Pith. sign in

REVIEW 4 major objections 5 minor 68 references

EVOS: Efficient Implicit Neural Training via EVOlutionary Selector

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read EVOS claims that selecting an evolutionary subset of coordinates each step cuts implicit-neural-representation training time 48–66% while keeping or improving quality — sparsification as implicit sample weighting.

desk verdict EVOS is a solid, incremental INR acceleration method whose default-configuration speedups are plausible, but the paper's unqualified 'consistently enhances' claim is contradicted by its own β=0.3 and β=0.05 experiments. read the letter →

arxiv 2412.10153 v3 pith:LNZNQNER submitted 2024-12-13 cs.CV cs.MMcs.NE

Weixiang Zhang , Shuzhao Xie , Chengwei Ren , Siyi Xie , Chen Tang , Shijia Ge , Mingzi Wang , Zhi Wang This is my paper · ORCID
classification cs.CVcs.MMcs.NE
keywords implicitneuralrepresentationtrainingaccelerationcoordinateselectionevolutionaryalgorithmspectralbiassparseforwardpassessampleweighting
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

Implicit neural representations — MLPs that map coordinates to signal values — are flexible but slow to train, because every iteration normally evaluates every coordinate. This paper proposes EVOS, which each step feeds the network only a subset of coordinates chosen by an evolutionary process: samples with the largest current reconstruction error are the 'fittest' survivors, they are recombined by a frequency-guided crossover, and a small random mutation keeps the selection honest. The paper's central claim is that this strategic sparsification cuts training time by roughly half (48–66% across settings) while matching or exceeding full-data reconstruction quality, so the acceleration is not a compromise but an improvement. If the claim holds, INR-based compression, novel-view synthesis, and physics simulation could run at about twice the speed without sacrificing fidelity.

What carries the argument

The load-bearing object is the evolved coordinate subset $z_t \subset x$ that replaces the full coordinate set in each forward pass, built as $z_t = M(x'_t \odot x''_t)$ where $\odot$ is crossover and $M$ is mutation. Selection is driven by the sparse fitness function $f_t(x) = \Gamma(t)\, D(F_\theta(x), y) + (1-\Gamma(t))\,\hat f(x)$ (Eq. 1): a full ranking of coordinates by reconstruction error is computed only at key iterations, with spacing shrinking linearly through $\Gamma(t)$ in Eq. 2, and the ranking is cached in between. To counter spectral bias, two parent sets are drawn per Eq. 3 — low-frequency parents ranked by square error and high-frequency parents ranked by Laplacian error — and crossed by a balancer $\Psi$ that allocates offspring slots according to the ratio $p = f^{\mathrm{low}}_t/(f^{\mathrm{low}}_t + f^{\mathrm{high}}_t)$ (Eqs. 4–5). A cross-frequency loss (Eq. 6) applies the Laplacian to predictions on $z_t$ augmented with cached predictions for unselected coordinates, since the spatial convolution filter needs the full volume of attributes. Augmented unbiased mutation (Eq. 7) then injects $\alpha k$ uniformly random non-survivors with $\alpha = 0.5$, preventing the deterministic cached rankings from entrenching selection bias.

What would settle it

Run EVOS on a signal whose hardest coordinates move quickly — a video whose sharp edges translate between frames — under the default evaluation schedule, comparing per-iteration PSNR with full-coordinate training at matched wall-clock time. If the selected subset stops tracking the true top-error set and reconstruction quality falls below the full-data baseline, the claim that sparse selection consistently beats full data is refuted; the paper's own statistic $G_t(\tau,\sigma)$ (Eq. 12) gives the direct measurement, and freezing the cache ($\Gamma(t) \equiv 0$) after some iteration shows exactly when the ranked coordinates diverge from the true hardest points.

Watch

Extended reading notes

Core claim

The central claim is that strategically dropping training samples does not trade away quality: it systematically improves per-iteration performance while cutting wall-clock cost. In direct comparison on 3×256 SIREN over DIV2K images, EVOS without its cross-frequency supervision reaches 37.49 dB PSNR at 5,000 iterations versus 37.10 dB for full-coordinate training while performing roughly half the forward passes, and the full EVOS reaches 37.81–38.43 dB depending on scheduler; time-to-target reductions are 47.7%, 59.4%, and 66.5% for 25, 30, and 35 dB PSNR. The paper explains the gain through sample weighting: reweighting coordinates by their current error implicitly regularizes the loss, and this benefit carries over from generalization tasks to the pure overfitting regime of signal fitting.

Load-bearing premise

The load-bearing premise is that the error ranking cached at the last key iteration stays accurate enough, between evaluations, to identify the coordinates the network currently fits worst, and that injecting 50% random points each step is enough to keep any staleness in that cache from steering training astray.

Editorial extensions

If this is right

  • Wall-clock training time drops by roughly half across all tested backbones (SIREN, PEMLP, WIRE, GAUSS, FINER) and network sizes, with the relative gain growing as the network gets larger, so existing INR applications inherit the speedup without changing architecture.
  • The sample-weighting view implies that 'use every training point' is not automatically the right default for signal fitting: per-iteration quality can rise when hard coordinates are over-sampled, even though fitting a signal is a pure overfitting task with no test set.
  • The same evolutionary selection transfers across modalities — 2D images (DIV2K and Kodak), 1D audio (LibriSpeech), 2D text, and 3D signed-distance shapes — with consistent efficiency gains, so the mechanism is not tuned to a single signal type.
  • Evaluating fitness too often ($\tau = 1$ in Eq. 2) costs time and degrades final quality, so sparse cached evaluation is not merely an engineering shortcut but part of what keeps training stable.
  • Selection overhead is only 1.34% of training time, so nearly all of the 48–66% saving appears in the bottom line, placing EVOS ahead of the closest prior method (INT), which shows unstable early-training quality.

Reading between the lines

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

  • Because the fitness ranking is just current per-coordinate error, the benefit may come mostly from the cache-plus-mutation scheme rather than the evolutionary framing; a stripped-down baseline that ranks coordinates by cached error and injects random points could isolate that, which the paper does not run.
  • The paper's own statistic $G_t(\tau,\sigma)$ (Eq. 12) could drive an adaptive evaluation schedule: when the measured top-error set shifts faster than the linear schedule assumes, key iterations should arrive sooner — a testable extension the paper leaves implicit.
  • Dense-coordinate fits where the hard set moves over time, such as physics-informed networks on collocation points or dynamic scenes, are natural next targets, since the same reweighting argument should apply there; nothing in the paper tests this.
  • The fixed hyperparameters — mutation ratio $\alpha = 0.5$ and the crossover balance $p$ — are tuned empirically, and an adaptive version that derives them from the gap between cached and fresh fitness could extend the gains to loss landscapes that change faster than the default schedule.
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 proposes EVOS, a training-time coordinate-selection method for implicit neural representations. At each iteration EVOS selects a subset of coordinates using sparse fitness evaluation with cached losses, a frequency-guided crossover based on low- and high-frequency fitness, and a mutation step that re-injects random coordinates; the network is trained with a cross-frequency loss on the selected subset. Experiments on image, audio, text, and SDF fitting tasks report 48%--66% training-time reductions relative to full-coordinate training, with PSNR/SSIM improvements at the default beta=0.5/step-wise configuration, and the paper claims that strategic sparsification 'consistently enhances' INR training.

Significance. EVOS is a simple and practical idea, and the paper contains a substantial experimental effort: multiple backbones, network sizes, modalities, ablations of each component, comparisons with four sampling-based baselines, and released code. The default configuration (constant beta=0.5 or the step-wise scheduler of INT) delivers acceleration close to the advertised range and quality at least matching or exceeding standard training in most reported settings. These strengths make the method worth pursuing. The contribution is nevertheless overstated: the headline 'consistently enhances' claim is contradicted by the authors' own beta=0.3 and beta=0.05 experiments, and the empirical evaluation lacks error bars and relies on reimplemented baselines.

major comments (4)
  1. [Abstract; Sec. 1; Sec. 4.2; App. H, Table 16] The unqualified claim that strategic sparsification 'consistently enhances' training performance is contradicted by the manuscript's own results. In Appendix H, Table 16 (beta=0.3 constant scheduler), EVOS reaches 36.24 dB at 5k iterations versus 37.10 dB for Standard; the appendix text explicitly concedes that at 30% selection EVOS 'fails to surpass standard training performance.' In addition, Table 12 (beta=0.05) reports EVOS at 29.74 dB, below Uniform (30.14 dB) and Soft Mining (30.73 dB) and far below the Standard reference of 37.10 dB from Table 1. The Abstract, Sec. 1, Sec. 4.2 ('Sample Weighting ALSO Benefits Overfitting'), and Sec. 5 therefore need to be revised to state that the benefit is conditional on the selection ratio/scheduler, and the operating regime (e.g., beta >= 0.5 with constant, or step-wise) should be characterized explicitly. The 48%-66% acceleration claim at default settings is not refuted by this comment, but the current framing is.
  2. [Sec. 3.5; Eq. (6)] The claim in Sec. 3.5 that EVOS achieves acceleration 'without introducing additional storage overhead' is inconsistent with Eq. (6). The cross-frequency loss term Lh_t uses \hat F_theta(x \ z_t), cached full-coordinate network outputs from the last key iteration, and Eq. (1) also caches per-coordinate fitness values. Both require storing O(N) values, which is additional memory relative to standard INR training. The authors should either remove the 'no storage overhead' claim or quantify the memory cost and explain why it is negligible.
  3. [Sec. 4.2; Tables 1-17] All quantitative tables report single runs without error bars or variance estimates. Given the consistent-improvement claim, this is a load-bearing issue: several margins are small (e.g., Table 4 at 10k iterations: 38.71 vs 38.60 dB; Table 1 at 5k: SSIM 0.962 vs 0.964 for Standard), and the comparison includes baselines EGRA and Expan. that were reimplemented by the authors because official implementations are unavailable (Supp. F). Please report multiple seeds or state and justify determinism, and provide confidence intervals or seed-level tables for the main comparisons.
  4. [Sec. 3.2; Appendix I] The design of Gamma(t) is motivated by the G_t curves in Appendix I, but the connection between G_t and the accuracy of cached fitness rankings is never made quantitative. Eq. (1) relies on cached values from the last key iteration; Table 5 shows that increasing evaluation frequency (tau=1, tau=10) actually degrades performance, which the paper attributes to instability. This counterintuitive result needs a more direct analysis: for example, measure how often the top-k selected set changes between key iterations, or ablate cache staleness while controlling for evaluation cost. Without this, the sparse-evaluation schedule remains an unexplained empirical choice rather than a validated mechanism.
minor comments (5)
  1. [Sec. 3.2, Eq. (2)] With tau=100 and theta=0.01, the interval tau - theta*t/T changes from 100 to 99.99 over T=5000 iterations, which cannot produce the 'more frequent evaluations as training progresses' described in Sec. 3.2. Please check the formula or the reported hyperparameter.
  2. [Sec. 3.3, Eqs. (4)-(5)] The crossover operation is not fully specified: it is not clear how the balancer Psi guarantees that exactly k offspring are produced, and the notation U_p l and U_(1-p)l is undefined. Please spell out the sampling procedure.
  3. [Table 1] The caption and table layout make it difficult to see which rows use the constant vs. step-wise scheduler; consider splitting into two tables or adding a scheduler column.
  4. [Abstract; Sec. 4.1] The 'link' for code in the Abstract and Sec. 4.1 is not a URL; provide the actual repository address.
  5. [Sec. 4.1] It is not stated whether the DIV2K and Kodak numbers are averages over a set of images or results on a single image; if averages, report the number of images and the variance.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; EVOS claims are empirical and benchmarked against external baselines.

full rationale

The paper's central claims are empirical: EVOS accelerates INR training by selecting a subset of coordinates via loss-based fitness, crossover, and mutation, and the reported 48%-66% time reduction and PSNR gains are measured against Standard training and external baselines (INT, Soft Mining, EGRA, Expansive Supervision, Uniform). No equation in the paper reduces a prediction to a fitted input. The fitness function f(x) = D(F_theta(x), y) is the training loss itself, but using hardest-sample selection to define the training set is a methodological choice, not a derived claim that the selection mechanism must improve accuracy by construction; the improvement is an experimentally tested outcome. The key-iteration schedule Gamma(t) in Eq. 2 is motivated by the measured G_t curves in Appendix I, but those curves are diagnostic observations, not fitted parameters used to produce the reported numbers, and the paper also ablates the schedule (Table 5) rather than treating it as a theorem. The two self-citations ([63] Expansive Supervision and [64] power transformation) appear as baselines or as related-work categories, not as load-bearing justification for EVOS's validity; the main comparison against INT [61] and full-data Standard training is external and reproducible from the paper's tables. The inconsistency between the unqualified 'consistently enhances' claim and the paper's own beta=0.3 and beta=0.05 results (Tables 16 and 12) is a correctness/over-claiming concern, not circularity: the method's benefit is conditional on selection ratio, but that does not mean any result is equivalent to its inputs by construction. Therefore no circular step meets the evidentiary standard required by the task.

Assumptions & free parameters 6 free parameters · 3 assumptions · 0 invented entities

The method's performance rests on several hand-set hyperparameters (evaluation interval, mutation ratio, selection ratio, loss weights) and on standard domain assumptions about spectral bias and error-based selection. No new entities are introduced, and no derivation is forced by a fitted target.

free parameters (6)
  • tau (key evaluation interval base) = 100
    Controls how often fitness is recomputed; larger means less overhead but staler cache. Set by hand based on experiments.
  • theta (linear schedule coefficient) = 0.01
    Controls the rate at which evaluation intervals shrink over training. Chosen to match the observed linear drift of the fitness distribution.
  • alpha (mutation ratio) = 0.5
    Fraction of extra random coordinates added to the offspring set; tuned to balance exploration and selection bias.
  • beta (constant selection ratio) = 0.5
    Fraction of coordinates used under the constant scheduler; also the initial and final values of the step-wise scheduler are taken from INT (0.2N to N).
  • lambda_l (low-frequency loss weight) = 1.0
    Weight on the L2 reconstruction term in the cross-frequency loss.
  • lambda_h (high-frequency loss weight) = 1e-5
    Weight on the Laplacian loss term; chosen to balance the scale difference between the two terms.
assumptions (3)
  • domain assumption Neural networks trained with coordinate MLPs exhibit spectral bias, preferring low-frequency content.
    Used in Sec. 3.3 to justify the frequency-guided crossover. Supported by the cited literature (Rahaman et al., 2019) and by the paper's own figures.
  • domain assumption Coordinates with larger reconstruction error are the most useful ones to train on next.
    This is the core selection heuristic behind Eq. 3 and is standard in hard mining. The paper validates it through the ablation w/o Eval., which collapses early.
  • standard math The Laplacian operator applied to predicted and target attributes measures high-frequency discrepancy.
    Used in Eq. 3 and Eq. 6 to define a high-frequency fitness function and loss. This is a standard differential operator identity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of EVOS: Efficient Implicit Neural Training via EVOlutionary Selector." pith.science (2026). https://pith.science/paper/LNZNQNER

@misc{pith2026241210153,
  author       = {Pith},
  title        = {Pith review of: EVOS: Efficient Implicit Neural Training via EVOlutionary Selector},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LNZNQNER}},
  note         = {Machine review of arXiv:2412.10153}
}
read the original abstract

We propose EVOlutionary Selector (EVOS), an efficient training paradigm for accelerating Implicit Neural Representation (INR). Unlike conventional INR training that feeds all samples through the neural network in each iteration, our approach restricts training to strategically selected points, reducing computational overhead by eliminating redundant forward passes. Specifically, we treat each sample as an individual in an evolutionary process, where only those fittest ones survive and merit inclusion in training, adaptively evolving with the neural network dynamics. While this is conceptually similar to Evolutionary Algorithms, their distinct objectives (selection for acceleration vs. iterative solution optimization) require a fundamental redefinition of evolutionary mechanisms for our context. In response, we design sparse fitness evaluation, frequency-guided crossover, and augmented unbiased mutation to comprise EVOS. These components respectively guide sample selection with reduced computational cost, enhance performance through frequency-domain balance, and mitigate selection bias from cached evaluation. Extensive experiments demonstrate that our method achieves approximately 48%-66% reduction in training time while ensuring superior convergence without additional cost, establishing state-of-the-art acceleration among recent sampling-based strategies.

Figures

Figures reproduced from arXiv: 2412.10153 by the authors.

Figure 1
Figure 1. Overview of EVOS framework. The proposed method aims to optimize a MLP for implicit signal representation via evolution￾ary selector. The process comprises three key components: (1) (Sparse) Fitness Evaluation (Sec. 3.2) for efficiently guiding coordinate selection, (2) (Frequency-Guided) Crossover (Sec. 3.3) for improving performance by balancing frequency domain preferences, and (3) (Augmented Unbiased) Mutation (… view at source ↗
Figure 2
Figure 2. Visualization of Constant (Top) and Step-wise (Bottom) [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. Comparison of reconstruction quality (PSNR & SSIM) [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (6 more)
Figure 5
Figure 5. Figure 5: Visualization of performance improvements across different backbone architectures when integrated EVOS. All experiments are [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Visual comparison of Mel spectrogram reconstructions [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Visual comparison of reconstruction error for 1D audio [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Visual comparison for 2D text fitting task. We em [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: Visual comparison of 3D shape fitting with fixed 90- [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10 [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

68 extracted references · 62 canonical work pages

  1. [1]

    https : / / graphics

    The Stanford 3D Scanning Repository. https : / / graphics . stanford . edu / data / 3Dscanrep/,

    work page

  2. [2]

    Ntire 2017 challenge on single image super-resolution: Dataset and study

    Eirikur Agustsson and Radu Timofte. Ntire 2017 challenge on single image super-resolution: Dataset and study. In Pro- ceedings of the IEEE conference on computer vision and pat- tern recognition workshops, pages 126–135, 2017. 5, 2

    work page 2017

  3. [3]

    SketchINR: A First Look into Sketches as Implicit Neural Representations

    Hmrishav Bandyopadhyay, Ayan Kumar Bhunia, Pinaki Nath Chowdhury, Aneeshan Sain, Tao Xiang, Timothy M. Hospedales, and Yi-Zhe Song. Sketchinr: A first look into sketches as implicit neural representations. CoRR, abs/2403.09344, 2024. 1, 2

    work page Pith review arXiv 2024

  4. [4]

    Batch normalization alleviates the spectral bias in coor- dinate networks

    Zhicheng Cai, Hao Zhu, Qiu Shen, Xinran Wang, and Xun Cao. Batch normalization alleviates the spectral bias in coor- dinate networks. In CVPR, pages 25160–25171. IEEE, 2024. 3

    work page 2024

  5. [5]

    Deep local shapes: Learning local sdf priors for detailed 3d reconstruction

    Rohan Chabra, Jan E Lenssen, Eddy Ilg, Tanner Schmidt, Julian Straub, Steven Lovegrove, and Richard Newcombe. Deep local shapes: Learning local sdf priors for detailed 3d reconstruction. In Computer Vision–ECCV 2020: 16th Eu- ropean Conference, Glasgow, UK, August 23–28, 2020, Pro- ceedings, Part XXIX 16, pages 608–625. Springer, 2020. 2

    work page 2020

  6. [6]

    Tensorf: Tensorial radiance fields

    Anpei Chen, Zexiang Xu, Andreas Geiger, Jingyi Yu, and Hao Su. Tensorf: Tensorial radiance fields. InEuropean con- ference on computer vision, pages 333–350. Springer, 2022. 1, 2, 3

    work page 2022

  7. [7]

    Joint implicit neural represen- tation for high-fidelity and compact vector fonts

    Chia-Hao Chen, Ying-Tian Liu, Zhifei Zhang, Yuan-Chen Guo, and Song-Hai Zhang. Joint implicit neural represen- tation for high-fidelity and compact vector fonts. In ICCV, pages 5515–5525. IEEE, 2023. 2

    work page 2023

  8. [8]

    Nerv: Neural representations for videos

    Hao Chen, Bo He, Hanyu Wang, Yixuan Ren, Ser-Nam Lim, and Abhinav Shrivastava. Nerv: Neural representations for videos. In NeurIPS, pages 21557–21568, 2021. 2

    work page 2021

Show all 68 references
  1. [9]

    Bidirectional multi-scale implicit neural representations for image derain- ing

    Xiang Chen, Jinshan Pan, and Jiangxin Dong. Bidirectional multi-scale implicit neural representations for image derain- ing. In CVPR, pages 25627–25636. IEEE, 2024. 1

    2024

  2. [10]

    Transformers as meta- learners for implicit neural representations

    Yinbo Chen and Xiaolong Wang. Transformers as meta- learners for implicit neural representations. In ECCV (17), pages 170–187. Springer, 2022. 1, 2

    2022

  3. [11]

    Dauphin, Razvan Pascanu, C ¸ aglar G ¨ulc ¸ehre, KyungHyun Cho, Surya Ganguli, and Yoshua Bengio

    Yann N. Dauphin, Razvan Pascanu, C ¸ aglar G ¨ulc ¸ehre, KyungHyun Cho, Surya Ganguli, and Yoshua Bengio. Iden- tifying and attacking the saddle point problem in high- dimensional non-convex optimization. InNIPS, pages 2933– 2941, 2014. 2

    2014

  4. [12]

    Coin: Compression with implicit neural representations

    Emilien Dupont, Adam Goli ´nski, Milad Alizadeh, Yee Whye Teh, and Arnaud Doucet. Coin: Compression with implicit neural representations. arXiv preprint arXiv:2103.03123 ,

    arXiv

  5. [13]

    Emilien Dupont, Hyunjik Kim, S. M. Ali Eslami, Danilo Jimenez Rezende, and Dan Rosenbaum. From data to functa: Your data point is a function and you can treat it like one. In ICML, pages 5694–5725. PMLR, 2022. 1, 2

    2022

  6. [14]

    Kodak dataset, 1999

    E.Kodak. Kodak dataset, 1999. 1

    1999

  7. [15]

    Egra-nerf: Edge-guided ray allocation for neural radiance fields

    Zhenbiao Gai, Zhenyang Liu, Min Tan, Jiajun Ding, Jun Yu, Mingzhao Tong, and Junqing Yuan. Egra-nerf: Edge-guided ray allocation for neural radiance fields. Image and Vision Computing, 134:104670, 2023. 1, 3, 5, 6, 2, 4, 7

    2023

  8. [16]

    Model-agnostic random weighting for out-of-distribution generalization

    Yue He, Pengfei Tian, Renzhe Xu, Xinwei Shen, Xingxuan Zhang, and Peng Cui. Model-agnostic random weighting for out-of-distribution generalization. In KDD, pages 1050–

  9. [17]

    Adaptation in natural and artificial sys- tems: an introductory analysis with applications to biology, control, and artificial intelligence

    John H Holland. Adaptation in natural and artificial sys- tems: an introductory analysis with applications to biology, control, and artificial intelligence. MIT press, 1992. 1

    1992

  10. [18]

    Compressing multidi- mensional weather and climate data into neural networks

    Langwen Huang and Torsten Hoefler. Compressing multidi- mensional weather and climate data into neural networks. In ICLR. OpenReview.net, 2023. 1, 2

    2023

  11. [19]

    Inverting the imaging process by learning an implicit camera model

    Xin Huang, Qi Zhang, Ying Feng, Hongdong Li, and Qing Wang. Inverting the imaging process by learning an implicit camera model. In CVPR, pages 21456–21465. IEEE, 2023. 1, 2

    2023

  12. [20]

    Jones, J

    Mark W. Jones, J. Andreas Bærentzen, and Milos Sr ´amek. 3d distance fields: A survey of techniques and applications. IEEE Trans. Vis. Comput. Graph., 12(4):581–599, 2006. 2

    2006

  13. [21]

    Accelerating neural field training via soft mining

    Shakiba Kheradmand, Daniel Rebain, Gopal Sharma, Hossam Isack, Abhishek Kar, Andrea Tagliasacchi, and Kwang Moo Yi. Accelerating neural field training via soft mining. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 20071– 20080, 20...

    2024

  14. [22]

    Scal- able neural video representations with learnable positional features

    Subin Kim, Sihyun Yu, Jaeho Lee, and Jinwoo Shin. Scal- able neural video representations with learnable positional features. In NeurIPS, 2022. 2

    2022

  15. [23]

    Lanzend ¨orfer and Roger Wattenhofer

    Luca A. Lanzend ¨orfer and Roger Wattenhofer. Siamese SIREN: audio compression with implicit neural representa- tions. CoRR, abs/2306.12957, 2023. 1

    2023 arXiv

  16. [24]

    Tay- lor, Mathias Unberath, Ming-Yu Liu, and Chen-Hsuan Lin

    Zhaoshuo Li, Thomas M ¨uller, Alex Evans, Russell H. Tay- lor, Mathias Unberath, Ming-Yu Liu, and Chen-Hsuan Lin. Neuralangelo: High-fidelity neural surface reconstruction. In CVPR, pages 8456–8465. IEEE, 2023. 2

    2023

  17. [25]

    Lindell, Dave Van Veen, Jeong Joon Park, and Gor- don Wetzstein

    David B. Lindell, Dave Van Veen, Jeong Joon Park, and Gor- don Wetzstein. Bacon: Band-limited coordinate networks for multiscale scene representation. In CVPR, pages 16231– 16241. IEEE, 2022. 2

    2022

  18. [26]

    Partition speeds up learning implicit neural rep- resentations based on exponential-increase hypothesis

    Ke Liu, Feng Liu, Haishuai Wang, Ning Ma, Jiajun Bu, and Bo Han. Partition speeds up learning implicit neural rep- resentations based on exponential-increase hypothesis. In ICCV, pages 5451–5460. IEEE, 2023. 1, 2

    2023

  19. [27]

    Neural sparse voxel fields

    Lingjie Liu, Jiatao Gu, Kyaw Zaw Lin, Tat-Seng Chua, and Christian Theobalt. Neural sparse voxel fields. In NeurIPS,

  20. [28]

    FINER: flexible spectral-bias tuning in implicit neural represen- tation by variable-periodic activation functions

    Zhen Liu, Hao Zhu, Qi Zhang, Jingde Fu, Weibing Deng, Zhan Ma, Yanwen Guo, and Xun Cao. FINER: flexible spectral-bias tuning in implicit neural represen- tation by variable-periodic activation functions. CoRR, abs/2312.02434, 2023. 2, 5, 7, 1

    2023 arXiv

  21. [29]

    Maiya, Sharath Girish, Max Ehrlich, Hanyu Wang, Kwot Sin Lee, Patrick Poirson, Pengxiang Wu, Chen Wang, and Abhinav Shrivastava

    Shishira R. Maiya, Sharath Girish, Max Ehrlich, Hanyu Wang, Kwot Sin Lee, Patrick Poirson, Pengxiang Wu, Chen Wang, and Abhinav Shrivastava. NIRV ANA: neural implicit representations of videos with adaptive networks and au- toregressive patch-wise modeling. In CVPR, pages 1437...

    2023

  22. [30]

    Julien N. P. Martel, David B. Lindell, Connor Z. Lin, Eric R. Chan, Marco Monteiro, and Gordon Wetzstein. Acorn: adap- tive coordinate networks for neural scene representation. ACM Trans. Graph., 40(4):58:1–58:13, 2021. 1, 2

    2021

  23. [31]

    Mod- ulated periodic activations for generalizable local functional representations

    Ishit Mehta, Micha ¨el Gharbi, Connelly Barnes, Eli Shecht- man, Ravi Ramamoorthi, and Manmohan Chandraker. Mod- ulated periodic activations for generalizable local functional representations. In ICCV, pages 14194–14203. IEEE, 2021. 3

    2021

  24. [32]

    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 syn- thesis. In ECCV (1), pages 405–421. Springer, 2020. 1, 2, 3

    2020

  25. [33]

    Instant neural graphics primitives with a multires- olution hash encoding

    Thomas M ¨uller, Alex Evans, Christoph Schied, and Alexan- der Keller. Instant neural graphics primitives with a multires- olution hash encoding. ACM Trans. Graph. , 41(4):102:1– 102:15, 2022. 1, 2, 3

    2022

  26. [34]

    Evolutionary dynamic optimization: A survey of the state of the art

    Trung Thanh Nguyen, Shengxiang Yang, and Juergen Branke. Evolutionary dynamic optimization: A survey of the state of the art. Swarm and Evolutionary Computation , 6:1–24, 2012. 2

    2012

  27. [35]

    Librispeech: An ASR corpus based on public domain audio books

    Vassil Panayotov, Guoguo Chen, Daniel Povey, and Sanjeev Khudanpur. Librispeech: An ASR corpus based on public domain audio books. In ICASSP, pages 5206–5210. IEEE,

  28. [36]

    Dogyun Park, Sihyeon Kim, Sojin Lee, and Hyunwoo J. Kim. DDMI: domain-agnostic latent diffusion models for synthesizing high-quality implicit neural representations. CoRR, abs/2401.12517, 2024. 2

    2024 arXiv

  29. [37]

    Florence, Julian Straub, Richard A

    Jeong Joon Park, Peter R. Florence, Julian Straub, Richard A. Newcombe, and Steven Lovegrove. Deepsdf: Learning continuous signed distance functions for shape rep- resentation. In CVPR, pages 165–174. Computer Vision Foundation / IEEE, 2019. 1, 2

    2019

  30. [38]

    Yang, Zachary DeVito, Mar- tin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, Junjie Bai, and Soumith Chintala

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zem- ing Lin, Natalia Gimelshein, Luca Antiga, Alban Desmai- son, Andreas K ¨opf, Edward Z. Yang, Zachary DeVito, Mar- tin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit S...

    2019

  31. [39]

    Hamprecht, Yoshua Bengio, and Aaron C

    Nasim Rahaman, Aristide Baratin, Devansh Arpit, Felix Draxler, Min Lin, Fred A. Hamprecht, Yoshua Bengio, and Aaron C. Courville. On the spectral bias of neural networks. In ICML, pages 5301–5310. PMLR, 2019. 1, 2

    2019

  32. [40]

    Karniadakis

    Maziar Raissi, Paris Perdikaris, and George E. Karniadakis. Physics-informed neural networks: A deep learning frame- work for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys. , 378:686–707, 2019. 1, 2

    2019

  33. [41]

    Beyond periodicity: Towards a unifying framework for activations in coordinate- mlps

    Sameera Ramasinghe and Simon Lucey. Beyond periodicity: Towards a unifying framework for activations in coordinate- mlps. In ECCV (33), pages 142–158. Springer, 2022. 2, 5, 7, 1

    2022

  34. [42]

    Derf: Decom- posed radiance fields

    Daniel Rebain, Wei Jiang, Soroosh Yazdani, Ke Li, Kwang Moo Yi, and Andrea Tagliasacchi. Derf: Decom- posed radiance fields. In CVPR, pages 14153–14161. Com- puter Vision Foundation / IEEE, 2021. 2

    2021

  35. [43]

    Kilonerf: Speeding up neural radiance fields with thousands of tiny mlps

    Christian Reiser, Songyou Peng, Yiyi Liao, and Andreas Geiger. Kilonerf: Speeding up neural radiance fields with thousands of tiny mlps. InICCV, pages 14315–14325. IEEE,

  36. [44]

    Rix, John G

    Antony W. Rix, John G. Beerends, Michael P. Hollier, and Andries P. Hekstra. Perceptual evaluation of speech qual- ity (pesq)-a new method for speech quality assessment of telephone networks and codecs. In ICASSP, pages 749–752. IEEE, 2001. 1

    2001

  37. [45]

    Baraniuk, and Ashok Veeraraghavan

    Vishwanath Saragadam, Jasper Tan, Guha Balakrishnan, Richard G. Baraniuk, and Ashok Veeraraghavan. MINER: multiscale implicit neural representation. In ECCV (23) , pages 318–333. Springer, 2022. 1, 2

    2022

  38. [46]

    Bara- niuk

    Vishwanath Saragadam, Daniel LeJeune, Jasper Tan, Guha Balakrishnan, Ashok Veeraraghavan, and Richard G. Bara- niuk. WIRE: wavelet implicit neural representations. In CVPR, pages 18507–18516. IEEE, 2023. 2, 5, 7

    2023

  39. [47]

    In search of a data transformation that accelerates neural field training

    Junwon Seo, Sangyoon Lee, Kwang In Kim, and Jaeho Lee. In search of a data transformation that accelerates neural field training. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024. 3

    2024

  40. [48]

    In- ductive gradient adjustment for spectral bias in implicit neu- ral representations, 2024

    Kexuan Shi, Hai Chen, Leheng Zhang, and Shuhang Gu. In- ductive gradient adjustment for spectral bias in implicit neu- ral representations, 2024. 3

    2024

  41. [49]

    Improved im- plicity neural representation with fourier bases reparameter- ized training

    Kexuan Shi, Xingyu Zhou, and Shuhang Gu. Improved im- plicity neural representation with fourier bases reparameter- ized training. CoRR, abs/2401.07402, 2024. 3

    2024 arXiv

  42. [50]

    Vincent Sitzmann, Julien N. P. Martel, Alexander W. Bergman, David B. Lindell, and Gordon Wetzstein. Implicit neural representations with periodic activation functions. In NeurIPS, 2020. 1, 2, 3, 5, 7

    2020

  43. [51]

    Implicit neural representations for image compression

    Yannick Str ¨umpler, Janis Postels, Ren Yang, Luc Van Gool, and Federico Tombari. Implicit neural representations for image compression. In ECCV (26), pages 74–91. Springer,

  44. [52]

    Piczak, Przemyslaw Spurek, Jacek Tabor, and Tomasz Trzcinski

    Filip Szatkowski, Karol J. Piczak, Przemyslaw Spurek, Jacek Tabor, and Tomasz Trzcinski. Hypernetworks build implicit neural representations of sounds. InECML/PKDD (4), pages 661–676. Springer, 2023. 1, 3

    2023

  45. [53]

    Taal, Richard C

    Cees H. Taal, Richard C. Hendriks, Richard Heusdens, and Jesper Jensen. An algorithm for intelligibility prediction of time-frequency weighted noisy speech. IEEE Trans. Speech Audio Process., 19(7):2125–2136, 2011. 1

    2011

  46. [54]

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

    Matthew Tancik, Pratul P. Srinivasan, Ben Mildenhall, Sara Fridovich-Keil, Nithin Raghavan, Utkarsh Singhal, Ravi Ra- mamoorthi, Jonathan T. Barron, and Ren Ng. Fourier fea- tures let networks learn high frequency functions in low di- mensional domains. In NeurIPS, 2020. 1, 2, 5, 7

    2020

  47. [55]

    Srinivasan, Jonathan T

    Matthew Tancik, Ben Mildenhall, Terrance Wang, Divi Schmidt, Pratul P. Srinivasan, Jonathan T. Barron, and Ren Ng. Learned initializations for optimizing coordinate-based neural representations. In CVPR, pages 2846–2855. Com- puter Vision Foundation / IEEE, 2021. 1, 2

    2021

  48. [56]

    Bovik, Hamid R

    Zhou Wang, Alan C. Bovik, Hamid R. Sheikh, and Eero P. Simoncelli. Image quality assessment: from error visibility to structural similarity. IEEE Trans. Image Process., 13(4): 600–612, 2004. 5

    2004

  49. [57]

    DINER: disorder-invariant implicit neu- ral representation

    Shaowen Xie, Hao Zhu, Zhen Liu, Qi Zhang, You Zhou, Xun Cao, and Zhan Ma. DINER: disorder-invariant implicit neu- ral representation. In CVPR, pages 6143–6152. IEEE, 2023. 1, 2, 3

    2023

  50. [58]

    Point-nerf: Point- based neural radiance fields

    Qiangeng Xu, Zexiang Xu, Julien Philip, Sai Bi, Zhixin Shu, Kalyan Sunkavalli, and Ulrich Neumann. Point-nerf: Point- based neural radiance fields. In CVPR, pages 5428–5438. IEEE, 2022. 3

    2022

  51. [59]

    TINC: tree-structured implicit neural com- pression

    Runzhao Yang. TINC: tree-structured implicit neural com- pression. In CVPR, pages 18517–18526. IEEE, 2023. 3

    2023

  52. [60]

    A survey on evolutionary computation for complex continuous optimization

    Zhi-Hui Zhan, Lin Shi, Kay Chen Tan, and Jun Zhang. A survey on evolutionary computation for complex continuous optimization. Artificial Intelligence Review , 55(1):59–110,

  53. [61]

    Nonparametric teaching of implicit neural representations

    Chen Zhang, Steven Tin Sui Luo, Jason Chun Lok Li, Yik- Chung Wu, and Ngai Wong. Nonparametric teaching of implicit neural representations. In Forty-first International Conference on Machine Learning, ICML 2024, Vienna, Aus- tria, July 21-27, 2024. OpenReview.net, 2024. 1, 2, ...

    2024

  54. [62]

    Efros, Eli Shecht- man, and Oliver Wang

    Richard Zhang, Phillip Isola, Alexei A. Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In CVPR, pages 586–

  55. [63]

    Expansive supervision for neural radi- ance field

    Weixiang Zhang, Shuzhao Xie, Shijia Ge, Wei Yao, Chen Tang, and Zhi Wang. Expansive supervision for neural radi- ance field. arXiv preprint arXiv:2409.08056, 2024. 1, 3, 5, 6, 2, 4, 7

    2024 arXiv

  56. [64]

    Enhancing implicit neural rep- resentations via symmetric power transformation, 2024

    Weixiang Zhang, Shuzhao Xie, Chengwei Ren, Shijia Ge, Mingzi Wang, and Zhi Wang. Enhancing implicit neural rep- resentations via symmetric power transformation, 2024. 3

    2024

  57. [65]

    Deep stable learning for out-of- distribution generalization

    Xingxuan Zhang, Peng Cui, Renzhe Xu, Linjun Zhou, Yue He, and Zheyan Shen. Deep stable learning for out-of- distribution generalization. In CVPR, pages 5372–5382. Computer Vision Foundation / IEEE, 2021. 7

    2021

  58. [66]

    RINAS: training with dataset shuffling can be general and fast

    Tianle Zhong, Jiechen Zhao, Xindi Guo, Qiang Su, and Ge- offrey Fox. RINAS: training with dataset shuffling can be general and fast. CoRR, abs/2312.02368, 2023. 2, 5 EVOS: Efficient Implicit Neural Training via EVOlutionary Selector Supplementary Material A. 1D Audio Fitting T...

    2023 arXiv

  59. [595]

    Computer Vision Foundation / IEEE Computer Society,

  60. [2007]

    Accessed: 2024-11. 2

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.