pith. machine review for the scientific record.
sign in

arxiv: 2509.00747 · v2 · submitted 2025-08-31 · ❄️ cond-mat.dis-nn · cond-mat.mes-hall· cond-mat.soft· cs.ET· cs.LG

Self-Organising Memristive Networks as Physical Learning Systems

Francesco Caravelli , Gianluca Milano , Adam Z. Stieg , Carlo Ricciardi , Simon Anthony Brown , Zdenka Kuncic This is my paper

Pith reviewed 2026-05-18 20:11 UTC · model grok-4.3

classification ❄️ cond-mat.dis-nn cond-mat.mes-hallcond-mat.softcs.ETcs.LG
keywords self-organising memristive networksphysical learningadaptive dynamicsphase transitionsedge intelligencecontinual learningneuromorphic hardware
0
0 comments X

The pith

Self-organising memristive networks harness their own nonlinear dynamics to perform physical learning.

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

This perspective argues that networks built from nanoscale resistive memory elements can self-organise their electrical connections and thereby perform learning tasks directly in hardware. Experiments show these networks display collective adaptive behaviors and pass through critical states when switching conductance, behaviors that theoretical models from graph theory and disordered systems help explain. The authors connect these properties to biological plasticity and conclude that the same physics could support continual learning in low-power devices. A reader would care if the claim holds because it points to a way of building intelligence that avoids the high energy cost of running neural-network software on conventional chips.

Core claim

Self-organising memristive networks composed of dynamically reconfigurable resistive memory components exhibit non-trivial interactions whose collective nonlinear and adaptive dynamics can be used for learning, with both experiments and mean-field plus graph-theoretic models revealing criticality and dynamical phase transitions between conductance states that parallel plasticity in biological neuronal networks.

What carries the argument

Self-organising memristive networks (SOMNs), physical assemblies of resistive memory nanoscale components whose electrical circuitry reconfigures itself and whose conductance-state transitions produce the adaptive dynamics required for learning.

If this is right

  • Continuous learning becomes possible inside resource-limited hardware without external software updates.
  • Real-time decision-making for autonomous systems and dynamic sensing can be embedded directly at the edge.
  • Energy use for intelligence tasks drops because computation occurs in the physical substrate rather than in separate processors.
  • Personalised healthcare devices could run adaptive algorithms locally because the networks support continual plasticity-like behavior.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Different fabrication methods for the resistive components may produce networks with distinct critical exponents that could be matched to particular classes of learning problems.
  • Mean-field descriptions of the conductance transitions might be used to predict the minimal network size needed for a target task before fabrication.
  • Hybrid systems that combine these networks with conventional sensors could test whether the physical learning layer improves robustness to noisy inputs compared with purely digital approaches.

Load-bearing premise

The non-trivial interactions and adaptive dynamics seen in these networks can be steered to perform useful learning tasks in varied hardware implementations.

What would settle it

An experiment in which a self-organising memristive network is given repeated input patterns yet shows no measurable change in its output behavior or conductance distribution would falsify the claim that its dynamics can be harnessed for learning.

Figures

Figures reproduced from arXiv: 2509.00747 by Adam Z. Stieg, Carlo Ricciardi, Francesco Caravelli, Gianluca Milano, Simon Anthony Brown, Zdenka Kuncic.

Figure 1
Figure 1. Figure 1: Conceptual overview of self-organised memristive networks (SOMNs) as a platform for physical learning systems. a. Schematic of a SOMN chip device depicting synaptic conductance changes (red) in response to electrical signals; inset depicts a nanoscale electrical junction where resistive switching results from the formation of conductive filaments between the self-assembled elements of the network, driven b… view at source ↗
Figure 2
Figure 2. Figure 2: Memristive behaviour of SOMNs. (a) Memristive behaviour of a NW network under voltage sweep stimulation in two-terminal configuration (see inset), characterized by an hysteretic loop in the I-V plane typical of memristive systems. Adapted from ref.44. (b) Short-term plasticity effects in a NW network, characterized by potentiation of the conductance state during voltage stimulation followed by spontaneous … view at source ↗
Figure 3
Figure 3. Figure 3: Modeling emergent behaviour of memristive networks. (a) Experimental data on potentiation and relaxation dynamics related to short-term plasticity effects in NW networks under voltage stimulation (blue line), grid-graph modeling (black dashed line) and mean-field theory results (green dashed line) obtained by using methods outlined in sec. 3. (b) Results from grid-graph modeling allow to visualize spatiote… view at source ↗
Figure 4
Figure 4. Figure 4: Conductance transitions in memristive networks. a. Schematic illustration of conductance phase transitions in SOMNs, where the macroscopic order parameter (average conductance G) depends on control parameters such as applied voltage V, nanowire/nanoparticle (areal) density ρ, or average connectivity d. The shaded region represents the transition regime where emergent collective behaviour manifests. Dependi… view at source ↗
Figure 5
Figure 5. Figure 5: Physical learning paradigms for self-organising memristive networks. Arbitrary input data (left) is first pre-processed and encoded into input voltage signals (centre left) and delivered to a multi-terminal SOMN device (centre main), where conductance changes adapt to the input signal patterns. Dynamical reconfiguration of the SOMN, governed by internal transport dynamics and Kirchhoff constraints, physica… view at source ↗
read the original abstract

Learning with physical systems is an emerging paradigm that seeks to harness the intrinsic nonlinear dynamics of physical substrates for learning. The impetus for a paradigm shift in how hardware is used for computational intelligence stems largely from the unsustainability of artificial neural network software implemented on conventional transistor-based hardware. This Perspective highlights one promising approach using physical networks comprised of resistive memory nanoscale components with dynamically reconfigurable, self-organising electrical circuitry. Experimental advances have revealed the non-trivial interactions within these Self-Organising Memristive Networks (SOMNs), offering insights into their collective nonlinear and adaptive dynamics, and how these properties can be harnessed for learning using different hardware implementations. Theoretical approaches, including mean-field theory, graph theory, and concepts from disordered systems, reveal deeper insights into the dynamics of SOMNs, especially during transitions between different conductance states where criticality and other dynamical phase transitions emerge in both experiments and models. Furthermore, parallels between adaptive dynamics in SOMNs and plasticity in biological neuronal networks suggest the potential for realising energy-efficient, brain-like continual learning. SOMNs thus offer a promising route toward embedded edge intelligence, unlocking real-time decision-making for autonomous systems, dynamic sensing, and personalised healthcare, by embedding continuous learning in resource-constrained environments. The overarching aim of this Perspective is to show how the convergence of nanotechnology, statistical physics, complex systems, and self-organising principles offers a unique opportunity to advance a new generation of physical intelligence technologies.

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

1 major / 1 minor

Summary. This Perspective article reviews experimental observations of non-trivial interactions, collective nonlinear dynamics, and adaptive behavior in Self-Organising Memristive Networks (SOMNs) composed of resistive memory nanoscale components. It discusses theoretical tools including mean-field theory, graph theory, and concepts from disordered systems that illuminate criticality and dynamical phase transitions during conductance-state changes. The manuscript draws parallels between SOMN plasticity and biological neuronal networks, and argues that these properties can be harnessed to realize energy-efficient, brain-like continual learning, thereby offering a route to embedded edge intelligence for autonomous systems, dynamic sensing, and personalised healthcare.

Significance. If the perspective's synthesis holds, the convergence of nanotechnology, statistical physics, and self-organising principles could help establish physical learning systems as a viable complement to conventional hardware, potentially enabling more sustainable and adaptive computational intelligence in resource-constrained settings. The manuscript usefully collates recent experimental and modeling advances without introducing new data or derivations.

major comments (1)
  1. [Abstract] Abstract: The claim that observed non-trivial interactions and adaptive dynamics 'can be harnessed for learning using different hardware implementations' and that SOMNs 'offer a promising route toward embedded edge intelligence' is central to the perspective, yet the text provides no quantitative learning-task benchmarks (e.g., classification accuracy, regression error, or energy-per-inference metrics) or explicit control mechanisms that map the reviewed dynamics onto task-relevant outputs. This leaves the transition from 'promising dynamics' to functional learning as an untested assumption rather than a demonstrated capability.
minor comments (1)
  1. [Abstract] The abstract and closing paragraph could more explicitly separate the reviewed experimental and theoretical results from the forward-looking statements about applications in autonomous systems and healthcare.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive review and recommendation of minor revision. The single major comment is addressed point-by-point below. As this is a Perspective article synthesizing existing literature rather than reporting new experiments, our response focuses on clarifying scope and strengthening references to supporting studies.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that observed non-trivial interactions and adaptive dynamics 'can be harnessed for learning using different hardware implementations' and that SOMNs 'offer a promising route toward embedded edge intelligence' is central to the perspective, yet the text provides no quantitative learning-task benchmarks (e.g., classification accuracy, regression error, or energy-per-inference metrics) or explicit control mechanisms that map the reviewed dynamics onto task-relevant outputs. This leaves the transition from 'promising dynamics' to functional learning as an untested assumption rather than a demonstrated capability.

    Authors: We agree that a Perspective should clearly ground its forward-looking claims in the reviewed literature. The manuscript does not introduce new benchmarks because its purpose is to collate and interpret recent experimental and modeling advances (as noted in the referee summary). However, the claims are not untested assumptions; they rest on quantitative results reported in the cited works, including demonstrations of classification accuracies above 90% on simple tasks, energy efficiencies orders of magnitude below digital implementations, and explicit mapping of conductance dynamics to output via read-out layers or external control. To address the concern directly, we will revise the abstract to include a brief clause referencing these supporting demonstrations and will add a short paragraph in the main text summarizing key performance metrics from the literature. This revision will make the transition from observed dynamics to functional capability more explicit without altering the Perspective's scope. revision: yes

Circularity Check

0 steps flagged

No significant circularity; perspective summarizes independent advances without tautological derivations

full rationale

This Perspective reviews experimental observations of collective dynamics in SOMNs, theoretical frameworks (mean-field, graph theory, criticality from disordered systems), and biological parallels drawn from the broader literature. No derivation chain, equations, or first-principles predictions are presented that reduce by construction to self-fitted parameters, renamed inputs, or load-bearing self-citations. Forward-looking statements about embedded edge intelligence rest on cited external advances rather than internal tautologies, rendering the analysis self-contained against independent benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This perspective relies on prior experimental and theoretical literature without introducing new free parameters, axioms, or invented entities in the summary provided.

pith-pipeline@v0.9.0 · 5819 in / 1076 out tokens · 41237 ms · 2026-05-18T20:11:56.596629+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

  • IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    Theoretical approaches, including mean-field theory, graph theory, and concepts from disordered systems, reveal deeper insights into the dynamics of SOMNs, especially during transitions between different conductance states where criticality and other dynamical phase transitions emerge

  • IndisputableMonolith/Foundation/BranchSelection.lean branch_selection unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    Using the theoretical framework outlined in sec. 3.1, recent work has formulated a dynamical mean-field theory that attempts to characterize the voltage-induced dynamics of SOMNs by reducing equations of the type (4) to a single mean field dynamical equation. This is written in terms of an effective potential VΔv(⟨g⟩)

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

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

  1. Thermal Anomaly Detection using Physics Aware Neuromorphic Networks: Comparison between Raw and L1C Sentinel-2 Data

    eess.SP 2026-04 unverdicted novelty 5.0

    PANN achieves MCC 0.809 on raw Sentinel-2 L0 data for thermal anomaly detection with 2.44s latency per granule, below the 3.6s acquisition time, and projects even faster neuromorphic hardware performance.

Reference graph

Works this paper leans on

170 extracted references · 170 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    J., van der Wiel, W

    Kaspar, C., Ravoo, B. J., van der Wiel, W. G., Weg- ner, S. V . & Pernice, W. H. P. The rise of intelli- gent matter. Nature 594, 345–355, DOI: 10.1038/ s41586-021-03453-y (2021)

    work page 2021

  2. [2]

    & van der Wiel, W

    Jaeger, H., Noheda, B. & van der Wiel, W. G. Toward a formal theory for computing machines made out of whatever physics offers. Nat. Commun. 14, 4911 (2023)

    work page 2023

  3. [3]

    & Murugan, A

    Stern, M. & Murugan, A. Learning without neu- rons in physical systems. Annu. Rev. Condens. Mat- ter Phys. 14, 417–441, DOI: https://doi.org/10.1146/ annurev-conmatphys-040821-113439 (2023)

    work page 2023

  4. [4]

    Wang, Z. et al. Resistive switching materials for infor- mation processing. Nat. Rev. Mater. 5, 173–195, DOI: 10.1038/s41578-019-0159-3 (2020)

    work page doi:10.1038/s41578-019-0159-3 2020

  5. [5]

    Information and Self-Organization: A Macroscopic Approach to Complex Systems(Springer: Berlin/Heidelberg, Germany, 2006)

    Haken, H. Information and Self-Organization: A Macroscopic Approach to Complex Systems(Springer: Berlin/Heidelberg, Germany, 2006)

    work page 2006

  6. [6]

    Anderson, P. W. More is different: Broken symmetry and the nature of the hierarchical structure of science. Science 177, 393–396, DOI: 10.1126/science.177.4047. 393 (1972)

    work page doi:10.1126/science.177.4047 1972

  7. [7]

    Outline of a theory of thought-processes and thinking machines

    Caianiello, E. Outline of a theory of thought-processes and thinking machines. J. Theor. Biol. 1, 204–235, DOI: 10.1016/0022-5193(61)90046-7 (1961)

    work page doi:10.1016/0022-5193(61)90046-7 1961

  8. [8]

    The perceptron: a probabilistic model for information storage and organization in the brain

    Rosenblatt, F. The perceptron: a probabilistic model for information storage and organization in the brain. Psychol. Rev. 65, 386–408, DOI: 10.1037/h0042519 (1958)

    work page doi:10.1037/h0042519 1958

  9. [9]

    Principles of Neurodynamics: Percep- trons and the Theory of Brain Mechanisms (Spartan Books, 1962)

    Rosenblatt, F. Principles of Neurodynamics: Percep- trons and the Theory of Brain Mechanisms (Spartan Books, 1962)

    work page 1962

  10. [10]

    Hopfield, J. J. Neural networks and physical systems with emergent collective computational abilities. Proc. Natl. Acad. Sci. 79, 2554–2558, DOI: 10.1073/pnas.79. 8.2554 (1982)

    work page doi:10.1073/pnas.79 1982

  11. [11]

    Approximation by superpositions of a sigmoidal function

    Cybenko, G. Approximation by superpositions of a sigmoidal function. Math. Control. Signals Syst. 2, 303– 314, DOI: 10.1007/BF02551274 (1989)

    work page doi:10.1007/bf02551274 1989

  12. [12]

    Stieg, A. Z. et al. Emergent criticality in complex tur- ing b-type atomic switch networks. Adv. Mater. 24, 286–293, DOI: 10.1002/adma.201103053 (2011)

    work page doi:10.1002/adma.201103053 2011

  13. [13]

    V .et al

    Avizienis, A. V .et al. Morphological transitions from dendrites to nanowires in the electroless deposition of silver. Cryst. Growth Des. 13, 465–469, DOI: 10.1021/ cg301692n (2013)

    work page 2013

  14. [14]

    Stieg, A. Z. et al. Self-organization and Emergence of Dynamical Structures in Neuromorphic Atomic Switch Networks. In Adamatzky, A. & Chua, L. (eds.) Memristor Networks, 173–209, DOI: 10.1007/ 978-3-319-02630-5_10 (Springer International Publish- ing, 2014)

    work page 2014

  15. [15]

    J., Gimzewski, J

    Sandouk, E. J., Gimzewski, J. K. & Stieg, A. Z. Mul- tistate resistive switching in silver nanoparticle films. Sci. Technol. Adv. Mater. 16, 045004, DOI: 10.1088/ 1468-6996/16/4/045004 (2015)

    work page 2015

  16. [16]

    Mallinson, J. B. et al. Avalanches and criticality in self-organized nanoscale networks. Sci. Adv. 5, DOI: 10.1126/sciadv.aaw8438 (2019)

    work page doi:10.1126/sciadv.aaw8438 2019

  17. [17]

    V .et al

    Christensen, D. V .et al. 2022 roadmap on neuromor- phic computing and engineering. Neuromorphic Com- put. Eng. 2, 022501, DOI: 10.1088/2634-4386/ac4a83 (2022)

    work page doi:10.1088/2634-4386/ac4a83 2022

  18. [18]

    & Yang, J

    Xia, Q. & Yang, J. J. Memristive crossbar arrays for brain-inspired computing. Nat. materials 18, 309–323 (2019)

    work page 2019

  19. [19]

    Momeni, A. et al. Training of physical neural networks. arXiv preprint arXiv:2406.03372 (2024). Available at https://arxiv.org/abs/2406.03372

    work page arXiv 2024

  20. [20]

    Demis, E. C. et al. Atomic switch net- works—nanoarchitectonic design of a complex system for natural computing. Nanotechnology 26, 204003 (2015)

    work page 2015

  21. [21]

    Mallinson, J. B. et al. Avalanches and criticality in self-organized nanoscale networks. Sci. advances 5, eaaw8438 (2019)

    work page 2019

  22. [22]

    Dynamic models of large-scale brain activity

    Breakspear, M. Dynamic models of large-scale brain activity. Nat. Neurosci. 20, 340–352, DOI: 10.1038/nn. 4497 (2017)

    work page doi:10.1038/nn 2017

  23. [23]

    Lynn, C. W. & Bassett, D. S. The physics of brain network structure, function and control. Nat. Rev. Phys. 1, 318–332, DOI: 10.1038/s42254-019-0040-8 (2019)

    work page doi:10.1038/s42254-019-0040-8 2019

  24. [24]

    Diaz-Alvarez, A. et al. Emergent dynamics of neu- romorphic nanowire networks. Sci. Reports 9, DOI: 10.1038/s41598-019-51330-6 (2019)

    work page doi:10.1038/s41598-019-51330-6 2019

  25. [25]

    Topological properties of neuromor- phic nanowire networks

    Loeffler, A.et al. Topological properties of neuromor- phic nanowire networks. Front. Neurosci. 14, DOI: 10.3389/fnins.2020.00184 (2020)

    work page doi:10.3389/fnins.2020.00184 2020

  26. [26]

    E., Richards, B

    Suárez, L. E., Richards, B. A., Lajoie, G. & Misic, B. Learning function from structure in neuromorphic networks. Nat. Mach. Intell. 3, 771–786, DOI: 10.1038/ s42256-021-00376-1 (2021)

    work page 2021

  27. [27]

    & Ricciardi, C

    Milano, G., Miranda, E. & Ricciardi, C. Connectome of memristive nanowire networks through graph theory. Neural Networks 150, 137–148 (2022)

    work page 2022

  28. [28]

    Vahl, A., Milano, G., Kuncic, Z., Brown, S. A. & Mi- lani, P. Brain-inspired computing with self-assembled 14/20 networks of nano-objects. J. Phys. D: Appl. Phys. 57, 503001, DOI: 10.1088/1361-6463/ad7a82 (2024)

    work page doi:10.1088/1361-6463/ad7a82 2024

  29. [29]

    Kim, T. H. et al. Nanoparticle Assemblies as Memris- tors. Nano Lett. 9, 2229–2233, DOI: 10.1021/nl900030n (2009)

    work page doi:10.1021/nl900030n 2009

  30. [30]

    & Miranda, E

    Milano, G., Michieletti, F., Pilati, D., Ricciardi, C. & Miranda, E. Self-organizing neuromorphic nanowire networks as stochastic dynamical systems. Nat. Com- mun. 16, 3509 (2025)

    work page 2025

  31. [31]

    Zhu, R. et al. Information dynamics in neuromorphic nanowire networks. Sci. reports 11, 13047 (2021)

    work page 2021

  32. [32]

    B., Snider, G

    Strukov, D. B., Snider, G. S., Stewart, D. R. & Williams, R. S. The missing memristor found. Nature 453, 80–83, DOI: 10.1038/nature06932 (2008)

    work page doi:10.1038/nature06932 2008

  33. [33]

    & Carbajal, J

    Caravelli, F. & Carbajal, J. P. Memristors for the cu- rious outsiders. Technologies 6, 118, DOI: 10.3390/ technologies6040118 (2018)

    work page 2018

  34. [34]

    Terasa, M.-I. et al. Pathways towards truly brain-like computing primitives. Mater. Today 69, 41–53, DOI: 10.1016/j.mattod.2023.07.019 (2023)

    work page doi:10.1016/j.mattod.2023.07.019 2023

  35. [35]

    & Bak, P

    Chialvo, D. & Bak, P. Learning from mistakes. Neuro- science 90, 1137–1148, DOI: 10.1016/s0306-4522(98) 00472-2 (1999)

    work page doi:10.1016/s0306-4522(98 1999

  36. [36]

    P., Dambre, J., Hermans, M

    Carbajal, J. P., Dambre, J., Hermans, M. & Schrauwen, B. Memristor models for machine learning.Neural Com- put. 27, 725–747, DOI: 10.1162/neco_a_00694 (2015)

    work page doi:10.1162/neco_a_00694 2015

  37. [37]

    P., Martin, D

    Carbajal, J. P., Martin, D. A. & Chialvo, D. R. Learning by mistakes in memristor networks. Phys. Rev. E 105, DOI: 10.1103/physreve.105.054306 (2022)

    work page doi:10.1103/physreve.105.054306 2022

  38. [38]

    Barrows, F., Sheldon, F. C. & Caravelli, F. Network analysis of memristive device circuits: dynamics, stabil- ity and correlations, DOI: 10.48550/ARXIV .2402.16015 (2024)

    work page internal anchor Pith review doi:10.48550/arxiv 2024

  39. [39]

    Zhou, Z. et al. Edge intelligence: Paving the last mile of artificial intelligence with edge computing. Proc. IEEE 107, 1738–1762, DOI: 10.1109/JPROC.2019.2918951 (2019)

    work page doi:10.1109/jproc.2019.2918951 2019

  40. [40]

    Fixing AI’s energy crisis

    Bourzac, K. Fixing AI’s energy crisis. Nature DOI: 10.1038/d41586-024-03408-z (2024)

    work page doi:10.1038/d41586-024-03408-z 2024

  41. [41]

    & Ricciardi, C

    Milano, G., Porro, S., Valov, I. & Ricciardi, C. Recent developments and perspectives for memristive devices based on metal oxide nanowires.Adv. Electron. Mater.5, 1800909, DOI: https://doi.org/10.1002/aelm.201800909 (2019)

    work page doi:10.1002/aelm.201800909 2019

  42. [42]

    & Nakayama, T

    Kuncic, Z. & Nakayama, T. Neuromorphic nanowire networks: principles, progress and future prospects for neuro-inspired information processing. Adv. Physics: X 6, DOI: 10.1080/23746149.2021.1894234 (2021)

    work page doi:10.1080/23746149.2021.1894234 2021

  43. [43]

    Manning, H. G. et al. Emergence of winner-takes-all connectivity paths in random nanowire networks. Nat. communications 9, 3219 (2018)

    work page 2018

  44. [44]

    Milano, G., Pedretti, G., Fretto, M. & et al. Brain- inspired structural plasticity through reweighting and rewiring in multi-terminal self-organizing memristive nanowire networks. Adv. Intell. Syst. 2, 2000096, DOI: 10.1002/aisy.202000096 (2020)

    work page doi:10.1002/aisy.202000096 2020

  45. [45]

    Sillin, H. O. et al. A theoretical and experimental study of neuromorphic atomic switch networks for reservoir computing. Nanotechnology 24, 384004 (2013)

    work page 2013

  46. [46]

    Li, Q. et al. Dynamic electrical pathway tuning in neu- romorphic nanowire networks. Adv. Funct. Mater. 30, 2003679, DOI: https://doi.org/10.1002/adfm.202003679 (2020). https://advanced.onlinelibrary.wiley.com/doi/ pdf/10.1002/adfm.202003679

    work page doi:10.1002/adfm.202003679 2020

  47. [47]

    Lilak, S. et al. Spoken digit classification by in-materio reservoir computing with neuromorphic atomic switch networks. Front. Nanotechnol. 3, DOI: 10.3389/fnano. 2021.675792 (2021)

    work page doi:10.3389/fnano 2021

  48. [48]

    Zhu, R. et al. Online dynamical learning and sequence memory with neuromorphic nanowire networks. Nat. Commun. 14, 6697 (2023)

    work page 2023

  49. [49]

    Kotooka, T. et al. Thermally Stable Ag2Se Nanowire Network as an Effective In-Materio Physical Reservoir Computing Device. Adv. Electron. Mater. 10, 2400443, DOI: 10.1002/aelm.202400443 (2024)

    work page doi:10.1002/aelm.202400443 2024

  50. [50]

    & Kozicki, M

    Valov, I., Waser, R., Jameson, J. & Kozicki, M. Electro- chemical metallization memories—fundamentals, appli- cations, prospects. Nanotechnology 22, 254003, DOI: 10.1088/0957-4484/22/28/289502 (2011)

    work page doi:10.1088/0957-4484/22/28/289502 2011

  51. [51]

    Electrochemical dynamics of nanoscale metallic inclusions in dielectrics

    Yang, Y .et al. Electrochemical dynamics of nanoscale metallic inclusions in dielectrics. Nat. communications 5, 4232 (2014)

    work page 2014

  52. [52]

    Milano, G. et al. Electrochemical rewiring through quan- tum conductance effects in single metallic memristive nanowires. Nanoscale Horizons 9, 416–426 (2024)

    work page 2024

  53. [53]

    K., Mallinson, J

    Bose, S. K., Mallinson, J. B., Gazoni, R. M. & Brown, S. A. Stable self-assembled atomic-switch networks for neuromorphic applications. IEEE Transactions on Electron Devices 64, 5194–5201 (2017)

    work page 2017

  54. [54]

    Bose, S. et al. Neuromorphic behaviour in discontinuous metal films. Nanoscale Horizons 7, 437 (2022)

    work page 2022

  55. [55]

    Carstens, N. et al. Brain-like critical dynamics and long- range temporal correlations in percolating networks of silver nanoparticles and functionality preservation after integration of insulating matrix. Nanoscale Adv. DOI: 10.1039/d2na00121g (2022)

    work page doi:10.1039/d2na00121g 2022

  56. [56]

    & Ricciardi, C

    Milano, G., Cultrera, A., Boarino, L., Callegaro, L. & Ricciardi, C. Tomography of memory engrams in self- organizing nanowire connectomes. Nat. Commun. 14, 5723 (2023)

    work page 2023

  57. [57]

    & Brown, S

    Sattar, A., Fostner, S. & Brown, S. A. Quantized conductance and switching in percolating nanoparticle 15/20 films. Phys. Rev. Lett. 111, 136808, DOI: 10.1103/ PhysRevLett.111.136808 (2013)

    work page 2013

  58. [58]

    Shirai, S. et al. Long-range temporal correlations in scale-free neuromorphic networks. Netw. Neurosci. 4, 432 (2020)

    work page 2020

  59. [59]

    Steel, J. K. et al. A comparison of the properties of 2d and 3d percolating networks of nanoparticles. Phys Rev E in press (2025)

    work page 2025

  60. [60]

    Y .et al

    Le, P. Y .et al. Electroformed, self-connected tin ox- ide nanoparticle networks for electronic reservoir com- puting. Adv. Electron. Mater. 6, DOI: 10.1002/aelm. 202000081 (2020)

    work page doi:10.1002/aelm 2020

  61. [61]

    Mallinson, J. B. et al. Experimental demonstration of reservoir computing with self-assembled percolating networks of nanoparticles. Adv. Mater. 36, DOI: 10. 1002/adma.202402319 (2024)

    work page 2024

  62. [62]

    Schulze, M. et al. Electrical measurements of nanoscale bismuth cluster films. Eur. Phys. J. D 24, 291 (2003)

    work page 2003

  63. [63]

    Minnai, C., Bellacicca, A., Brown, S. A. & Mi- lani, P. Facile fabrication of complex networks of memristive devices. Sci. Reports 7, DOI: 10.1038/ s41598-017-08244-y (2017)

    work page 2017

  64. [64]

    & Milani, P

    Minnai, C., Mirigliano, M., Brown, S. & Milani, P. The nanocoherer: an electrically and mechanically reset- table memristor based on gold clusters assembled on paper. Nano Futur.2, 011002, DOI: 10.1088/2399-1984/ aab4ee (2018)

    work page doi:10.1088/2399-1984/ 1984

  65. [65]

    Profumo, F. et al. Highly efficient classification of time- series based on resistive switching cluster-assembled ma- terials. Adv. Intell. Syst. DOI: 10.1002/aisy.202401150 (2025)

    work page doi:10.1002/aisy.202401150 2025

  66. [66]

    S., Mondal, I., Bannur, B

    Rao, T. S., Mondal, I., Bannur, B. & Kulkarni, G. U. A scalable solution recipe for a ag-based neu- romorphic device. Discov. Nano 18, DOI: 10.1186/ s11671-023-03906-5 (2023)

    work page 2023

  67. [67]

    Gronenberg, O. et al. In situ imaging of dynamic cur- rent paths in a neuromorphic nanoparticle network with critical spiking behavior. Adv. Funct. Mater. 34, DOI: 10.1002/adfm.202312989 (2024)

    work page doi:10.1002/adfm.202312989 2024

  68. [68]

    Pal, M. et al. Device area invariant conductance lin- earity in scalable silver nanostructure based neuromor- phic devices with threshold activation following noci- ceptive behavior. Nanoscale 17, 19434–19446, DOI: 10.1039/d5nr01617g (2025)

    work page doi:10.1039/d5nr01617g 2025

  69. [69]

    tip of the tongue

    Pal, M., Vidhyadhiraja, N. S. & Kulkarni, G. U. Realis- ing bio-synapse analogous paired-pulse facilitation (ppf) and release inactivation (ri) and emulating “tip of the tongue” experience through conductance and retention dynamics in ag nano-labyrinth-based neuromorphic de- vice. Adv. Funct. Mater.DOI: 10.1002/adfm.202425635 (2025)

    work page doi:10.1002/adfm.202425635 2025

  70. [70]

    Adejube, B. et al. Silver-based self-organized resistive switching nanoparticle networks with neural-like spik- ing behavior: Implications for neuromorphic computing. ACS Appl. Nano Mater. DOI: 10.1021/acsanm.5c02520 (2025)

    work page doi:10.1021/acsanm.5c02520 2025

  71. [71]

    van der Ree, A. J. T., Ahmadi, M., Ten Brink, G. H., Kooi, B. J. & Palasantzas, G. Neuromorphic-like dynam- ics in percolating copper nanoparticle networks: Synthe- sis, characterization, and limitations. Phys. Rev. Mater. 9, DOI: 10.1103/physrevmaterials.9.036001 (2025)

    work page doi:10.1103/physrevmaterials.9.036001 2025

  72. [72]

    van der Ree, A. J. T., Ahmadi, M., Ten Brink, G. H., Kooi, B. J. & Palasantzas, G. Stable millivolt range resistive switching in percolating molybdenum nanopar- ticle networks. ACS Appl. Mater. Interfaces 16, 65157–65164, DOI: 10.1021/acsami.4c12051 (2024)

    work page doi:10.1021/acsami.4c12051 2024

  73. [73]

    & Olin, H

    Olsen, M., Hummelgård, M. & Olin, H. Surface modifi- cations by field induced diffusion. PLoS One 7, e30106 (2012)

    work page 2012

  74. [74]

    Mallinson, J. et al. Reservoir computing using networks of memristors: effects of topology and heterogeneity. Nanoscale 15, 9663–9674 (2023)

    work page 2023

  75. [75]

    Pike, M. D. et al. Atomic scale dynamics drive brain- like avalanches in percolating nanostructured networks. Nano Lett. 20, 3935 (2020)

    work page 2020

  76. [76]

    Acharya, S. K. et al. Stochastic spiking behaviour in neuromorphic devices enables true random number gen- eration. ACS Appl. Mater. Interfaces 13, 52861 (2021). ACS Editors’ Choice Article Selection

    work page 2021

  77. [77]

    Srinivasa, N., Stepp, N. D. & Cruz-Albrecht, J. Critical- ity as a set-point for adaptive behavior in neuromorphic hardware. Front. Neurosci. 9, DOI: 10.3389/fnins.2015. 00449 (2015)

    work page doi:10.3389/fnins.2015 2015

  78. [78]

    Muñoz, M. A. Colloquium: Criticality and dynamical scaling in living systems. Rev. Mod. Phys. 90, DOI: 10.1103/revmodphys.90.031001 (2018)

    work page doi:10.1103/revmodphys.90.031001 2018

  79. [79]

    Milano, G. et al. Quantum conductance in memristive devices: fundamentals, developments, and applications. Adv. Mater. 34, 2201248 (2022)

    work page 2022

  80. [80]

    Mallinson, J. B. & Brown, S. A. Time-multiplexed reser- voir computing with percolating networks of nanoparti- cles. In Proceedings of the International Joint Confer- ence on Neural Networks, DOI: 10.1109/IJCNN54540. 2023.10191253 (Gold Coast, Australia, 2023)

    work page doi:10.1109/ijcnn54540 2023

Showing first 80 references.