Pith. sign in

REVIEW 6 major objections 6 minor 33 references

LuKAN: A Kolmogorov-Arnold Network Framework for 3D Human Motion Prediction

T0 review · 6 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a Kolmogorov-Arnold layer using Lucas polynomial activations on wavelet-encoded joint trajectories predicts 3D human motion as accurately as current state-of-the-art baselines while using far fewer parameters.

desk verdict LuKAN is a coherent KAN-plus-DWT architecture worth knowing about, but its superiority claims rest on sub-millimeter MPJPE gaps with no error bars, and the paper overstates its own ablation results. read the letter →

arxiv 2508.04847 v1 pith:UXYKMMPH submitted 2025-08-06 cs.CV

Md Zahidul Hasan , A. Ben Hamza , Nizar Bouguila This is my paper
classification cs.CV
keywords 3DhumanmotionpredictionKolmogorov-ArnoldnetworksLucaspolynomialsdiscretewavelettransformMPJPEposeforecastingefficientarchitecturestemporalmodeling
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 LuKAN, an architecture for forecasting future 3D body poses from a short observed history. It encodes each joint trajectory with the discrete wavelet transform, mixes information across joints with two spatial projections, and models temporal dependencies with a Kolmogorov-Arnold layer whose per-edge functions are Lucas polynomials, before reconstructing poses in the time domain. The central claim is that this compact design matches or beats strong baselines such as SiMLPe on Human3.6M, AMASS, and 3DPW while using far fewer parameters and less compute. The authors further argue that Lucas polynomials evaluate more cheaply than the B-spline functions of standard KANs and are better suited to oscillatory motion. If correct, the work offers an efficient KAN-based alternative to transformer-, GCN-, and deep-MLP motion predictors.

What carries the argument

The central object is the KAN layer inside the Temporal Dependency Learner. Each learnable edge function is a weighted combination of Lucas polynomials: $$\varphi_{q,p}(x_p)=\sum_{r=0}^{R}\gamma_{q,p,r}P_r(x_p)$$ with $P_0(x)=2$, $P_1(x)=x$, and the recurrence $P_r(x)=xP_{r-1}(x)+P_{r-2}(x)$. The recurrence makes evaluation linear in the polynomial degree, and the oscillatory shapes of the polynomials are invoked to capture high-frequency motion. Around this layer, a three-level DWT splits each joint trajectory into approximation and detail coefficients, two spatial projection matrices mix information across joints, and IDWT returns the processed coefficients to the time domain.

What would settle it

Run LuKAN and SiMLPe under the same protocol with ten random seeds and report per-horizon MPJPE means with standard deviations or confidence intervals. If the intervals at 720 ms and 1000 ms on Human3.6M overlap or favor SiMLPe, the claimed superiority over the strongest baseline is not supported; likewise, repeat Table 4's Lucas-versus-B-spline comparison and check whether its 2.8 mm gap at 320 ms persists across seeds.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a KAN layer parameterized by Lucas polynomials, inserted between two spatial projections and fed with three-level Daubechies-4 wavelet coefficients of joint trajectories, can forecast 3D poses at the level of the strongest current baselines. LuKAN reports the best average MPJPE among compared methods at every horizon on Human3.6M (109.3 mm vs 109.4 mm for SiMLPe at 1000 ms) and the best results on 3DPW at most horizons, with close-to-best values on AMASS-BMLrub. The ablations attribute consistent small gains to DWT over DCT encoding and larger gains to Lucas polynomials over B-spline, Chebyshev, Legendre, and Hermite bases. The authors conclude

Load-bearing premise

The load-bearing premise is that the MPJPE values quoted from different papers in Tables 1 and 2 are directly comparable and that the 0.1-0.5 mm differences between LuKAN and SiMLPe are real effects rather than run-to-run noise, since no variance, seeds, or significance tests are provided.

Editorial extensions

If this is right

  • If LuKAN's numbers hold, KAN-based motion predictors become a credible lightweight baseline: the model matches the strongest MLP competitor while using far fewer parameters and a fraction of the compute.
  • Encoding motion in the wavelet domain separates fast joint movements from slow postural drift, so the model can refine high-frequency detail without diffusing it away — the property the authors cite against MotionWavelet.
  • Lucas polynomials' linear recurrence keeps the learnable activations cheap to evaluate, which is what allows a 48-block temporal learner to stay small enough for an 8 GB GPU.
  • The practical gap over SiMLPe is small (0.1-0.2 mm at 720-1000 ms on Human3.6M), so the contribution is better read as competitive accuracy at lower cost than as a large accuracy jump.

Reading between the lines

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

  • Because no variance or significance tests are reported, this review's inference is that the head-to-head with SiMLPe should be read as statistical parity until multi-seed results appear; the architecture's value would then be efficiency, not accuracy.
  • The paper's own supplementary discussion (Section F) concedes that robustness on large-scale noisy or incomplete motion data and real-time suitability are untested; that concession narrows the efficiency claim to the evaluated benchmarks.
  • The Table 4 ablation suggests the polynomial basis, not the wavelet encoding, drives most of the gain: Lucas over B-spline is 2.8 mm at 320 ms, while DWT over DCT is 0.4 mm in the same setting.
  • The same recipe — DWT, polynomial KAN, IDWT — is a generic learned filter-bank forecaster; transferring it to vehicle trajectories or physiological signals would be a natural test the paper does not attempt.
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

6 major / 6 minor

Summary. The paper proposes LuKAN, a Kolmogorov-Arnold Network architecture for 3D human motion prediction. The model applies a discrete wavelet transform (DWT) to the input motion sequence, projects the coefficients into an embedding space, processes them with B blocks of a KAN layer whose activations are Lucas polynomials, and reconstructs the time-domain prediction via inverse DWT. Experiments on Human3.6M, AMASS-BMLrub, and 3DPW report MPJPE against several baselines, with ablations on temporal encoding, polynomial basis, and embedding dimension. The central claim is that LuKAN is competitive or superior to state-of-the-art methods while having significantly fewer parameters and lower computational cost. Code is provided.

Significance. If the performance and efficiency claims survive scrutiny, LuKAN is a plausible lightweight KAN-based baseline for 3D motion prediction. The architecture is simple, the ablation covers relevant design choices, and the public code is a practical asset. However, the empirical support is currently fragile: the reported accuracy advantages over the strongest baseline are mostly in the 0.1–0.5 mm range with no uncertainty quantification, and the efficiency claim is not measured. The paper is best viewed as a promising empirical study whose claims need to be tightened before publication.

major comments (6)
  1. [Section 4.2, Tables 1 and 2] The claim of 'consistently lower' MPJPE and 'superior' performance is not supported at the reported precision. On Human3.6M, LuKAN's advantage over SiMLPe is 0.1–0.4 mm (e.g., 89.9 vs 90.1 mm at 720 ms; 109.3 vs 109.4 mm at 1000 ms). On AMASS-BMLrub, LuKAN is worse than SiMLPe at every horizon from 320 ms to 1000 ms (e.g., 57.6 vs 57.3 at 720 ms; 66.4 vs 65.7 at 1000 ms). No variance, number of runs, or significance tests are reported, and the baselines are not rerun under identical conditions. These gaps are likely within run-to-run noise. Please report mean ± std over multiple seeds and either demonstrate statistical significance or soften the conclusion to 'competitive'.
  2. [Section 4.3, Table 3] The ablation text states that DWT 'demonstrates consistent improvements over DCT across all datasets,' but the same table shows DWT worse than DCT on Human3.6M at 160 ms (21.5 vs 21.4), 320 ms (46.2 vs 45.8), and 400 ms (57.2 vs 56.8). This internal inconsistency directly weakens the paper's motivation for DWT. The claim should be revised to acknowledge that DWT is favorably comparable or better on most settings but not uniformly consistent, or the experiments need to be rerun with error bars to establish consistency.
  3. [Section 4.4 and Conclusion] The computational-efficiency claim ('significantly fewer parameters and lower computational cost') is not substantiated by measurements. The complexity analysis is asymptotic only, and no parameter count, FLOPs, inference time, or throughput is reported for LuKAN or any baseline. The supplementary Figure 3 is a qualitative scatter plot without numerical parameter counts. Please provide actual measured parameter counts and runtime comparisons under a consistent protocol.
  4. [Section 3.3 and Section 4.4, Eqs. (5), (7)] The complexity formulas need dimension bookkeeping. Equation (7) operates on an (La + Ld) × D tensor, and the KAN layer is described as applied along the temporal dimension. However, Section 4.4 reports time complexity O(BDRL^2) and parameter complexity O(BRL^2). These formulas imply the KAN weights are shared across the D embedding channels, but sharing is not stated in Eq. (5). If the KAN weights are not shared, the parameter complexity is O(BD R L^2), not O(BR L^2). Please clarify the sharing pattern and recompute the complexity accordingly.
  5. [Supplementary Table 6 and Section 4.1] The embedding dimension D=200 is selected as the best performer on Human3.6M, but the paper does not state whether this is a validation or test split. If D is chosen from the same test table used to report final results, the reported MPJPE is optimistically biased. Similarly, B=48 is fixed without a sensitivity study. Please clarify the hyperparameter selection procedure or perform selection on a validation split.
  6. [Section 4.3, Table 4] The claim that Lucas polynomials yield 'significant improvements' over other polynomial bases is largely supported by sub-millimeter differences against the next-best basis, Hermite (e.g., 46.2 vs 46.3 mm at 320 ms; 57.2 vs 57.3 mm at 400 ms; 109.3 vs 110.1 mm at 1000 ms). With single-run MPJPE, these gaps are not distinguishable from noise. Please report variance or significance tests, and temper the language unless the advantage is statistically robust.
minor comments (6)
  1. [Supplementary Section F] The Limitations paragraph explicitly states that robustness on large-scale and noisy motion data requires further investigation. This should be reflected in the abstract/conclusion, which currently overstates generality.
  2. [Section 3.3] Typo: 'In it important to mention' should be 'It is important to mention'.
  3. [Section 4.2] Typo: 'test it on on the AMASS-BMLrub' has a duplicated 'on'.
  4. [Figure 3 (supplementary)] Axis labels contain 'uni00A0' unicode artifacts. Also, the plotted parameter counts for each method should be given numerically in the text or caption.
  5. [Full text] There are duplicated paragraphs and repeated Figure 2 instances in the main body; the final version should be cleaned.
  6. [References] Reference [26] has 'Maching Learning' instead of 'Machine Learning'; reference [9] has 'Fenga' rather than 'Feng'; the title in [25] should be in sentence case.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: LuKAN is an empirical architecture paper whose claims rest on test-set MPJPE comparisons, not on equations fitted to the target result or on load-bearing self-citation.

full rationale

The paper's chain is architectural construction plus empirical evaluation, not a derivation whose output equals an input by construction. The KAN layer is defined by a learnable combination of Lucas polynomials (Eq. 5) whose recurrence is cited to an external source [25]; the DWT/IDWT pair (Eqs. 2-3) is an invertible transform applied to input trajectories, and the spatial projections (Eqs. 4, 8) are learned linear maps trained with the loss in Eq. 10 on standard training splits. Predictions are then evaluated on held-out test data (Human3.6M S5, AMASS-BMLrub, 3DPW), so the reported MPJPE values are not fitted quantities renamed as predictions. Hyperparameter choices such as embedding dimension D=200 and number of blocks B=48 are selected via ablations on the same benchmarks; this is model selection, not definitional circularity, and the paper does not present those choices as independent predictions. No load-bearing self-citation appears: the cited Kolmogorov-Arnold theorem, KAN papers, and Lucas polynomial recurrence are all external prior work, and no uniqueness theorem from the authors is invoked to force the architecture. The strongest concern—that the advantage over SiMLPe is often 0.1-0.4 mm without variance or significance tests, and that Table 3's DWT-vs-DCT comparison is not consistently better on Human3.6M short horizons—is a statistical robustness and reporting concern, not a circularity concern. The supplementary limitation statement that robustness on large-scale, noisy, or incomplete data 'requires further investigation' also acknowledges an empirical gap rather than a circular step. Therefore no circular step can be quoted or reduced to the paper's own inputs, and the honest finding is no significant circularity.

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

Central claim rests on architecture hyperparameters and standard signal-processing assumptions, not on new physical entities. The most notable unstated choices are D=200 and B=48, plus the unreported polynomial degree R. No independent benchmarks are used outside the paper's own experiments.

free parameters (5)
  • embedding dimension D = 200
    Spatial projection width; selected by ablation on Human3.6M (Table 6) using test-set MPJPE.
  • Lucas polynomial degree R = not reported
    Defined in Eq. (5) and used in complexity O(BDRL^2), but its value is not stated anywhere.
  • number of temporal blocks B = 48
    Set in implementation details; no ablation in main text.
  • wavelet type and decomposition levels = Daubechies 4, 3 levels
    Chosen transform configuration; no sensitivity analysis.
  • pose and velocity loss balance = 1.0
    Eq. (10) sums pose and velocity L2 terms with equal weight; no tuning reported.
assumptions (4)
  • domain assumption Kolmogorov-Arnold representation theorem applies to the motion-prediction functions being approximated
    Invoked in Section 3 to motivate KANs; the theorem is about exact representation, not learnability with Lucas polynomials on this data.
  • domain assumption Lucas polynomials form a sufficiently expressive basis for the needed univariate functions
    Eq. (5) assumes each edge function is a finite weighted sum of Lucas polynomials; convergence for motion trajectories is not analyzed.
  • standard math DWT followed by IDWT reconstructs the trajectory after coefficient modification
    DWT/IDWT is invertible for unmodified coefficients with proper boundary handling, but the network modifies coefficients and boundary effects are not discussed (Section 3.1).
  • domain assumption Joint indices are consistently aligned across Human3.6M (22 joints), AMASS (23), and 3DPW (18) under the standard protocol
    The model fixes K=3J and evaluates across datasets with different J; alignment is delegated to protocol [22] (Section 4.1).

how reviews work

0 comments
Cite this review

Pith. "Pith review of LuKAN: A Kolmogorov-Arnold Network Framework for 3D Human Motion Prediction." pith.science (2026). https://pith.science/paper/UXYKMMPH

@misc{pith2026250804847,
  author       = {Pith},
  title        = {Pith review of: LuKAN: A Kolmogorov-Arnold Network Framework for 3D Human Motion Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UXYKMMPH}},
  note         = {Machine review of arXiv:2508.04847}
}
read the original abstract

The goal of 3D human motion prediction is to forecast future 3D poses of the human body based on historical motion data. Existing methods often face limitations in achieving a balance between prediction accuracy and computational efficiency. In this paper, we present LuKAN, an effective model based on Kolmogorov-Arnold Networks (KANs) with Lucas polynomial activations. Our model first applies the discrete wavelet transform to encode temporal information in the input motion sequence. Then, a spatial projection layer is used to capture inter-joint dependencies, ensuring structural consistency of the human body. At the core of LuKAN is the Temporal Dependency Learner, which employs a KAN layer parameterized by Lucas polynomials for efficient function approximation. These polynomials provide computational efficiency and an enhanced capability to handle oscillatory behaviors. Finally, the inverse discrete wavelet transform reconstructs motion sequences in the time domain, generating temporally coherent predictions. Extensive experiments on three benchmark datasets demonstrate the competitive performance of our model compared to strong baselines, as evidenced by both quantitative and qualitative evaluations. Moreover, its compact architecture coupled with the linear recurrence of Lucas polynomials, ensures computational efficiency.

Figures

Figures reproduced from arXiv: 2508.04847 by the authors.

Figure 1
Figure 1. Overview of Model Architecture. LuKAN processes input 3D motion data by applying DWT to encode temporal information. A spatial projection is applied both before and after the Temporal Dependency Learner block (repeated B times). Each block consists of a KAN layer, LayerNorm, and a residual skip connection. The inverse DWT (IDWT) reconstructs the motion in the time domain, outputting a sequence of predicted 3D poses.… view at source ↗
Figure 2
Figure 2. Visual comparison results of our model and the SiMLPe baseline on two ac￾tions: Directions (top) and Eating (bottom). Predicted poses from our model are depicted in red and blue, while those from SiMLPe are shown in yellow and green. Ground truth poses, represented by dashed lines, are overlaid with the predictions to highlight deviations [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Comparison of performance and model complexity. Our model is bench￾marked against state-of-the-art methods, including LTD [21], Hisrep [22], MSR-GCN [6], ST-DGCN [19], CIST-GCN [24], MotionMixer [3], and SiMLPe [11]. Performance is as￾sessed using the Mean Per Joint Position Error (MPJPE), where lower values indicate supe￾rior prediction performance. All evaluations are performed on the Human3.6M dataset. D Experime… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

33 extracted references · 1 linked inside Pith

  1. [1]

    A spatio-temporal transformer for 3D human motion prediction

    Emre Aksan, Manuel Kaufmann, Peng Cao, and Otmar Hilliges. A spatio-temporal transformer for 3D human motion prediction. In Proc. International Conference on 3D Vision, pages 565–574, 2021

    2021

  2. [2]

    Make- An-Animation: Large-scale text-conditional 3D human motion generation

    Samaneh Azadi, Akbar Shah, Thomas Hayes, Devi Parikh, and Sonal Gupta. Make- An-Animation: Large-scale text-conditional 3D human motion generation. In Proc. IEEE International Conference on Computer Vision, pages 15039–15048, 2023

    2023

  3. [3]

    MotionMixer: MLP-based 3D human body pose forecasting

    Arij Bouazizi, Adrian Holzbock, Ulrich Kressel, Klaus Dietmayer, and Vasileios Bela- giannis. MotionMixer: MLP-based 3D human body pose forecasting. In Proc. Inter- national Joint Conference on Artificial Intelligence, pages 791–798, 2022

    2022

  4. [4]

    On a constructive proof of Kolmogorov’s superpo- sition theorem

    Jürgen Braun and Michael Griebel. On a constructive proof of Kolmogorov’s superpo- sition theorem. Constructive Approximation, 30:653–675, 2009

    2009

  5. [5]

    Learning progressive joint propagation for human motion prediction

    Yujun Cai, Lin Huang, Yiwei Wang, Tat-Jen Cham, Jianfei Cai, Junsong Yuan, Jun Liu, Xu Yang, Yiheng Zhu, Xiaohui Shen, Ding Liu, Jing Liu, and Nadia Magnenat Thalmann. Learning progressive joint propagation for human motion prediction. In Proc. European Conference on Computer Vision, 2020

    2020

  6. [6]

    MSR- GCN: Multi-scale residual graph convolution networks for human motion prediction

    Lingwei Dang, Yongwei Nie, Chengjiang Long, Qing Zhang, and Guiqing Li. MSR- GCN: Multi-scale residual graph convolution networks for human motion prediction. In Proc. IEEE International Conference on Computer Vision, pages 11447–11456, 2021

    2021

  7. [7]

    CG-HOI: Contact-guided 3D human-object interac- tion generation

    Christian Diller and Angela Dai. CG-HOI: Contact-guided 3D human-object interac- tion generation. In Proc. IEEE Conference on Computer Vision and Pattern Recogni- tion, pages 19888–19901, 2024

    2024

  8. [8]

    Uncertainty-aware short- term motion prediction of traffic actors for autonomous driving

    Nemanja Djuric, Vladan Radosavljevic, Henggang Cui, Thi Nguyen, Fang-Chieh Chou, Tsung-Han Lin, Nitin Singh, and Jeff Schneiders. Uncertainty-aware short- term motion prediction of traffic actors for autonomous driving. In Proc. IEEE Winter Conference on Applications of Computer Vision, 2020

    2020

Show all 33 references
  1. [9]

    MotionWavelet: Human motion prediction via wavelet manifold learning

    Yuming Fenga, Zhiyang Dou, Ling-Hao Chen, Yuan Liu, Tianyu Li, Jingbo Wang, Zeyu Cao, Wenping Wang, Taku Komura, and Lingjie Liu. MotionWavelet: Human motion prediction via wavelet manifold learning. arXiv:2411.16964, 2024

    2024 arXiv

  2. [10]

    Recurrent net- work models for human dynamics

    Katerina Fragkiadaki, Sergey Levine, Panna Felsen, and Jitendra Malik. Recurrent net- work models for human dynamics. Proc. IEEE International Conference on Computer Vision, pages 4346–4354, 2015

    2015

  3. [11]

    Back to MLP: A simple baseline for human motion pre- diction

    Wen Guo, Yuming Du, Xi Shen, Vincent Lepetit, Xavier Alameda-Pineda, and Francesc Moreno-Noguer. Back to MLP: A simple baseline for human motion pre- diction. In Proc. IEEE Winter Conference on Applications of Computer Vision, pages 4809–4819, 2023

    2023

  4. [12]

    Human3.6M: Large scale datasets and predictive methods for 3D human sensing in natural envi- ronments

    Catalin Ionescu, Dragos Papava, Vlad Olaru, and Cristian Sminchisescu. Human3.6M: Large scale datasets and predictive methods for 3D human sensing in natural envi- ronments. IEEE Transactions on Pattern Analysis and Machine Intelligence , 36(7): 1325–1339, 2014. 12 HASAN ET AL...

    2014

  5. [13]

    Zamir, Silvio Savarese, and Ashutosh Saxena

    Ashesh Jain, Amir R. Zamir, Silvio Savarese, and Ashutosh Saxena. Structural-RNN: Deep learning on spatio-temporal graphs. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, pages 5308–5317, 2016

    2016

  6. [14]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In International Conference on Learning Representations, 2015

    2015

  7. [15]

    Convolutional sequence to sequence model for human dynamics

    Chen Li, Zhen Zhang, Wee Sun Lee, and Gim Hee Lee. Convolutional sequence to sequence model for human dynamics. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, pages 5226–5234, 2018

    2018

  8. [16]

    Dynamic multiscale graph neural networks for 3D skeleton-based human motion pre- diction

    Maosen Li, Siheng Chen, Yangheng Zhao, Ya Zhang, Yanfeng Wang, and Qi Tian. Dynamic multiscale graph neural networks for 3D skeleton-based human motion pre- diction. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, pages 215–223, 2020

    2020

  9. [17]

    Skeleton- parted graph scattering networks for 3D human motion predictiont

    Maosen Li, Siheng Chen, Zijing Zhang, Lingxi Xie, Qi Tian, and Ya Zhang. Skeleton- parted graph scattering networks for 3D human motion predictiont. In Proc. European Conference on Computer Vision, pages 18–36, 2022

    2022

  10. [18]

    Hou, and Max Tegmark

    Ziming Liu, Yixuan Wang, Sachin Vaidya, Fabian Ruehle, James Halverson, Marin Soljacic, Thomas Y . Hou, and Max Tegmark. KAN: Kolmogorov-arnold networks. In International Conference on Learning Representations, 2025

    2025

  11. [19]

    Pro- gressively generating better initial guesses towards next stages for high-quality human motion prediction

    Tiezheng Ma, Yongwei Nie, Chengjiang Long, Qing Zhang, and Guiqing Li. Pro- gressively generating better initial guesses towards next stages for high-quality human motion prediction. In Proc. IEEE Conference on Computer Vision and Pattern Recog- nition, pages 6437–6446, 2022

    2022

  12. [20]

    Troje, Gerard Pons-Moll, and Michael J

    Naureen Mahmood, Nima Ghorbani, Nikolaus F. Troje, Gerard Pons-Moll, and Michael J. Black. AMASS: Archive of motion capture as surface shapes. In Proc. IEEE International Conference on Computer Vision, 2019

    2019

  13. [21]

    Learning trajectory dependencies for human motion prediction

    Wei Mao, Miaomiao Liu, Mathieu Salzmann, and Hongdong Li. Learning trajectory dependencies for human motion prediction. In Proc. IEEE International Conference on Computer Vision, pages 9489–9497, 2019

    2019

  14. [22]

    History repeats itself: Human motion prediction via motion attention

    Wei Mao, Miaomiao Liu, and Mathieu Salzmann. History repeats itself: Human motion prediction via motion attention. In Proc. European Conference on Computer Vision , pages 474–489, 2020

    2020

  15. [23]

    Black, and Javier Romero

    Julieta Martinez, Michael J. Black, and Javier Romero. On human motion prediction using recurrent neural networks. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, pages 4674–4683, 2017

    2017

  16. [24]

    Context-based interpretable spatio-temporal graph convolutional network for human motion forecasting

    Edgar Medina, Leyong Loh, Namrata Gurung, Kyung Hun Oh, and Niels Heller. Context-based interpretable spatio-temporal graph convolutional network for human motion forecasting. In Proc. IEEE Winter Conference on Applications of Computer Vision, 2024

    2024

  17. [25]

    Ömer Oruç. A new algorithm based on lucas polynomials for approximate solution of 1D and 2D nonlinear generalized Benjamin-Bona-Mahony-Burgers equation.Comput- ers and Mathematics with Applications, 74:3042–3057, 2017. HASAN ET AL.: LUKAN FOR 3D HUMAN MOTION PREDICTION 13

    2017

  18. [26]

    Hamprecht, Yoshua Bengio, and Aaron Courville

    Nasim Rahaman, Aristide Baratin, Devansh Arpit, Felix Draxler, Min Lin, Fred A. Hamprecht, Yoshua Bengio, and Aaron Courville. On the spectral bias of neural net- works. In Proc. International Conference on Maching Learning, 2019

    2019

  19. [27]

    The Kolmogorov-Arnold representation theorem revisited

    Johannes Schmidt-Hieber. The Kolmogorov-Arnold representation theorem revisited. Neural Networks, 137:119–126, 2021

    2021

  20. [28]

    De- FeeNet: Consecutive 3D human motion prediction with deviation feedback

    Xiaoning Sun, Huaijiang Sun, Bin Li, Dong Wei, Weiqing Li, and Jianfeng Lu. De- FeeNet: Consecutive 3D human motion prediction with deviation feedback. In Proc. IEEE Conference on Computer Vision and Pattern Recognition , pages 5527–5536, 2023

    2023

  21. [29]

    Recovering accurate 3D human pose in the wild using IMUs and a moving camera

    Timo von Marcard, Roberto Henschel, Michael Black Bodo J., Rosenhahn, and Gerard Pons-Moll. Recovering accurate 3D human pose in the wild using IMUs and a moving camera. In Proc. European Conference on Computer Vision, 2018

    2018

  22. [30]

    Siegel, Ziming Liu, and Thomas Y

    Yixuan Wang, Jonathan W. Siegel, Ziming Liu, and Thomas Y . Hou. On the expres- siveness and spectral bias of KANs. In International Conference on Learning Repre- sentations, 2025

    2025

  23. [31]

    NeRMo: Learning implicit neural representations for 3D human motion prediction

    Dong Wei, Huaijiang Sun, Xiaoning Sun, and Shengxiang Hug. NeRMo: Learning implicit neural representations for 3D human motion prediction. In Proc. European Conference on Computer Vision, pages 409–427, 2024

    2024

  24. [32]

    MotionNet: Joint perception and motion prediction for autonomous driving based on bird’s eye view maps

    Pengxiang Wu, Siheng Chen, and Dimitris Metaxas. MotionNet: Joint perception and motion prediction for autonomous driving based on bird’s eye view maps. In Proc. IEEE Conference on Computer Vision and Pattern Recognition , pages 11385–11395, 2020

    2020

  25. [33]

    T2M-GPT: Generating human motion from tex- tual descriptions with discrete representations

    Jianrong Zhang, Yangsong Zhang, Xiaodong Cun, Shaoli Huang, Yong Zhang, Hong- wei Zhao, Hongtao Lu, and Xi Shen. T2M-GPT: Generating human motion from tex- tual descriptions with discrete representations. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, 202...

    2023

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.