline cannot be checked or reused. The reference to Milano et al.'s connectome paper appears in the reference list but is never cited in the text, which is odd because that is the closest prior work. None of these are fatal; they are gaps a revision could close.

Who is this for? Experimental groups working on silver nanowire or similar self-assembled networks who want to move from images to simulation-ready graphs. The paper deserves a serious referee—the core idea is useful and the authors are not overselling the physical realism—but it needs a major revision that states the crossing-to-edge rule, validates or at least calibrates the extraction, and adds a baseline comparison. As it stands, the topology conclusions should be read as preliminary.

Referee Report

3 major / 4 minor

Summary. The manuscript proposes an image-processing pipeline that converts stitched, zenithal dark-field photomicrographs of silver nanowire networks (AgNWNs) into an abstract graph, with nanowire segments coded as nodes and intersections as edges. For three experimental samples, the authors report basic graph attributes (N, e, average degree) and topological metrics (degree distribution, clustering coefficient, path length, community size distribution), comparing them against small-world, modular, and scale-free reference networks of size N=70000 and mean degree k=4. Because the optical images cannot distinguish true junctions from mere cross-points of vertically separated wires, the authors also study the effect of randomly removing edges with probability p_er on clustering coefficient and path length. The central claim is that the extracted graphs reveal small-world-like degree distributions and modular community structure, while the electrode geometry affects the measured path length.

Significance. If the pipeline is correctly specified, it would provide a practical bridge between optical characterization of nanowire networks and simulation-ready graph models, and the comparison with archetypal network topologies would be of interest to the neuromorphic nanowire-network community. The paper is honest about the zenithal-view limitation, and the p_er sweep is a useful sensitivity check. At the same time, the central graph-construction step is currently under-specified, and the topological classification rests mainly on visual comparison. The approach has clear value, but the present manuscript does not yet make the claimed results reproducible or quantitatively supported.

major comments (3)

[Pipeline for graph generation] The crossing-to-edge rule is not defined. The text states that 'segments will be coded as nodes and the junctions as edges' and Table 1 reports e=121217 for N=66435, giving 2e/N=3.65, exactly the quoted average degree. Thus each crossing contributes one edge. But a skeleton crossing of two nanowires has four incident segment ends; choosing which pair of segment-nodes is joined by that single edge (collinear pass-through, perpendicular, or an arbitrary pair) is never stated. This choice directly fixes the degree distribution, clustering coefficient, and path length — the quantities used for the small-world/modular classification. Moreover, under the stated segment-as-node definition, a segment bounded by two crossings has at most two incident crossings, so the average degree should be near 2, not 3.65, unless multiple edges per crossing are allowed or the node definition is different. Ple
[Topological analysis, Fig. 4] The claim that the experimental graphs are 'small-world-like' and 'modular' is based on visual inspection of degree histograms, adjacency matrices, and community-size distributions. No quantitative comparison with the reference networks is provided, and the reference networks are generated with N=70000, k=4 while the experimental samples have N between 26222 and 77477 and mean degree between 3.20 and 3.65, so finite-size and mean-degree effects are uncontrolled. Please provide quantitative measures (e.g., degree-distribution distance, clustering coefficient relative to Erdős–Rényi graphs of the same N and k, modularity Q, or a small-world index) and state the community-detection algorithm and its parameters, which are needed to reproduce the community-size distributions in Fig. 4(g)-(i). Without this, the central topological conclusion is not established.
[Topological analysis, Fig. 5] The p_er sweep is explicitly an arbitrary robustness probe, not a calibrated noise model. The headline conclusion — similarity to small-world and modular topologies — is computed on the unperturbed graph, which the paper itself says overestimates the number of true junctions. The figures show how clustering coefficient and path length evolve with p_er, but not how the degree distribution or community structure (the basis of the topological classification) change. If the small-world/modular resemblance disappears at a plausible p_er, the conclusion should be restricted to the 'all-crossings-are-junctions' idealization; if it persists, that should be shown. Please report the relevant topology metrics as a function of p_er or explicitly limit the claim.

minor comments (4)

[Pipeline for graph generation] The text says the graph is modified by 'scanning every node and applying a previously defined p_er'; since the object being removed is an edge, the description should refer to edges. Also, 'the maximum removal probability can be thought of as the inverse of the graph's minimum density' is unclear; the intended relation to the percolation threshold should be stated precisely.
[Topological analysis, Eq. (2)] The path-length estimator uses m=1000 seed nodes, but no confidence intervals or standard errors are reported. Calling it 'a robust estimator' requires some quantification. Also, when the graph becomes disconnected under p_er, the text discards unreachable pairs; the exact rule for which components/pairs are included should be stated.
[Topological analysis, Fig. 5] Fig. 5(e) shows path length only for Sample A. If Samples B and C exhibit the same behavior, a sentence stating that would help; otherwise the representative nature of the figure is unclear.
[Abstract/Conclusions] The phrase 'hints of the three types could be found' in the conclusions is vague. If scale-free characteristics are present, they should be identified explicitly; if not, the sentence should be removed or qualified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the graph extraction and topology analyses are not derived from fitted parameters or self-referential definitions.

full rationale

No circularity found. The pipeline converts photomicrographs into graphs through image processing steps (binarization, cleaning, skeletonization, intersection detection) and reports the resulting graph features in Table 1. The reference networks in Fig. 3 are explicitly illustrative, generated with N = 70000 and k = 4 as a comparable size, not fitted to reproduce the experimental data. The edge-removal probability p_er is introduced as an arbitrary robustness sweep to model the known zenithal-view ambiguity, and the paper candidly states that this overestimates the number of interconnections; it is not a fitted parameter renamed as a prediction. The self-citations ([9], [18]) concern synthesis and prior characterization methods and are not load-bearing for the graph-extraction or topology-classification claims. The skeptical concern about the unspecified crossing-to-edge pairing rule is a real underspecification or validation gap, but it is not circular reasoning: no equation or construction in the paper defines the conclusion in terms of the input. The paper's own stated limitations about false junctions and the need for further experimental improvement are consistent with an honest, non-circular report.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

No new physical entities (particles, forces, dimensions) are introduced. p_er is an analysis parameter, and 'effective graph'/'effective electrode graph' are operational re-definitions of the graph, not new physical postulates.

free parameters (4)

p_er — edge-removal probability for false junctions = swept 0 → near percolation (Fig. 5)
Arbitrary uniform removal probability introduced to model zenithal over-counting of junctions; uncalibrated against any 3D measurement (Section 'Topological analysis', Fig. 5).
Image-processing hyperparameters (binarization threshold, cleaning window, skeletonization details)
Chosen by hand and never specified in the text ('Pipeline for graph generation'); they determine the node/edge counts of Table 1.
m — seed-node sample size for path-length estimation = 1000
Computational approximation parameter for the all-pairs shortest-path mean (Eq. 2); estimator precision and bias are not reported (Section 'Topological analysis').
Reference-network size and mean degree = N = 70000, k ≈ 4
Archetype networks generated 'comparable to' the experimental graphs (Fig. 3); hand-chosen for visualization, not fit to data.

assumptions (5)

domain assumption A zenithal 2D projection of the assembly, after skeletonization, is a valid graph abstraction of the physical interconnection scheme (nodes = wire segments, edges = crossings).
Central modeling premise of the pipeline; the paper itself flags that it overcounts junctions (abstract; Section 'Pipeline for graph generation').
domain assumption Wire segments are passive conductors and crossings are the only electrically active components, so the segments-as-nodes graph is the right electrical abstraction.
Justifies the node/edge duality in 'Pipeline for graph generation'; rests on the memristive-junction picture of AgNWNs from Refs. [9,14,16].
standard math Standard graph-theoretic definitions: local clustering coefficient (Eq. 1), mean path length (Eq. 2), and the Watts-Strogatz / Girvan-Newman / Barabasi-Albert archetypes.
Background from Refs. [20-24]; not proved in the paper and not controversial.
domain assumption The mean shortest path over m=1000 random seeds in the largest connected component is a robust estimator of the global path length.
Statistical approximation in 'Topological analysis'; no convergence or variance analysis beyond the 20-run spread in Fig. 5(e).
ad hoc to paper Each skeleton crossing maps to graph edges by a well-defined deterministic rule connecting exactly one pair of the four incident segment-ends.
Needed for the reported 2e/N ≈ 3.6 mean degree (Table 1) but never stated — which two of the four ends are paired (same wire, perpendicular pair, or random) is left open; this determines the degree distribution used for the small-world classification.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Nanowire networks' interconnection graphs from their photomicrographs." pith.science (2026). https://pith.science/paper/5K4SEB6R

@misc{pith2026260716445,
  author       = {Pith},
  title        = {Pith review of: Nanowire networks' interconnection graphs from their photomicrographs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5K4SEB6R}},
  note         = {Machine review of arXiv:2607.16445}
}

read the original abstract

Self-assemblies of tunable units are being intensively studied as physical systems with signal-processing capabilities. Specifically, silver nanowire networks (AgNWNs) have demonstrated accumulation, non-linearity, and memory retention with multiple timescales, features that enable a wide variety of neuromorphic implementations. In this study, we aim to extract the interconnection scheme to analyze the experimentally obtained network architecture and, eventually, use it as input to a previously developed simulation platform. By post-processing photomicrographs of AgNWN, we present a pipeline optimized to extract the interconnection diagram, recognizing the intersections formed among the nanowires, to determine the associated graph for each physical sample. A graph is a collection of nodes and edges whose properties can be associated with different electrical responses. It is thus possible to study graphs' metrics, such as the degree distribution, community size, clustering coefficient, and path-length, to compare the experimental assemblies' attributes to those of topological models of reference. Small-world, modular, and scale-free are well-known structures in the field of mathematical graphs. By analyzing the degree distribution, the adjacency matrices, among other useful representation means, the experimental assemblies reveal similarities to both small-world and modular topologies. All the mentioned analyses were conducted considering the interconnection scheme obtained from zenithal-view optical images, which overestimates the number of NWs' interconnections (due to the impossibility of distinguishing real junctions from spurious cross-points between vertically displaced NWs). For that reason, this communication also studies the impact of artificially removing junctions from the resulting graphs on the previously calculated clustering coefficient and path length.

Figures

Figures reproduced from arXiv: 2607.16445 by the authors.

**Figure 1.** Infograph of a typical nanowire network (NWN) depicting the constituent units. Each NWN comprises the network of nanowires (NWs) intersecting each other, enclosed by two macroscopic electrodes placed 1 mm apart. A zoomed-in area is included as an inset where individual NWs and their junctions can be spotted. The image represents the so-called active area, which does not cover the full substrate surface but is a lat… view at source ↗

**Figure 2.** Sketch of the pipeline flow. Starting from the stitched picture, the procedure identifies the electrodes and binarizes the RGB-format image to facilitate the skeletonization conversion. Once the network has been converted to a set of one-pixel-wide contours, network extraction is performed. The wires and their mutual intersections (zenithal view) are interpreted as a graph comprised of nodes and edges, respectively,… view at source ↗

**Figure 3.** Different network schemes used as a reference: small-world, modular, and scalefree. The top row, (a), (b), and (c), comprises a degree histogram including an inset with sketches of a moderate size to help visualization. On the second row, (d), (e), and (f) show the adjacency matrix, while the bottom row, (g), (h), and (i), displays the distribution of community sizes for each case. Degree distributions, adjacency m… view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Network metrics of graphs generated from experimental images. Degree distributions with optical images as insets with a scale bar of 100 µm ((a), (b), and (c)), adjacency matrices ((d), (e), and (f)), and community distributions ((g), (h), and (i)) for three experimen…

**Figure 5.** Figure 5: Clustering coefficient and path length as a function of the edge-removal probability. (a), (b) and (c) depict schematics of the different graphs used to measure the path length. (d) shows the clustering coefficient as a function of the edge-removal probability for the…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 12 canonical work pages

[1]

Haykin.Neural Networks: A Comprehensive Foundation

Simon S. Haykin.Neural Networks: A Comprehensive Foundation. en. Prentice Hall, 1999.isbn: 978-0-13-273350-2

1999
[2]

Toward a formal theory for computing machines made out of whatever physics offers

Herbert Jaeger, Beatriz Noheda, and Wilfred G. van der Wiel. “Toward a formal theory for computing machines made out of whatever physics offers”. en. In:Nature Communications14.1 (Aug. 2023), p. 4911.issn: 2041-1723.doi:10 . 1038 / s41467 - 023 - 40533 - 1.url:https : //www.nature.com/articles/s41467-023-40533-1(visited on 05/15/2025)

2023
[3]

Deep learning

Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. “Deep learning”. en. In:Nature521.7553 (May 2015), pp. 436–444.issn: 1476-4687.doi:10.1038/nature14539.url:https://www. nature.com/articles/nature14539(visited on 08/07/2020)

work page doi:10.1038/nature14539.url:https://www 2015
[4]

The balance between integration and segregation drives network dynamics maximizing multistability and metastability

Javier Palma-Espinosa et al. “The balance between integration and segregation drives network dynamics maximizing multistability and metastability”. en. In:Scientific Reports15.1 (May 2025), p. 18811.issn: 2045-2322.doi:10.1038/s41598- 025- 01612- z.url:https://www. nature.com/articles/s41598-025-01612-z(visited on 02/05/2026)

doi:10.1038/s41598- 2025
[5]

Neuromorphic Atomic Switch Networks

Audrius V. Avizienis et al. “Neuromorphic Atomic Switch Networks”. en. In:PLOS ONE7.8 (2012), e42772.issn: 1932-6203.doi:10 . 1371 / journal . pone . 0042772.url:https : / / journals . plos . org / plosone / article ? id = 10 . 1371 / journal . pone . 0042772(visited on 11/28/2022)

2012
[6]

Brain-Inspired Structural Plasticity through Reweighting and Rewiring in Multi-Terminal Self-Organizing Memristive Nanowire Networks

Gianluca Milano et al. “Brain-Inspired Structural Plasticity through Reweighting and Rewiring in Multi-Terminal Self-Organizing Memristive Nanowire Networks”. en. In:Advanced Intelligent Systems2.8 (2020), p. 2000096.issn: 2640-4567.doi:https : / / doi . org / 10 . 1002 / aisy . 202000096.url:https://onlinelibrary.wiley.com/doi/abs/10.1002/aisy.202000096 ...

doi:10.1002/aisy.202000096 2020
[7]

Neuromorphic nanowire networks: principles, progress and future prospects for neuro-inspired information processing

Zdenka Kuncic and Tomonobu Nakayama. “Neuromorphic nanowire networks: principles, progress and future prospects for neuro-inspired information processing”. In:Advances in Physics: X6.1 (2021), p. 1894234.issn: null.doi:10.1080/23746149.2021.1894234.url:https://doi.org/ 10.1080/23746149.2021.1894234(visited on 03/03/2022)

arXiv 2021
[8]

Ferroelastic Domain Walls in BiFeO3 as Memristive Networks

Jan L. Rieck et al. “Ferroelastic Domain Walls in BiFeO3 as Memristive Networks”. en. In: Advanced Intelligent Systems5.1 (2023), p. 2200292.issn: 2640-4567.doi:10 . 1002 / aisy . 202200292.url:https://onlinelibrary.wiley.com/doi/abs/10.1002/aisy.202200292 (visited on 11/14/2022)

work page doi:10.1002/aisy.202200292 2023

Show all 24 references

[9]

Two-Junction Model in Different Percolation Regimes of Silver Nanowires Networks

Juan Ignacio Diaz Schneider et al. “Two-Junction Model in Different Percolation Regimes of Silver Nanowires Networks”. en. In:Advanced Functional Materials34.52 (2024), p. 2410766. issn: 1616-3028.doi:10.1002/adfm.202410766.url:https://onlinelibrary.wiley.com/ doi/abs/10.1002/...

2024 doi
[10]

Thermal evolution of low-temperature magnetic texture modulation in FePt thin films by direct visualization

Cynthia P Quinteros et al. “Thermal evolution of low-temperature magnetic texture modulation in FePt thin films by direct visualization”. en. In:Journal of Physics D: Applied Physics57.18 (Feb. 2024), p. 185001.issn: 0022-3727.doi:10.1088/1361-6463/ad2567.url:https://dx. doi.o...

2024 doi
[11]

Self-Assembly at All Scales

George M. Whitesides and Bartosz Grzybowski. “Self-Assembly at All Scales”. In:Science 295.5564 (Mar. 2002), pp. 2418–2421.doi:10.1126/science.1070821.url:https://www. science.org/doi/full/10.1126/science.1070821(visited on 04/05/2023)

2002 doi
[12]

Nanoarchitectonics: what’s coming next after nanotechnology?

Katsuhiko Ariga. “Nanoarchitectonics: what’s coming next after nanotechnology?” en. In:Nanoscale Horizons6.5 (May 2021), pp. 364–378.issn: 2055-6764.doi:10.1039/D0NH00680G.url:https: //pubs.rsc.org/en/content/articlelanding/2021/nh/d0nh00680g(visited on 04/04/2023). 9

2021 doi
[13]

Solution-processed silver sulphide nanocrystal film for resistive switching memories

Beatriz Mart ´ ın-Garc ´ ıa et al. “Solution-processed silver sulphide nanocrystal film for resistive switching memories”. en. In:Journal of Materials Chemistry C6.48 (Dec. 2018), pp. 13128– 13135.issn: 2050-7534.doi:10.1039/C8TC04068K.url:https://pubs.rsc.org/en/content/ arti...

2018 doi
[14]

Resistance of Single Ag Nanowire Junctions and Their Role in the Conductivity of Nanowire Networks

Allen T. Bellew et al. “Resistance of Single Ag Nanowire Junctions and Their Role in the Conductivity of Nanowire Networks”. In:ACS Nano9.11 (Nov. 2015), pp. 11422–11429.issn: 1936-0851.doi:10 . 1021 / acsnano . 5b05469.url:https : / / doi . org / 10 . 1021 / acsnano . 5b05469...

2015
[15]

Reservoir computing with 3D nanowire networks

R. K. Daniels et al. “Reservoir computing with 3D nanowire networks”. In:Neural Networks 154 (Oct. 2022), pp. 122–130.issn: 0893-6080.doi:10.1016/j.neunet.2022.07.001.url: https://www.sciencedirect.com/science/article/pii/S089360802200257X(visited on 12/11/2024)

2022 doi
[16]

Connectome of memristive nanowire networks through graph theory

Gianluca Milano, Enrique Miranda, and Carlo Ricciardi. “Connectome of memristive nanowire networks through graph theory”. en. In:Neural Networks150 (June 2022), pp. 137–148.issn: 0893-6080.doi:10.1016/j.neunet.2022.02.022.url:https://www.sciencedirect.com/ science/article/pii/...

2022 doi
[17]

Ultimate conductivity performance in metallic nanowire net- works

Claudia Gomes da Rocha et al. “Ultimate conductivity performance in metallic nanowire net- works”. en. In:Nanoscale7.30 (July 2015), pp. 13011–13016.issn: 2040-3372.doi:10.1039/ C5NR03905C.url:https://pubs.rsc.org/en/content/articlelanding/2015/nr/c5nr03905c (visited on 06/19/2026)

2015
[18]

Resistive Switching of Self-Assembled Silver Nanowire Net- works Governed by Environmental Conditions

Juan Ignacio Diaz Schneider et al. “Resistive Switching of Self-Assembled Silver Nanowire Net- works Governed by Environmental Conditions”. en. In:Advanced Electronic Materials8.11 (2022), p. 2200631.issn: 2199-160X.doi:10.1002/aelm.202200631.url:https://onlinelibrary. wiley.c...

2022 doi
[19]

A Fast Parallel Algorithm for Thinning Digital Patterns

T. Y. Zhang and C. Y. Suen. “A Fast Parallel Algorithm for Thinning Digital Patterns”. In: Commun. ACM27.3 (1984).doi:10.1145/357994.358023

1984
[20]

Collective dynamics of ‘small-world’ networks

Duncan J. Watts and Steven H. Strogatz. “Collective dynamics of ‘small-world’ networks”. en. In:Nature393.6684 (June 1998), pp. 440–442.issn: 1476-4687.doi:10 . 1038 / 30918.url: https://www.nature.com/articles/30918(visited on 11/07/2018)

1998
[21]

Accedido: 2026-06-18

NetworkX Developers.clustering — NetworkX Reference. Accedido: 2026-06-18. 2024.url: https : / / networkx . org / documentation / stable / reference / algorithms / generated / networkx.algorithms.cluster.clustering.html

2026
[22]

Intensity and Coherence of Motifs in Weighted Complex Networks

Jukka-Pekka Onnela et al. “Intensity and Coherence of Motifs in Weighted Complex Networks”. In:Physical Review E71.6 (June 13, 2005).doi:10.1103/PhysRevE.71.065103

2005 doi
[23]

Community structure in social and biological networks

M. Girvan and M. E. J. Newman. “Community structure in social and biological networks”. In:Proceedings of the National Academy of Sciences99.12 (June 2002), pp. 7821–7826.doi: 10 . 1073 / pnas . 122653799.url:https : / / www . pnas . org / doi / 10 . 1073 / pnas . 122653799 (v...

2002
[24]

Emergence of Scaling in Random Networks

Albert-L´ aszl´ o Barab´ asi and R´ eka Albert. “Emergence of Scaling in Random Networks”. In: Science286.5439 (Oct. 1999), pp. 509–512.doi:10.1126/science.286.5439.509.url:https: //www.science.org/doi/10.1126/science.286.5439.509(visited on 06/09/2026). 10

1999 doi

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.

[1] [1]

Haykin.Neural Networks: A Comprehensive Foundation

Simon S. Haykin.Neural Networks: A Comprehensive Foundation. en. Prentice Hall, 1999.isbn: 978-0-13-273350-2

1999

[2] [2]

Toward a formal theory for computing machines made out of whatever physics offers

Herbert Jaeger, Beatriz Noheda, and Wilfred G. van der Wiel. “Toward a formal theory for computing machines made out of whatever physics offers”. en. In:Nature Communications14.1 (Aug. 2023), p. 4911.issn: 2041-1723.doi:10 . 1038 / s41467 - 023 - 40533 - 1.url:https : //www.nature.com/articles/s41467-023-40533-1(visited on 05/15/2025)

2023

[3] [3]

Deep learning

Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. “Deep learning”. en. In:Nature521.7553 (May 2015), pp. 436–444.issn: 1476-4687.doi:10.1038/nature14539.url:https://www. nature.com/articles/nature14539(visited on 08/07/2020)

work page doi:10.1038/nature14539.url:https://www 2015

[4] [4]

The balance between integration and segregation drives network dynamics maximizing multistability and metastability

Javier Palma-Espinosa et al. “The balance between integration and segregation drives network dynamics maximizing multistability and metastability”. en. In:Scientific Reports15.1 (May 2025), p. 18811.issn: 2045-2322.doi:10.1038/s41598- 025- 01612- z.url:https://www. nature.com/articles/s41598-025-01612-z(visited on 02/05/2026)

doi:10.1038/s41598- 2025

[5] [5]

Neuromorphic Atomic Switch Networks

Audrius V. Avizienis et al. “Neuromorphic Atomic Switch Networks”. en. In:PLOS ONE7.8 (2012), e42772.issn: 1932-6203.doi:10 . 1371 / journal . pone . 0042772.url:https : / / journals . plos . org / plosone / article ? id = 10 . 1371 / journal . pone . 0042772(visited on 11/28/2022)

2012

[6] [6]

Brain-Inspired Structural Plasticity through Reweighting and Rewiring in Multi-Terminal Self-Organizing Memristive Nanowire Networks

Gianluca Milano et al. “Brain-Inspired Structural Plasticity through Reweighting and Rewiring in Multi-Terminal Self-Organizing Memristive Nanowire Networks”. en. In:Advanced Intelligent Systems2.8 (2020), p. 2000096.issn: 2640-4567.doi:https : / / doi . org / 10 . 1002 / aisy . 202000096.url:https://onlinelibrary.wiley.com/doi/abs/10.1002/aisy.202000096 ...

doi:10.1002/aisy.202000096 2020

[7] [7]

Neuromorphic nanowire networks: principles, progress and future prospects for neuro-inspired information processing

Zdenka Kuncic and Tomonobu Nakayama. “Neuromorphic nanowire networks: principles, progress and future prospects for neuro-inspired information processing”. In:Advances in Physics: X6.1 (2021), p. 1894234.issn: null.doi:10.1080/23746149.2021.1894234.url:https://doi.org/ 10.1080/23746149.2021.1894234(visited on 03/03/2022)

arXiv 2021

[8] [8]

Ferroelastic Domain Walls in BiFeO3 as Memristive Networks

Jan L. Rieck et al. “Ferroelastic Domain Walls in BiFeO3 as Memristive Networks”. en. In: Advanced Intelligent Systems5.1 (2023), p. 2200292.issn: 2640-4567.doi:10 . 1002 / aisy . 202200292.url:https://onlinelibrary.wiley.com/doi/abs/10.1002/aisy.202200292 (visited on 11/14/2022)

work page doi:10.1002/aisy.202200292 2023