Pith. sign in

REVIEW 3 major objections 5 minor 60 references

Autonomous Physical Computation: A Categorical Closure Criterion for Physical and Neuromorphic Reservoirs

T0 review · 3 major / 5 minor · reviewed 2026-07-31 · deepseek-v4-flash

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

Pith's one-line read The paper claims that autonomous physical computation requires an internal physical readout state to select each next operation, and that the wave-particle walker, despite its memory and feedback, fails this closure test.

desk verdict Useful conceptual framework with a real soft spot: the closure criterion as written is too permissive to support the walker classification without an added causal-distinctness axiom. read the letter →

arxiv 2607.23902 v1 pith:HME6LQJH submitted 2026-07-27 cs.ET cond-mat.dis-nnmath-phmath.MP

Nima Dehghani This is my paper · ORCID
classification cs.ETcond-mat.dis-nnmath-phmath.MP
keywords autonomouscomputationclosurecriterionphysicalreservoircomputingwave-memorywalkingdropletcoarse-grainingneuromorphic
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

The paper establishes a sharp boundary between physical memory, transition-preserving computation, and autonomous physical computation. It argues that memory and feedback are not enough; a system computes autonomously only when a physical readout state inside the system selects the next operation. Applying this to the walking-droplet system, the paper shows that the walker genuinely writes, stores, reads, and erases its own wave memory, but its erasing phase shift is externally imposed, so it does not qualify as a closed autonomous physical computer. The criterion turns this classification into a design principle for building closed physical computers from wave-memory reservoirs.

What carries the argument

The load-bearing object is the closure criterion for autonomous physical computation: the closed-loop map Φ = Act ∘ ⟨C ∘ p_Y, id_X⟩, which internalizes morphism selection so that the next physical process is chosen by the system's own readout state rather than by an external agent. This is coupled with a coarse-grained transition-preservation condition (Π ∘ F = G ∘ Π) that distinguishes genuine computation from arbitrary observer labeling.

What would settle it

The decisive test is the paper's own 'autonomous eraser': couple a physical detector of wave energy to the forcing phase so the system triggers its own π-shift. If this modified walker is still classified as not closed, or the original externally triggered walker is classified as closed, the necessity of the closure criterion is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that a physical system computes autonomously exactly when its dynamics takes the form x_{n+1} = F_{C(y_n)}(x_n), where y_n is an internal physical readout state and C maps that state to a physical operation. Under this closure criterion, the wave-particle walker is a wave-memory machine with genuine Turing-like primitives—writing, storage, reading, feedback, and erasure—but not a closed autonomous physical computer, because the phase shift that produces erasure is imposed externally rather than selected by an internal readout.

Load-bearing premise

The definition of autonomy as internal operation selection is a stipulated criterion, not a derived theorem; if autonomy is instead defined as self-organization or homeostatic closure, the walker's classification could change.

Editorial extensions

If this is right

  • The wave-particle walker is classified as a wave-memory physical machine with Turing-like primitives, not a closed autonomous Turing machine.
  • Standard reservoir computing, in which a readout is trained externally and not fed back into the reservoir, does not satisfy closure unless the readout is physically internalized.
  • Physical memory and feedback do not by themselves imply computation; robust coarse-grained transition preservation is a separate requirement that must be demonstrated.
  • Closure is independent of computational power: a closed system can compute nothing of interest, and nontrivial computation requires additional structure such as a finite alphabet and programmable transitions.
  • The framework yields concrete designs for closing the wave-memory loop: multistable readout traps, thresholded wave detectors, boundary-controlled Faraday baths, and coupled-walker controllers.

Reading between the lines

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

  • If the closure criterion is accepted, many neuromorphic and reservoir-computing devices described as 'computing' are technically externally read dynamical systems; their computational status depends on whether readouts are internalized.
  • The criterion suggests an operational benchmark for cortical traveling waves: they contribute to computation only if read out by downstream circuits that in turn select subsequent wave-generating operations.
  • The paper's distinction implies a three-way experimental comparison—external phase schedule, observer readout without feedback, and physical readout with feedback—as a practical test for autonomy in any proposed physical computer.
  • One could extend the criterion to define a hierarchy of autonomous physical computation in which the readout basin itself is learned or tuned, connecting closure with trainability and adaptation.
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.

Referee Report

3 major / 5 minor

Summary. The paper develops a conceptual hierarchy for autonomous physical computation: physical memory, transition-preserving computation via a commuting coarse-graining Π∘F=G∘Π, robust symbolization via separated basins and noise tolerance, and closed autonomous computation via an internal readout state selecting the next physical operation, x_{n+1}=F_{C(y_n)}(x_n). The framework is applied to the Perrard–Fort–Couder walking-droplet system, which is modeled as a stroboscopic reservoir with wave-field memory. The paper concludes that the walker is a wave-memory physical machine with writing, storage, reading, feedback, finite-time reversal, and externally triggered erasure, but not a closed autonomous physical computer, because the erasing π-shift is externally imposed. A categorical reformulation and several constructive designs for closing the loop (autonomous eraser, branch, coupled walkers) are also presented.

Significance. The paper addresses a genuine foundational question: when does physical memory become autonomous computation? Its main strengths are the explicit formalization of computation as quotient-functorial transition preservation, the clean separation of memory, transition preservation, robustness, and closure, and the self-critical application to the walker. Proposition 1 is elementary and correct, and the classification of the walker follows from the stipulated definitions rather than from fitted parameters or circular empirical claims. The paper also makes a useful design point: physical readout states must be coupled back into operation selection. However, the central closure criterion is stipulated, and, as discussed below, it is under-specified in a way that makes the walker classification boundary-relative. If the authors add a formal causal-independence condition, the framework would be substantially more convincing.

major comments (3)
  1. [§VII, Eq. (30); §A5.7; §IX] The closure criterion is under-specified. No formal restrictions are imposed on the operation set O, the readout subsystem X_Y, or the decomposition X=X_R×X_Y×X_Z. After the one-time external π-shift in the stroboscopic model, take X_Y={+1,−1} as the σ-coordinate, O={F_+,F_-} with F_σ the map (1)–(5) with the phase held at σ, and C(σ_n)=F_{σ_n}. Then x_{n+1}=F_{C(σ_n)}(x_n) and |Im(C)|=2, so Eq. (30) and the A5.7 non-vacuity condition hold at every subsequent step. This contradicts the Sec. IX assertion that no internal readout state satisfies C(y_n)=o_π. The contradiction is resolved only by the unformalized choice to exclude F_± from O and σ_n from X_Y. The paper needs an explicit causal-independence axiom: the actuator executing C(y_n) must be a distinct physical channel from the variable reporting y_n, and C(y_n) must not be a mere re-labeling of the readout's own update law. Without
  2. [§IX; §XI.B; Table A7.13] The positive classification 'wave-memory physical machine with genuine Turing-like primitives' is not tied to any formal notion of 'Turing-like'. The paper lists strict Turing-machine requirements (finite alphabet, internal states, transition table, robust encoding) in §XI.B and A7.12, and admits the walker lacks them, yet the classification table credits the walker with 'Turing-like primitives'. Since this phrase appears in the main result, it should either be defined precisely (e.g., a finite set of operations that can compose, under a stated encoding, to Turing-complete transitions) or replaced by neutral language such as 'read/write/store/erase primitives'. As written, the positive claim is stronger than the formalism supports.
  3. [§X, Eqs. (50)–(54)] The constructive closed-loop architectures are presented as the main payoff of the closure criterion, but no quantitative demonstration is given that they satisfy the paper's own robust symbolization conditions: basin separation (22), transition stability (24)–(25), and operation-selection reliability (A36). For example, the autonomous eraser uses hysteresis thresholds Θ_E and Δ, but no noise scale, barrier height, or coupling strength is estimated; the claim that this device 'would test the closure criterion directly' is therefore a design sketch rather than a supported prediction. The authors should either label these explicitly as open modeling proposals or provide elementary feasibility estimates.
minor comments (5)
  1. [§VIII] The sentence 'categorical treatments of information such as ?' contains a missing citation/reference placeholder. This should be completed before publication.
  2. [§A7.13] The classification table in the appendix has no caption and is not referenced from the main text; the reader encounters it only after the prose classification. Add a numbered caption and a cross-reference.
  3. [§XI.B] The distinction between the loose operational sense and the strict computability-theoretic sense of 'Turing machine' is useful, but the loose sense is never formalized. The term 'Turing-like primitives' should be defined in the same section where this distinction is drawn.
  4. [§A3.8 and §A5.4] The symbol U is used both for a domain U⊆X and for an abstract operation set U. This creates avoidable confusion in the appendix; rename one of them.
  5. [General] The paper is extremely long, and the appendices repeat much of the main text nearly verbatim. Condensing the appendices to definitions, proofs, and tables not already in the main text would improve readability without changing content.

Circularity Check

1 steps flagged · score 4.0 of 10

The walker 'not autonomous' classification is carried by the stipulated closure definition's unformalized boundary choice of O and X_Y, a partial self-definitional circularity; the Ref. 19 self-citation is scaffolding, not load-bearing.

  1. self definitional [Sec. VII (Criterion 1, Eq. 30), Sec. IX (Classification), Appendix A5.7 (Eq. A38), Remark 1; walker state in Sec. III (Eqs. 1–5) / A1.1 (x_n=(r_n,v_n,σ_n,H_n))]
    "To claim autonomous physical computation, one must identify an internal physical readout state y_n and an operation-selection map C : X_Y → O such that x_{n+1} = F_{C(y_n)}(x_n). In the walker experiment, the π-shift is imposed externally. There is no internal readout state satisfying C(y_n) = o_π."

    The walker’s own state contains σ_n ∈ X with σ_{n+1}=σ_n (A1.6). Taking y_n=σ_n, X_Y={±1}, O={F_+,F_-} where F_σ is the stroboscopic map with frozen phase, and C(σ_n)=F_{σ_n}, the post-flip dynamics satisfies Eq. (30) with |Im(C)|=2, meeting the A5.7 nontriviality condition (A38). Remark 1 even admits a two-state phase-flipping controller is “closed, transition-preserving, and autonomous.” Thus the Sec. IX claim “there is no internal readout state satisfying C(y_n)=o_π” is not entailed by Criterion 1; it depends on unstated exclusions (σ not counted as X_Y, F_± not counted as distinct operations, forcing apparatus excluded from X). The negative classification therefore reduces by construction to these informal boundary choices; a causal-independence or non-vacuity axiom is needed but never

full rationale

No fitted parameter is renamed as a prediction, and no empirical result is derived from an assumed conclusion: the stroboscopic walker model (Eqs. 1–5) and the memory/erasure analysis (Eqs. 7–17) are self-contained reductions of prior experimental work, with no circularity. The category-theoretic framing is imported from the author’s own Ref. 19, but the closure criterion itself (Criterion 1, Eq. 30) is introduced in this paper, so the self-citation is scaffolding, not load-bearing for the walker classification. The paper is also unusually candid: §XI.I says the criterion is ‘intentionally minimal’ and ‘not an automatic procedure’, and A5.7 concedes that conditional entropy alone cannot certify closure. The one genuine weakness is definitional, not empirical: the criterion stipulates what counts as an operation (O) and a readout subsystem (X_Y) without formalizing causal independence or non-vacuity. Because σ_n is part of the paper’s own state space, the closure equation (30) can be satisfied by declaring y_n=σ_n and C(σ_n)=F_{σ_n}; conversely, enlarging X to include the forcing apparatus internalizes the π-shift. Hence the central classification (‘wave-memory machine but not closed autonomous computer’) is a direct application of a deliberately minimal stipulation and is boundary-relative to the unformalized choice of O and X_Y: a partial self-definitional circularity, not a fitted-input or self-citation-chain circularity. Score 4 reflects this partial definitional carry while recognizing the independent content of the walker decomposition and the erasure-backtracking analysis.

Assumptions & free parameters 0 free parameters · 5 assumptions · 1 invented entities

The ledger shows the central claim rests mainly on a stipulated definition (closure) plus the modeling assumption that the wave field is the only memory readout. There are no fitted parameters; the constructive architectures use illustrative thresholds. Independent evidence for the readout subsystem is absent—it is a design target.

assumptions (5)
  • domain assumption Wave memory updates as H_{n+1}=λH_n+σ_nψ_{r_n} with λ=e^{-1/M_e}, and the droplet reads ∇H_n(r_n).
    Sec. III Eqs. (1)–(5); a discrete reduction of the Oza–Rosales–Bush trajectory equation, not re-derived or experimentally validated here.
  • domain assumption A physically meaningful metric d_X and noise scale ε_X exist, allowing basins separated by dist(B_i,B_j)>2ε_X.
    Sec. VI/A4; required for robust symbolization, but the paper does not supply a concrete metric for the wave-field component beyond existence.
  • ad hoc to paper Physical states decompose as X=X_R×X_Y×X_Z and an operation-selection map C:X_Y→O exists for closed systems.
    Sec. VII Criterion 1; this is the paper's stipulated definition of autonomy, not a theorem or empirical fact.
  • standard math Standard set-theoretic/categorical facts: quotient maps, composition, Galois connections, and the existence of the unique induced map G in Proposition 1.
    Sec. V and Appendix A3/A6; these are standard and unproblematic.
  • domain assumption In the Perrard–Fort–Couder experiment the π-shift is externally imposed and no internal physical state satisfies C(y_n)=o_π.
    Sec. IX and §A7.11; based on the experimental description, this is the empirical premise that makes the classification negative.
invented entities (1)
  • Physical readout-control subsystem X_Y
    purpose: Provides the internal state whose value selects the next operation, making a physical computer autonomous.
    Introduced as a required architectural component; no concrete realization is built. Proposed implementations in Sec. X are hypothetical.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Autonomous Physical Computation: A Categorical Closure Criterion for Physical and Neuromorphic Reservoirs." pith.science (2026). https://pith.science/paper/HME6LQJH

@misc{pith2026260723902,
  author       = {Pith},
  title        = {Pith review of: Autonomous Physical Computation: A Categorical Closure Criterion for Physical and Neuromorphic Reservoirs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HME6LQJH}},
  note         = {Machine review of arXiv:2607.23902}
}
read the original abstract

Physical reservoirs, neuromorphic devices, and wave-mediated systems often possess memory, feedback, and rich state-dependent dynamics, but these properties do not by themselves establish autonomous computation. Here we develop a closure criterion for autonomous physical computation, motivated by the wave--particle walker. We formulate the walker as a stroboscopic reservoir with state, where the wave field stores an exponentially decaying trace of previous droplet impacts and guides future motion through local slope coupling. This model separates physical writing, storage, reading, feedback, and externally triggered erasure. We then define computation as robust coarse-grained transition preservation: a physical map implements an abstract transition only when a coarse-graining satisfies compositionality, with abstract states realized by separated physical basins and transitions stable under noise. Autonomous physical computation requires a further closure condition: an internal physical readout state must select the next physical operation. This criterion classifies the wave--particle walker as a wave-memory machine with genuine Turing-like primitives, but not as a closed autonomous physical computer, because the erasing phase shift is externally imposed. The framework turns this distinction into a design principle: memory becomes autonomous computation when physical readout basins are coupled back to operation selection.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 25 canonical work pages

  1. [1]

    and Fort, E

    Perrard, S. and Fort, E. and Couder, Y. , journal =. Wave-Based. 2016 , month =. doi:10.1103/PhysRevLett.117.094502 , url =

    work page doi:10.1103/physrevlett.117.094502 2016

  2. [2]

    2024 , month =

    Dehghani, Nima and Caterina, Gianluca , title =. 2024 , month =. doi:10.1088/2632-072X/ad6051 , url =

    work page doi:10.1088/2632-072x/ad6051 2024

  3. [3]

    and Proti

    Couder, Y. and Proti. Walking and orbiting droplets , journal=. 2005 , month=. doi:10.1038/437208a , url=

    doi:10.1038/437208a 2005

  4. [4]

    Particle–wave association on a fluid interface , volume=

    Protiere, Suzie and Boudaoud, Arezki and Couder, Yves , year=. Particle–wave association on a fluid interface , volume=. doi:10.1017/S0022112006009190 , journal=

    work page doi:10.1017/s0022112006009190

  5. [5]

    Information stored in

    Eddi, Antonin and Sultan, Eric and Moukhtar, Julien and Fort, Emmanuel and Rossi, Maurice and Couder, Yves , year=. Information stored in. doi:10.1017/S0022112011000176 , journal=

    work page doi:10.1017/s0022112011000176

  6. [6]

    and Rosales, Rodolfo R

    Oza, Anand U. and Rosales, Rodolfo R. and Bush, John W. M. , year=. A trajectory equation for walking droplets: hydrodynamic pilot-wave theory , volume=. doi:10.1017/jfm.2013.581 , journal=

    work page doi:10.1017/jfm.2013.581 2013

  7. [7]

    Interaction of two walkers: Wave-mediated energy and force , author =. Phys. Rev. E , volume =. 2014 , month =. doi:10.1103/PhysRevE.90.063017 , url =

    work page doi:10.1103/physreve.90.063017 2014

  8. [8]

    Pilot-Wave Hydrodynamics

    Bush, John W.M. Pilot-Wave Hydrodynamics. Annual Review of Fluid Mechanics. 2015. doi:https://doi.org/10.1146/annurev-fluid-010814-014506

    doi:10.1146/annurev-fluid-010814-014506 2015

Show all 60 references
  1. [9]

    Self-organization into quantized eigenstates of a classical wave-driven particle , journal=

    Perrard, St. Self-organization into quantized eigenstates of a classical wave-driven particle , journal=. 2014 , month=. doi:10.1038/ncomms4219 , url=

    2014 doi

  2. [10]

    Chaos Driven by Interfering Memory , author =. Phys. Rev. Lett. , volume =. 2014 , month =. doi:10.1103/PhysRevLett.113.104101 , url =

    2014 doi

  3. [11]

    Proceedings of the Royal Society of London

    Benjamin, Thomas Brooke and Ursell, Fritz Joseph , title =. Proceedings of the Royal Society of London. A. Mathematical and Physical Sciences , volume =. 1954 , month =. doi:10.1098/rspa.1954.0218 , url =

    1954

  4. [12]

    Douady, S. , year=. Experimental study of the Faraday instability , volume=. doi:10.1017/S0022112090003603 , journal=

    doi

  5. [13]

    Time-reversed imaging as a diagnostic of wave and particle chaos , author =. Phys. Rev. E , volume =. 1998 , month =. doi:10.1103/PhysRevE.58.5668 , url =

    1998 doi

  6. [14]

    and Pilipetsky, Nikolai F

    Zel'dovich, Boris Ya. and Pilipetsky, Nikolai F. and Shkunov, Vladimir V. , title =. 1985 , edition =. doi:10.1007/978-3-540-38959-0 , isbn =

    1985 doi

  7. [15]

    Robust Acoustic Time Reversal with High-Order Multiple Scattering , author =. Phys. Rev. Lett. , volume =. 1995 , month =. doi:10.1103/PhysRevLett.75.4206 , url =

    1995 doi

  8. [16]

    2001 , month =

    Mathias Fink and Claire Prada , title =. 2001 , month =. doi:10.1088/0266-5611/17/1/201 , url =

    2001 doi

  9. [17]

    One-Channel Time Reversal of Elastic Waves in a Chaotic

    Draeger, Carsten and Fink, Mathias , journal =. One-Channel Time Reversal of Elastic Waves in a Chaotic. 1997 , month =. doi:10.1103/PhysRevLett.79.407 , url =

    1997 doi

  10. [18]

    Time Reversal of Electromagnetic Waves , author =. Phys. Rev. Lett. , volume =. 2004 , month =. doi:10.1103/PhysRevLett.92.193904 , url =

    2004 doi

  11. [19]

    Nature Physics , year=

    Bacot, Vincent and Labousse, Matthieu and Eddi, Antonin and Fink, Mathias and Fort, Emmanuel , title=. Nature Physics , year=. doi:10.1038/nphys3810 , url=

    doi

  12. [20]

    Order and chaos in reversible dynamical systems , journal =

    Chaos and time-reversal symmetry. Order and chaos in reversible dynamical systems , journal =. 1992 , issn =. doi:https://doi.org/10.1016/0370-1573(92)90163-T , url =

    1992 doi

  13. [21]

    Devaney , journal =

    Robert L. Devaney , journal =. Reversible Diffeomorphisms and Flows , urldate =. 1976 , doi =

    1976

  14. [22]

    and Kendon, Viv , title =

    Horsman, Dominic and Stepney, Susan and Wagner, Rob C. and Kendon, Viv , title =. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences , volume =. 2014 , month =. doi:10.1098/rspa.2014.0182 , url =

    2014

  15. [23]

    Abstraction/Representation Theory and the Natural Science of Computation , booktitle=

    Horsman, Dominic and Kendon, Viv and Stepney, Susan , editor=. Abstraction/Representation Theory and the Natural Science of Computation , booktitle=. 2018 , pages=

    2018

  16. [24]

    and Paquette, \'E ric.O

    Coecke, Bob. and Paquette, \'E ric.O. Categories for the Practising Physicist. New Structures for Physics. 2010. doi:10.1007/978-3-642-12821-9_3

    2010 doi

  17. [25]

    Fong, Brendan and Spivak, David I. , year=. An Invitation to Applied Category Theory: Seven Sketches in Compositionality , publisher=

  18. [26]

    1978 , edition =

    Mac Lane, Saunders , title =. 1978 , edition =. doi:10.1007/978-1-4757-4721-8 , isbn =

    1978 doi

  19. [27]

    Turing, A. M. , title =. Proceedings of the London Mathematical Society , volume =. doi:https://doi.org/10.1112/plms/s2-42.1.230 , url =

    doi

  20. [28]

    Journal of Symbolic Logic , author=

    A note on the. Journal of Symbolic Logic , author=. 1936 , pages=. doi:10.2307/2269326 , number=

    1936 doi

  21. [29]

    Proceedings of the Royal Society of London

    Deutsch, David , title =. Proceedings of the Royal Society of London. A. Mathematical and Physical Sciences , volume =. 1985 , month =. doi:10.1098/rspa.1985.0070 , url =

    1985

  22. [30]

    , title=

    Chalmers, David J. , title=. Minds and Machines , year=. doi:10.1007/BF00974166 , url=

    doi

  23. [31]

    , title=

    Chalmers, David J. , title=. Synthese , year=. doi:10.1007/BF00413692 , url=

    doi

  24. [32]

    Dimensions of Mind , editor =

    Putnam, Hilary , title =. Dimensions of Mind , editor =. 1960 , pages =

    1960

  25. [33]

    Robots: Machines or Artificially Created Life? , urldate =

    Hilary Putman and Hilary Putnam , journal =. Robots: Machines or Artificially Created Life? , urldate =

  26. [34]

    1988 , doi =

    Putnam, Hilary , title =. 1988 , doi =

    1988

  27. [35]

    Philosophy of Science , author=

    Computing Mechanisms , volume=. Philosophy of Science , author=. 2007 , pages=. doi:10.1086/522851 , number=

    2007 doi

  28. [36]

    information processing: why their difference matters to cognitive science , journal =

    Computation vs. information processing: why their difference matters to cognitive science , journal =. 2010 , note =. doi:https://doi.org/10.1016/j.shpsa.2010.07.012 , url =

    2010 doi

  29. [37]

    Piccinini, Gualtiero , title =. The. 2025 , edition =

    2025

  30. [38]

    Introduction , booktitle =

    Shagrir, Oron , isbn =. Introduction , booktitle =. 2022 , month =. doi:10.1093/oso/9780197552384.003.0001 , url =

    2022

  31. [39]

    Synthese , year=

    Campbell, Douglas Ian and Yang, Yi , title=. Synthese , year=. doi:10.1007/s11229-019-02275-w , url=

    doi

  32. [40]

    International Journal of Theoretical Physics , year=

    Fredkin, Edward and Toffoli, Tommaso , title=. International Journal of Theoretical Physics , year=. doi:10.1007/BF01857727 , url=

    doi

  33. [41]

    1984 , issn =

    Physics-like models of computation , journal =. 1984 , issn =. doi:https://doi.org/10.1016/0167-2789(84)90252-5 , url =

    1984 doi

  34. [42]

    Computing Inside the Billiard Ball Model

    Durand-Lose, J \'e r \^o me. Computing Inside the Billiard Ball Model. Collision-Based Computing. 2002. doi:10.1007/978-1-4471-0129-1_6

    2002 doi

  35. [43]

    echo state

    Jaeger, Herbert , year=. The “echo state” approach to analysing and training recurrent neural networks , url=. doi:10.24406/publica-fhg-291111 , publisher=

    doi

  36. [44]

    Science , volume =

    Herbert Jaeger and Harald Haas , title =. Science , volume =. 2004 , doi =

    2004

  37. [45]

    Neural Computation , volume =

    Maass, Wolfgang and Natschläger, Thomas and Markram, Henry , title =. Neural Computation , volume =. 2002 , month =. doi:10.1162/089976602760407955 , url =

    2002 doi

  38. [46]

    2019 , issn =

    Recent advances in physical reservoir computing: A review , journal =. 2019 , issn =. doi:https://doi.org/10.1016/j.neunet.2019.03.005 , url =

    2019 doi

  39. [47]

    Physics for neuromorphic computing , journal=

    Markovi. Physics for neuromorphic computing , journal=. 2020 , month=. doi:10.1038/s42254-020-0208-2 , url=

    2020 doi

  40. [48]

    , title=

    Crutchfield, James P. , title=. Nature Physics , year=. doi:10.1038/nphys2190 , url=

    doi

  41. [49]

    , title=

    Shalizi, Cosma Rohilla and Crutchfield, James P. , title=. Journal of Statistical Physics , year=. doi:10.1023/A:1010388907793 , url=

    doi

  42. [50]

    2001 , issn =

    Traveling Electrical Waves in Cortex: Insights from Phase Dynamics and Speculation on a Computational Role , journal =. 2001 , issn =. doi:https://doi.org/10.1016/S0896-6273(01)00178-7 , url =

    2001 doi

  43. [51]

    Muller and Bartosz Telenczuk and Eric Halgren and Sydney Cash and Nicholas G

    Michel Le Van Quyen and Lyle E. Muller and Bartosz Telenczuk and Eric Halgren and Sydney Cash and Nicholas G. Hatsopoulos and Nima Dehghani and Alain Destexhe , title =. Proceedings of the National Academy of Sciences , volume =. 2016 , doi =

    2016

  44. [52]

    Cortical travelling waves: mechanisms and computational principles , journal=

    Muller, Lyle and Chavane, Fr. Cortical travelling waves: mechanisms and computational principles , journal=. 2018 , month=. doi:10.1038/nrn.2018.20 , url=

    2018 doi

  45. [53]

    Anderson and Welling, Max , title =

    Keller, T. Anderson and Welling, Max , title =. Proceedings of the 40th International Conference on Machine Learning , articleno =. 2023 , publisher =

    2023

  46. [54]

    , journal=

    Chua, L. , journal=. Memristor-The missing circuit element , year=

  47. [55]

    Pershin and Massimiliano Di Ventra , title =

    Yuriy V. Pershin and Massimiliano Di Ventra , title =. Advances in Physics , volume =. 2011 , publisher =. doi:10.1080/00018732.2010.544961 , URL =

    2011

  48. [56]

    Frontiers in Computational Neuroscience , VOLUME=

    Dehghani, Nima , TITLE=. Frontiers in Computational Neuroscience , VOLUME=. 2018 , URL=. doi:10.3389/fncom.2018.00081 , ISSN=

    2018

  49. [57]

    PLOS Complex Systems , publisher =

    Systematizing cellular complexity: A. PLOS Complex Systems , publisher =. 2024 , month =. doi:10.1371/journal.pcsy.0000013 , author =

    2024 doi

  50. [58]

    Ashby, W. Ross. Requisite Variety and Its Implications for the Control of Complex Systems. Facets of Systems Science. 1991. doi:10.1007/978-1-4899-0718-9_28

    1991 doi

  51. [59]

    Complexity , volume =

    Bar-Yam, Yaneer , title =. Complexity , volume =. doi:https://doi.org/10.1002/cplx.20014 , url =

    doi

  52. [60]

    2026 , eprint=

    Evolutionary Optimization Reveals Structural Constraints on Reservoir Architecture for Spatiotemporal Chaos , author=. 2026 , eprint=

    2026

Pith tools

Reviewed July 31, 2026 · model on record in the stance chip above.