Pith. sign in

REVIEW 4 major objections 5 minor 38 references

Neural Field Turing Machine: A Differentiable Spatial Computer

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · 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 the Neural Field Turing Machine—a neural controller with movable local read/write heads over a continuous spatial field—is Turing complete under bounded error and unifies symbolic, physical, and perceptual computation.

desk verdict An honest architecture mashup whose central Turing-completeness claim does not hold on the finite-grid issue; the demos are too small to compensate. read the letter →

arxiv 2509.03370 v1 pith:TKJYAOHW submitted 2025-08-27 cs.NE cs.AI

Akash Malhotra , Nacéra Seghouani This is my paper · ORCID
classification cs.NEcs.AI MSC 68Q0568Q1068Q8068T07
keywords NeuralFieldTuringMachinecompletenessRule110cellularautomatacontinuousspatialmemoryPDEsolvingimageinpaintingdifferentiablecomputation
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 introduces the Neural Field Turing Machine (NFTM), a differentiable architecture in which a neural controller reads local patches of a continuous spatial field, computes updates, writes them back, and moves read/write heads. Its central claim is that this single architecture is Turing complete under bounded error, via a reduction to the universal cellular automaton Rule 110. The authors argue this matters because the same substrate can express exact symbolic computation (cellular automata), continuous physical simulation (the 2D heat equation), and iterative perceptual refinement (CIFAR-10 inpainting), with per-step cost linear in field size. The demonstrations show learned controllers reproducing Rule 110 over 100 steps, recovering global and spatially varying diffusion coefficients, and improving inpainting quality when rolled out beyond the training horizon.

What carries the argument

The central object is the coupled update pair $f_{t+1}(x) = g(\int A_t(x,y) f_t(y)\,dy)$ and $h_{t+1} = h_t + \Delta h_t$, in which the controller reads a local patch $f_t[S(h_t)]$ and emits a spatial attention field $A_t(x,y)$ plus a head displacement $\Delta h_t$. The support region $S(h_t)$ is the paper's named device for choosing the read/write neighborhood, from a small ball to the whole field. The straight-through estimator keeps training differentiable while forcing Boolean field values: rounded values in the forward pass, unrounded gradients in the backward pass. Together these components allow one architecture to host both exact symbolic rules and continuous field physics.

What would settle it

Run a trained NFTM Rule-110 controller for an unbounded number of steps from a random binary initial field, comparing every site against true Rule 110 at each step; a single mismatched site, or a mismatch rate that grows with rollout length, refutes the exactness on which the bounded-error Turing-completeness claim rests.

Watch

Extended reading notes

Core claim

The paper's load-bearing statement is Proposition 1: NFTMs are Turing complete under bounded error. The argument restricts the support region to a radius-1 local neighborhood, discretizes the continuous field with straight-through-estimator binarization, and claims the neural controller can learn Rule 110's Boolean transition function exactly; since Rule 110 is Turing complete, NFTM inherits universality. The bounded-error qualifier acknowledges that quantization must recover the symbolic dynamics with arbitrarily small error. The same framework is then instantiated as a cellular-automaton simulator, a PDE solver that recovers diffusion coefficients, and an iterative image refiner, with each controller learning local rules whose repeated application produces global behavior.

Load-bearing premise

The central claim rests on the controller, trained with straight-through-estimator rounding, being able to implement Rule 110's Boolean update exactly at every site for every timestep while the rounding error stays bounded over arbitrarily long rollouts; the paper asserts this exactness but does not prove it.

Editorial extensions

If this is right

  • A single differentiable architecture can in principle express both exact algorithmic computation and continuous field dynamics, so tasks that mix discrete logic with spatial reasoning need not switch between separate models.
  • With fixed-radius neighborhoods, per-step cost is $O(N)$ in field size, placing NFTM in the same asymptotic class as convolutional networks and finite-difference solvers.
  • Controllers trained on short rollouts can continue to improve when rolled out further, giving a concrete form of test-time compute scaling for spatial refinement tasks.
  • Because NCA is a special case of NFTM, any existing neural cellular automaton can in principle run inside the framework while gaining explicit controller logic and movable heads.
  • If the bounded-error universality claim holds, it connects continuous neural-field computation to classical computability, implying neural field models can in principle simulate arbitrary algorithms.

Reading between the lines

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

  • Beyond the paper, a rigorous tightening of Proposition 1 would produce an explicit bound on how rounding errors compound with rollout length; without such a bound, the practical guarantee is 'universal up to a finite horizon with small error.'
  • Beyond the paper, a natural next experiment is to train a single controller on all three domains and test whether the same weights transfer rules across tasks, which would distinguish a unified substrate from task-specific instantiations.
  • Beyond the paper, the framework implies a spatial analogue of adaptive computation time: the model could keep refining a field until a confidence threshold is met, spending more steps on hard inputs.
  • Beyond the paper, equivariant controllers that enforce conservation laws are only sketched; a concrete test is to measure long-rollout energy drift with and without translation or rotation equivariance.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The manuscript introduces the Neural Field Turing Machine (NFTM), a differentiable architecture consisting of a neural controller, a continuous spatial memory field, and movable read/write heads. The central formal claim is Proposition 1: NFTMs are Turing complete under bounded error, argued by reduction to the cellular automaton Rule 110. The paper also reports three proof-of-concept instantiations: learning Rule 110 and Conway's Game of Life truth tables, recovering global and spatially varying diffusion coefficients in 2D heat equation rollouts, and iterative image inpainting on CIFAR-10. The authors position NFTM as a unifying differentiable substrate for symbolic, physical, and perceptual computation, with linear scaling in field size for fixed-radius updates and with NCA as a special case.

Significance. If the Turing completeness claim were correct, NFTM would be a noteworthy conceptual contribution: a single differentiable architecture that connects discrete algorithmic computation with continuous field dynamics, while retaining linear per-step cost and the ability to learn local update rules. The manuscript is also honest about the exploratory nature of its experiments and makes code available. However, the load-bearing theoretical result is not established: as defined, the model is a finite-state system, so the claimed universality does not follow from the Rule 110 reduction. The empirical demonstrations are illustrative but lack error bars, baselines, and reported training/evaluation splits, so they do not compensate for the unsupported formal claim. The central advertised novelty therefore rests on an invalid argument, which makes the contribution, as written, not acceptable.

major comments (4)
  1. [Section 3, Proposition 1] The Turing completeness claim is invalid because the NFTM is defined over a finite spatial field. Rule 110 is universal only on an unbounded tape with a suitable infinite background, whereas Appendix A explicitly fixes N as the number of spatial sites and Section 4.1 sets the number of heads equal to the number of cells, treating N as a finite constant. A deterministic Boolean system on N cells has at most 2^N reachable configurations and is eventually periodic; it cannot simulate computations that require more than N cells of memory or recognize non-regular predicates. The manuscript supplies no mechanism for an unbounded or growing field and no quantification of N as a function of the simulated computation. Therefore, even granting that the controller implements Rule 110 exactly, Proposition 1's conclusion does not follow under the model as defined.
  2. [Section 3, Proposition 1] The qualifier 'under bounded error' is never quantified. No epsilon, no horizon-dependent error bound, and no error metric are provided. The text asserts that quantization 'ensures symbolic dynamics can be recovered with arbitrarily small error,' but this is an assertion, not a proof. The paper does not show that the continuous-field approximation remains within a bounded error over unbounded rollouts, nor does it specify whether the bound is on a single step or on the entire trajectory. Without such a bound, the phrase 'Turing complete under bounded error' has no precise meaning, and the reduction cannot be evaluated.
  3. [Section 5.2 and 5.3] The reported diffusion coefficient recoveries are single numbers without error bars, number of seeds, or a statement of train/test splits. Because the controller and the recovered alpha are fitted on the same rollouts that are used to evaluate recovery, the reported MAE and PSNR values are fitting results rather than evidence of independent identification. For example, the global alpha values in Section 5.2 (0.067 for true 0.05, then 0.100, 0.150, 0.200) are presented as exact numbers, and Section 5.3 reports a single mean PSNR of 40.89 dB. The paper should report means and standard deviations over repeated initializations and, crucially, evaluate on held-out trajectories or at least state explicitly whether any holdout exists.
  4. [Section 5.4 and Figure 5] The inpainting result is reported as a single PSNR curve with no confidence intervals, no baselines, and no test-set size. The claim that PSNR 'improves monotonically' from 15.2 dB to 24.5 dB cannot be assessed without variance information, and the paper does not specify whether the curve is an average over a fixed test set or a single example. Since the authors explicitly decline head-to-head comparisons in Section 5.5, the inpainting experiment may be acceptable as a proof of concept, but the claim of monotonic improvement and generalization beyond the training horizon should be supported with error bars and a clear evaluation protocol.
minor comments (5)
  1. [Abstract and Section 1] 'Turing complete under bounded error' is used in the abstract and introduction before being defined; a precise definition should appear before Proposition 1, including the error metric and the quantification over time horizon and field size.
  2. [Section 4.1] The paper states the controller learns Rule 110 and Conway's Game of Life, but the results section only shows Rule 110; the Game of Life experiment is not reported, so the claim should be either removed or accompanied by experimental evidence.
  3. [Equation (7)] The heteroscedastic loss notation is incomplete: the variables δg_t, σ, β, and γ are not all defined before or immediately after the equation, and the relationship between δg_t and the earlier α∇²u is unclear.
  4. [Figure 5 caption] 'PSNR improves monotonically' should be phrased as 'non-decreasing' unless strict monotonicity is guaranteed, and the figure would benefit from error bars or shaded confidence intervals.
  5. [Section 5.1] The text says 'the task is ultimately trivial—Rule 110 reduces to a finite truth table'; this directly undercuts the Turing completeness claim in Section 3, because a finite truth table is not universal. The authors should acknowledge that the experiment demonstrates finite-horizon approximation only, not universality.

Circularity Check

1 steps flagged · score 5.0 of 10

The heat-equation 'recovery' of diffusion coefficients reports the loss's own fitted parameter; the Turing-completeness proof is an external reduction and not circular.

  1. fitted input called prediction [Section 5.2 and Eq. (7); parallel claim in Section 5.3]
    "For α = 0.05, the model learns 0.067 (absolute error 0.018). For α = 0.10, 0.15, and 0.20, the learned values are 0.100, 0.150, and 0.200, respectively, with negligible errors (≤ 0.001). This demonstrates that NFTM can reliably infer physical constants from spatio-temporal rollouts."

    The diffusion coefficient α is exactly the free parameter optimized in Eq. (7): L_NLL = ½(δ_gt − α L_phys(u))²/σ² + β/2 log σ². Minimizing this objective fits α to the same training transitions δ_gt used for evaluation; no held-out split or forward-prediction benchmark is reported. The reported MAE/PSNR therefore measure training fit, and calling the fitted α 'recovered' or 'inferred' restates the fit by construction. Proposition 1 is separate: it is an external reduction to Cook's Rule 110 theorem, not a fitted quantity.

full rationale

The paper's central theoretical claim, Proposition 1, is an external reduction: NFTM's discretized local update is asserted to replicate Rule 110, whose universality is cited to Cook (2004). That is a genuine import of an independent theorem, not a self-citation or a definitional equivalence, although its correctness on a finite field is a separate soundness concern outside the circularity question. No load-bearing self-citation, imported uniqueness theorem, or ansatz-by-citation was found. The one circular element is experimental: in the heat-equation demonstrations, the diffusion coefficient α appears directly in the training loss (Eq. 7) as the parameter being optimized, and the reported 'recovered' values are those fitted values measured on the same rollouts. The paper presents this fitting result as evidence that NFTM can 'reliably infer physical constants,' which is a fitted input renamed as a predictive capability; without a held-out split or forward-prediction check, that evidence is forced by construction. The cellular-automata and inpainting demos are rollouts rather than fitted-parameter predictions, though the inpainting evaluation also lacks an explicit train/test split for its horizon-generalization claim. Overall, the main derivation is not circular, but one reported result is a fit masquerading as inference, giving a moderate circularity score.

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

The central claims rely on the external Rule 110 theorem and on unproven assumptions about the controller's ability to implement exact local rules with bounded quantization error. The empirical recovery of diffusion coefficients is a fitting exercise on the training rollouts, adding a circularity burden. No new physical entities are introduced.

free parameters (4)
  • Global diffusion coefficient alpha = 0.067 for true 0.05; 0.100, 0.150, 0.200 for true 0.10, 0.15, 0.20
    Learned from the same PDE rollouts used to train the controller (Section 5.2); no held-out split specified.
  • Spatially varying diffusion field alpha(x,y) = Central square 0.15 on background 0.05, smoothed by 3x3 average pool
    Recovered via heteroscedastic loss (Sections 4.2 and 5.3); it is a fitted parameter field, not a prediction on unseen data.
  • Heteroscedastic loss weights (beta, sigma, gamma) = Not quantified in text
    Hand-chosen hyperparameters in Eq. (7); values not reported.
  • TV weight lambda_TV and step-size beta for inpainting = Not quantified
    Hand-chosen in Eq. (8) and Eq. (9); schedules described qualitatively.
assumptions (4)
  • standard math Rule 110 is Turing complete
    External theorem by Cook (2004), cited as [8]; used in Proposition 1.
  • domain assumption A neural controller can represent the Boolean transition function of Rule 110 with bounded error
    Assumed in Proposition 1; no construction or error bound is given.
  • domain assumption STE binarization preserves differentiability and yields discrete dynamics
    Used in Eq. (6) for cellular automata; the quantization error is asserted to be arbitrarily small but not derived.
  • domain assumption The field and heads can be constrained to a fixed local neighborhood radius r=1
    Required for the Rule 110 reduction in Section 3.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Neural Field Turing Machine: A Differentiable Spatial Computer." pith.science (2026). https://pith.science/paper/TKJYAOHW

@misc{pith2026250903370,
  author       = {Pith},
  title        = {Pith review of: Neural Field Turing Machine: A Differentiable Spatial Computer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TKJYAOHW}},
  note         = {Machine review of arXiv:2509.03370}
}
read the original abstract

We introduce the Neural Field Turing Machine (NFTM), a differentiable architecture that unifies symbolic computation, physical simulation, and perceptual inference within continuous spatial fields. NFTM combines a neural controller, continuous memory field, and movable read/write heads that perform local updates. At each timestep, the controller reads local patches, computes updates via learned rules, and writes them back while updating head positions. This design achieves linear O(N) scaling through fixed-radius neighborhoods while maintaining Turing completeness under bounded error. We demonstrate three example instantiations of NFTM: cellular automata simulation (Rule 110), physics-informed PDE solvers (2D heat equation), and iterative image refinement (CIFAR-10 inpainting). These instantiations learn local update rules that compose into global dynamics, exhibit stable long-horizon rollouts, and generalize beyond training horizons. NFTM provides a unified computational substrate bridging discrete algorithms and continuous field dynamics within a single differentiable framework.

Figures

Figures reproduced from arXiv: 2509.03370 by the authors.

Figure 1
Figure 1. Neural Field Turing Machine (NFTM). The system maintains a spatial field ft at time t, which evolves over discrete timesteps. Read/write heads ht specify positions within this field, and each update is computed over a local support region centered at ht . In the simplest case this is a ball of radius r, S(ht) = {x ∈ X : ∥x − ht∥ ≤ r}, (1) but more generally S(ht) can be defined by an arbitrary locality kernel K(x, h… view at source ↗
Figure 2
Figure 2. NFTM reproduces Rule 110 cellular automaton dynamics. Left: ground-truth evolution. Right: [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Learning global diffusion coefficient α with heteroscedastic loss. Left: learned vs. true coefficients. Right: absolute error per case. NFTM recovers α with mean absolute error below 0.01 in most cases. 5.3 Heat Equation: Variable Diffusion Coefficient NFTM also handles spatially varying coefficients α(x, y). In this experiment, α(x, y) was set to 0.05 everywhere except in a central square region where it was 0.15, … view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Learning spatially varying diffusion coefficient [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Step-wise average PSNR on CIFAR-10 inpainting task. NFTM exhibits monotonic improvement, [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Qualitative image inpainting results on CIFAR-10. Left to right: ground truth, initialization, and [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 24 canonical work pages

  1. [1]

    Battaglia, Jessica B

    Peter W. Battaglia, Jessica B. Hamrick, and Joshua B. Tenenbaum. Simulation as an engine of physical scene understanding. Proceedings of the National Academy of Sciences, 110(45):18327–18332, 2013. doi: 10.1073/pnas.1306572110

    doi:10.1073/pnas.1306572110 2013

  2. [2]

    Scheduled sampling for sequence prediction with recurrent neural networks

    Samy Bengio, Oriol Vinyals, Navdeep Jaitly, and Noam Shazeer. Scheduled sampling for sequence prediction with recurrent neural networks. In Advances in Neural Information Processing Systems (NeurIPS), pages 1171–1179, 2015

    work page 2015

  3. [3]

    On a theory of computation and complexity over the real numbers: NP-completeness, recursive functions and universal machines

    Lenore Blum, Michael Shub, and Steve Smale. On a theory of computation and complexity over the real numbers: NP-completeness, recursive functions and universal machines. In Ciprian Foias and Georg Sell, editors, Complexity of Computation, pages 170–195. Springer, 1989

    work page 1989

  4. [4]

    Bronstein, Joan Bruna, Taco Cohen, and Petar Veliˇckovi´c

    Michael M. Bronstein, Joan Bruna, Taco Cohen, and Petar Veliˇckovi´c. Geometric deep learning: Grids, groups, graphs, geodesics, and gauges. arXiv preprint arXiv:2104.13478, 2021

    arXiv 2021

  5. [5]

    Gabriel B´ena, Maxence Faldor, Dan F. M. Goodman, and Antoine Cully. A path to universal neural cellular automata. arXiv preprint, 2025. arXiv:2505.13058. 11

    work page Pith review arXiv 2025

  6. [6]

    Ricky T. Q. Chen, Yulia Rubanova, Jesse Bettencourt, and David K. Duvenaud. Neural ordinary differential equations. In NeurIPS, 2018

    work page 2018

  7. [7]

    Cohen and Max Welling

    Taco S. Cohen and Max Welling. Group equivariant convolutional networks. In Proceedings of the 33rd International Conference on Machine Learning (ICML), volume 48 of Proceedings of Machine Learning Research, pages 2990–2999. PMLR, 2016

    work page 2016

  8. [8]

    Universality in elementary cellular automata

    Matthew Cook. Universality in elementary cellular automata. Complex Systems, 15(1):1–40, 2004

    work page 2004

Show all 38 references
  1. [9]

    Generalizing convolutional neural networks for equivariance to lie groups

    Marc Finzi, Max Welling, et al. Generalizing convolutional neural networks for equivariance to lie groups. In International Conference on Machine Learning (ICML), 2020

    2020

  2. [10]

    Adaptive computation time for recurrent neural networks

    Alex Graves. Adaptive computation time for recurrent neural networks. In International Conference on Learning Representations (ICLR), 2016. arXiv:1603.08983

    2016 arXiv

  3. [11]

    Neural turing machines.arXiv preprint arXiv:1410.5401, 2014

    Alex Graves, Greg Wayne, and Ivo Danihelka. Neural turing machines.arXiv preprint arXiv:1410.5401, 2014

    2014 arXiv

  4. [12]

    Hybrid computing using a neural network with dynamic external memory

    Alex Graves, Greg Wayne, Malcolm Reynolds, Tim Harley, Ivo Danihelka, Agnieszka Grabska- Barwi´nska, Sergio G ´omez Colmenarejo, Edward Grefenstette, Tiago Ramalho, John Agapiou, Adri`a Puigdom`enech Badia, Karl Moritz Hermann, Yori Zwols, Georg Ostrovski, Adam Cain, Helen Kin...

    2016 doi

  5. [13]

    Hamiltonian neural networks

    Samuel Greydanus, Misko Dzamba, and Jascha Yosinski. Hamiltonian neural networks. In NeurIPS, 2019

    2019

  6. [14]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. In NeurIPS, 2020. arXiv preprint arXiv:2006.11239

    2020 arXiv

  7. [15]

    J¨orn-Henrik Jacobsen, Arnold W. M. Smeulders, and Edouard Oyallon. i-revnet: Deep invertible networks. In International Conference on Learning Representations (ICLR) , 2018. URL https: //openreview.net/forum?id=HJsjkMb0Z

    2018

  8. [16]

    Kingma and Prafulla Dhariwal

    Diederik P. Kingma and Prafulla Dhariwal. Glow: Generative flow with invertible 1x1 convolutions. In Advances in Neural Information Processing Systems (NeurIPS), volume 31, pages 10236–10245, 2018

    2018

  9. [17]

    Control-nerf: Editable feature volumes for scene rendering and manipulation

    Verica Lazova, Vladimir Guzov, Kyle Olszewski, Sergey Tulyakov, and Gerard Pons-Moll. Control-nerf: Editable feature volumes for scene rendering and manipulation. arXiv preprint arXiv:2204.10850, 2022

    2022 arXiv

  10. [18]

    Randall J. LeVeque. Finite Difference Methods for Ordinary and Partial Differential Equations: Steady-State and Time-Dependent Problems. Society for Industrial and Applied Mathematics (SIAM), Philadelphia, 2007. ISBN 978-0-898716-29-0

    2007

  11. [19]

    Latent ODEs for irregularly-sampled time series

    Yulia Li et al. Latent ODEs for irregularly-sampled time series. NeurIPS, 2020

    2020

  12. [20]

    Fourier neural operator for parametric partial differential equations

    Zongyi Li, Nikola Kovachki, Kamyar Azizzadenesheli, Burigede Liu, Kaushik Bhattacharya, Andrew Stuart, and Anima Anandkumar. Fourier neural operator for parametric partial differential equations. arXiv preprint arXiv:2010.08895, 2020

    2010 arXiv

  13. [21]

    Deep operator networks (deeponets) for learning nonlinear operators

    Lu Lu, Pengzhan Jin, and George Em Karniadakis. Deep operator networks (deeponets) for learning nonlinear operators. arXiv preprint arXiv:1910.03193, 2019. 12

    1910 arXiv

  14. [22]

    Srinivasan, Matthew Tancik, Jonathan T

    Ben Mildenhall, Pratul P. Srinivasan, Matthew Tancik, Jonathan T. Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. In ECCV, 2020

    2020

  15. [23]

    Growing neural cellular automata

    Alexander Mordvintsev, Ettore Randazzo, Eyvind Niklasson, and Michael Levin. Growing neural cellular automata. Distill, 2020. doi: 10.23915/distill.00023. URL https://distill.pub/ 2020/growing-ca

    2020 doi

  16. [24]

    Neural cellular automata: From cells to pixels

    Ehsan Pajouheshgar, Yitao Xu, Ali Abbasi, Alexander Mordvintsev, Wenzel Jakob, and Sabine S¨usstrunk. Neural cellular automata: From cells to pixels. arXiv preprint, 2025. arXiv:2506.22899

    2025 arXiv

  17. [25]

    Barron, Sofien Bouaziz, Dan B

    Keunhong Park, Utkarsh Sinha, Jonathan T. Barron, Sofien Bouaziz, Dan B. Goldman, Steven M. Seitz, and Ricardo Martin-Brualla. Nerfies: Deformable neural radiance fields. arXiv preprint arXiv:2011.12948, 2020

    2011 arXiv

  18. [26]

    Attention is turing complete

    Jorge P´erez, Pablo Barcel´o, and Javier Marinkovi´c. Attention is turing complete. Journal of Machine Learning Research, 22(75):1–35, 2021

    2021

  19. [27]

    D-nerf: Neural radiance fields for dynamic scenes

    Albert Pumarola, Antonio Agudo, Wolfgang Heidrich, and Guillermo Vidal. D-nerf: Neural radiance fields for dynamic scenes. In CVPR, 2021

    2021

  20. [28]

    Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations

    Maziar Raissi, Paris Perdikaris, and George Em Karniadakis. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378:686–707, 2019

    2019

  21. [29]

    Battaglia

    Alvaro Sanchez-Gonzalez, Jonathan Godwin, Tobias Pfaff, Rex Ying, Jure Leskovec, and Peter W. Battaglia. Learning to simulate complex physics with graph networks. In Proceedings of the 37th International Conference on Machine Learning (ICML), volume 119 of PMLR, page 869–877, ...

    2020 arXiv

  22. [30]

    Claude E. Shannon. Mathematical theory of the differential analyzer. Journal of Mathematics and Physics, 20:337–354, 1941

    1941

  23. [31]

    Shepard and Jacqueline Metzler

    Roger N. Shepard and Jacqueline Metzler. Mental rotation of three-dimensional objects. Science, 171 (3972):701–703, 1971. doi: 10.1126/science.171.3972.701

    1971 doi

  24. [32]

    Martel, Julien N

    Vincent Sitzmann, P. Martel, Julien N. Alexander W. Bergman, David B. Lindell, and Gordon Wetzstein. Implicit neural representations with periodic activation functions. In NeurIPS, 2020

    2020

  25. [33]

    Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole

    Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. In ICLR, 2021. Outstanding Paper Award

    2021

  26. [34]

    Srinivasan, Ben Mildenhall, Sara Fridovich-Keil, Nithin Raghavan, Utkarsh Singhal, Ravi Ramamoorthi, Jonathan T

    Matthew Tancik, Pratul P. Srinivasan, Ben Mildenhall, Sara Fridovich-Keil, Nithin Raghavan, Utkarsh Singhal, Ravi Ramamoorthi, Jonathan T. Barron, and Ren Ng. Fourier features let networks learn high frequency functions in low-dimensional domains. In NeurIPS, 2020

    2020

  27. [35]

    Gomez,Łukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez,Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. In NeurIPS, 2017. 13 A Complexity Analysis of NFTM We analyze the computational complexity of the Neural Field Turing Machine...

    2017

  28. [36]

    Reading a local neighborhood at head positions

  29. [37]

    Applying the controllerC to compute updates

  30. [38]

    For a neighborhood of fixed sizek, the read/write cost isO(kN ), while the controller addsO(NP )

    Writing updates back into the field. For a neighborhood of fixed sizek, the read/write cost isO(kN ), while the controller addsO(NP ). Thus, the total time complexity is O(N·T· (k +P )), which is effectively linear in field sizeN and rollout lengthT . Memory usage isO(N) for f...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.