Pith. sign in

REVIEW 3 major objections 4 minor 52 references

Dyadic Mamba: Long-term Dyadic Human Motion Synthesis

T0 review · 3 major / 4 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 This paper claims that a Mamba-based diffusion model can synthesize text-described two-person motion with stable per-person quality far beyond the 10-second training horizon, where transformer-based generators degrade.

desk verdict Useful Mamba-based recipe for dyadic motion with a clever concatenation design, but the long-term superiority claim rests on a metric that doesn't measure interaction quality and a table typo. read the letter →

arxiv 2505.09827 v1 pith:P3AUCSP5 submitted 2025-05-14 cs.CV

Julian Tanke , Takashi Shibuya , Kengo Uchida , Koichi Saito , Yuki Mitsufuji This is my paper · ORCID
classification cs.CV
keywords humanmotionsynthesisdyadicinteractionstate-spacemodelsMambatext-to-motionlong-termgenerationdiffusionqualitybenchmark
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

This paper tries to establish that state-space models, not transformers, are the right backbone for generating text-described two-person (dyadic) motion, especially for long interactions. It introduces Dyadic Mamba, a diffusion model that runs each person's motion through its own Mamba layers and mixes the two streams by simple concatenation, and argues this reaches transformer-level quality on standard short-term benchmarks while keeping motion quality stable at 14 and 28 seconds—beyond the 10-second training length. If true, long social interactions could be synthesized in one pass without windowing tricks or positional-encoding fixes, and with fewer parameters. The paper also contributes a long-term evaluation benchmark based on per-person motion similarity measured over one-third-second windows.

What carries the argument

The load-bearing object is the Mamba layer, a selective state-space model that processes a sequence in linear time by updating a recurrent hidden state, so it has no positional encoding and no quadratic attention cost; in this paper it runs as two paired modules per block—a self-Mamba per person and a cross-Mamba that receives the concatenation of both persons' intermediate features. The same weights process both persons, and Adaptive LayerNorm injects the text embedding and diffusion step. This design is trained as a data-space denoising diffusion model, and the long-term claim is measured by per-person Normalized Directional Motion Similarity over one-third-second windows at 7, 14, and 28 seconds.

What would settle it

Generate 28-second motions with Dyadic Mamba on InterHuman test prompts and measure per-frame contact distance between the two bodies, the angle between their facing directions, and the consistency of interaction roles; compare these against ground-truth motions. If per-person NDMS stays around 0.37 while these interaction measures deviate from the real data just as much as InterGen's do, the claim that Dyadic Mamba outperforms transformers on long-term dyadic synthesis is falsified.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that a state-space backbone removes the length ceiling for dyadic motion generation. Dyadic Mamba is a denoising diffusion model whose denoiser is built from Mamba layers arranged in cooperative blocks: each person's motion passes first through a shared self-Mamba, the two streams are concatenated along the feature dimension and down-projected, and a second shared Mamba mixes them; conditioning on text and diffusion time step uses Adaptive LayerNorm. Trained on at most 10-second sequences, the model matches transformer-based short-term results and maintains an NDMS score around 0.37 at 7s, 14s, and 28s, whereas InterGen drops from 0.34 to 0.27 and InterGen with rotary positional embeddings drops from 0.32 to 0.21. The paper takes this as evidence that SSM-based architectures extrapolate to arbitrary length, eliminating the need for windowing or positional-encoding workarounds.

Load-bearing premise

The load-bearing premise is that per-person motion quality, averaged over one-third-second windows, is a sufficient indicator of long-term dyadic quality; if the interaction between the two people can degrade while each individual's short windows look fine, the paper's long-term advantage over transformers is not proven.

Editorial extensions

If this is right

  • Long social interactions can be generated in a single diffusion pass rather than stitched windows, because the state-space layer carries context without positional encodings.
  • Dyadic coordination does not require cross-attention: a shared Mamba over concatenated person streams is enough, which simplifies the architecture and roughly halves the parameter count relative to InterGen.
  • The known failure of rotary embeddings beyond roughly twice the training length transfers from language modeling to motion, explaining why the RoPE-enhanced transformer still collapses at 28 seconds.
  • The new per-person NDMS benchmark gives later methods a cheap, standardized way to report long-term stability on InterHuman.
  • Because Mamba's recurrence is linear in sequence length, generating minute-long interactions avoids the memory growth that attention would incur.

Reading between the lines

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

  • Editorial inference: the same concatenation-based interaction module may extend to three or more persons; the paper notes the addition variant is order-invariant, and an asymmetric concatenation could be replaced by a permutation-equivariant mixing operation.
  • Editorial inference: the benchmark's per-person focus leaves room for an interaction-level long-term metric; a fair next test would track contact distance, facing angle, and role consistency over 28-second generations.
  • Editorial inference: 'arbitrary length' is demonstrated only up to 28 seconds; genuinely arbitrary length would require checking whether the recurrent state saturates or drifts over minute-scale generations, which the paper does not report.
  • Editorial inference: the method's success suggests state-space models might also stabilize long single-person motion generation and motion in-betweening, since the length bottleneck is in the backbone, not the dyadic task.
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

3 major / 4 minor

Summary. The paper introduces Dyadic Mamba, a diffusion-based state-space model for text-to-dyadic human motion synthesis. The architecture processes the two persons' motion sequences with shared per-person Mamba blocks and exchanges information through concatenation with a learnable down-projection, avoiding cross-attention. The paper reports competitive short-term results on InterHuman and Inter-X, and proposes a long-term benchmark based on per-person per-frame Normalized Directional Motion Similarity (NDMS). The central claim is that Dyadic Mamba extrapolates beyond the 10-second training horizon and 'significantly outperforms transformer-based approaches' on longer dyadic sequences, while remaining competitive on short-term benchmarks. The paper also includes ablations of model size, conditioning, and cross-person information flow, and discusses failure cases.

Significance. The architectural direction is timely: demonstrating that a linear-time SSM backbone can handle dyadic motion synthesis and extrapolate temporally is a useful contribution, especially if it holds with parameter efficiency. The authors are also transparent in reporting a long-term evaluation protocol, which is valuable for future work. However, the strongest advertised claim — significant superiority over transformers for long-term dyadic synthesis — is not quantitatively established: the long-term metric measures only individual per-person motion quality, and the paper explicitly concedes that interaction quality is not evaluated. The short-term Inter-X results also contain an impossible reported number. If the metric issues and reporting errors are corrected and the claims are re-scoped to local single-person motion quality over long horizons, the contribution is solid but less headline-worthy. The use of NDMS from the first author's prior work is not circular, since it is an external quality metric rather than a learned evaluator fitted to the model, but it is not sufficient for the dyadic claim.

major comments (3)
  1. [Table 1, Inter-X block] The reported Inter-X Top-1 R-Precision for the proposed method is 3.658±0.007, which is larger than 1 and therefore cannot be an accuracy. This is an impossible value for a retrieval-based precision metric and makes the Inter-X short-term comparison uninterpretable as printed. The authors must correct the value (likely a decimal error, e.g., 0.3658) and re-evaluate the associated conclusions about competitive performance on Inter-X.
  2. [Section 4.4, Table 2, Figure 4] The central long-term claim is not supported by the reported metric. NDMS is computed per person over 1/3-second windows and measures local single-person motion similarity, not dyadic interaction quality. The paper itself states in Section 4.4: 'our approach only evaluates individual motion quality, and per-frame human interaction evaluation remains an open problem for future work.' The failure cases in Figure 7 (clipping, person-order changes, one person walking while the other is instructed to interact) are exactly the kinds of dyadic coherence failures that NDMS cannot detect. Consequently, Table 2 establishes only that per-person local motion quality degrades less for Dyadic Mamba than for InterGen beyond 10s; it does not establish the abstract's claim of 'significantly outperforming transformer-based approaches on longer sequences' for dyadic motion synthesis.
  3. [Table 2, long-term benchmark] The long-term comparison lacks both a real-data reference at 14s and 28s and any significance testing. The 'Real' row is given only at 7s (0.451±0.152), and at 14s and 28s the models are compared only against each other. Given the reported standard deviations — for example Ours 0.379±0.157 versus InterGen 0.290±0.118 at 14s, and Ours 0.376±0.155 versus InterGen (RoPE) 0.212±0.009 at 28s — the differences are not obviously significant, and the phrase 'significantly outperforms' is unsupported. The authors should report significance tests or confidence intervals, and ideally provide real-data NDMS values at the longer horizons to calibrate what good long-term performance means.
minor comments (4)
  1. [References] Reference [13] contains a typo: 'Internantional Conference' should be 'International Conference on Learning Representations.'
  2. [Table 1] Several citation labels in Table 1 appear inconsistent with the reference list: 'ComMDM [32]' and 'RIG [32]' both point to reference [32], but the reference list assigns [32] to the PriorMDM paper, while RIG appears to be reference [36]. Please verify all citation keys in the table.
  3. [Figure 2] The caption says 'single-step denoising' while the model is trained with T=1000 diffusion steps and sampled with 50 DDIM steps; consider rewording to 'one denoising network evaluation' to avoid confusion.
  4. [Table 3] The column header row for the ablation table is difficult to parse, especially the '#Param S M L' and 'Prepending+ ⊕' entries; a clearer layout with explicit sub-headers would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the long-term evaluation uses an external, published NDMS metric; the admitted lack of interaction-level evaluation is a validity limitation, not a derivation loop.

full rationale

The paper's central claims—competitive short-term results and better long-term motion quality than transformer baselines—are supported by standard external benchmarks (InterHuman and Inter-X protocols, Table 1) and by a published similarity metric, NDMS [37], applied in Section 4.4. The only self-citation is to NDMS, a metric on which the first author is a co-author. This is not load-bearing circularity: NDMS is a fixed, externally defined metric with a specified 1/3-second window, it is not fitted to or derived from Dyadic Mamba, and it is applied symmetrically to all compared methods. The paper does not redefine NDMS, fit its parameters, or use it inside the model's training objective. The 'new benchmark' is a protocol that reuses this independent metric over longer horizons, not a renamed version of the model's own output. The paper itself states in Section 4.4: 'our approach only evaluates individual motion quality, and per-frame human interaction evaluation remains an open problem for future work.' That passage is a substantive validity caveat—the long-term dyadic claim is not fully established because interaction-level coherence is unmeasured—but a measurement gap is not a circular derivation. No equation in the paper reduces a prediction to an input, no fitted parameter is renamed as a prediction, and the cited long-horizon limitations of RoPE/positional encodings are from external prior work and are not used to define the model's outputs. Therefore, under the hard rules requiring a specific reduction, no circular step is present.

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

No derivation is offered; all design choices are empirical. The central long-term claim depends on the chosen architecture size and evaluation protocol, neither of which is justified theoretically. No new physical or parametric entities are introduced; the proposed long-term benchmark is an evaluation protocol, not an invented entity.

free parameters (5)
  • Architecture hyperparameters = N=8 blocks, latent dim h=512, expansion factor 2, conv kernel size 4
    Chosen by hand; the M variant is selected for a balance of R-Precision and FID, with no theoretical derivation.
  • Model size variant = M variant, 79M parameters
    Selected over S and L variants by prioritizing text-to-motion fidelity (R-Precision), as stated in Section 4.5.
  • Diffusion steps and schedule = T=1000 training, 50 DDIM steps inference, cosine schedule
    Standard choices from prior diffusion work; they affect generation quality but are not fitted to the long-term claim.
  • Text masking probability = 10%
    Follows InterGen's conditional training recipe; no ablation is provided for this value.
  • NDMS evaluation window = 1/3 second
    Taken from the NDMS paper [37] as the recommended window; it is an evaluation choice, not a model parameter.
assumptions (5)
  • domain assumption Mamba selective state-space layers process sequences of arbitrary length without learned positional encodings.
    The long-horizon extrapolation claim relies on this property of Mamba, assumed from prior work [8]; the paper does not derive or formally verify it.
  • domain assumption Frozen CLIP-ViT-L/14 text embeddings are a sufficient conditioning signal for dyadic motion.
    Adopted from InterGen in Section 3.2; no analysis of text-embedding failure modes is provided.
  • ad hoc to paper The two motion streams are temporally aligned frame by frame, so per-frame concatenation is a sufficient channel for cross-person information.
    This is the core design assumption in Equations 5-6 and Section 4.5; no proof is given that concatenation preserves interaction coherence over long horizons.
  • domain assumption Per-person NDMS over 1/3-second windows is a meaningful measure of long-term motion synthesis quality.
    Section 4.4 uses NDMS as the long-term benchmark while admitting it does not evaluate interaction quality.
  • standard math Standard DDPM/DDIM diffusion objectives hold for motion data.
    Background from [12, 34]; not load-bearing beyond standard practice in the field.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dyadic Mamba: Long-term Dyadic Human Motion Synthesis." pith.science (2026). https://pith.science/paper/P3AUCSP5

@misc{pith2026250509827,
  author       = {Pith},
  title        = {Pith review of: Dyadic Mamba: Long-term Dyadic Human Motion Synthesis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P3AUCSP5}},
  note         = {Machine review of arXiv:2505.09827}
}
read the original abstract

Generating realistic dyadic human motion from text descriptions presents significant challenges, particularly for extended interactions that exceed typical training sequence lengths. While recent transformer-based approaches have shown promising results for short-term dyadic motion synthesis, they struggle with longer sequences due to inherent limitations in positional encoding schemes. In this paper, we introduce Dyadic Mamba, a novel approach that leverages State-Space Models (SSMs) to generate high-quality dyadic human motion of arbitrary length. Our method employs a simple yet effective architecture that facilitates information flow between individual motion sequences through concatenation, eliminating the need for complex cross-attention mechanisms. We demonstrate that Dyadic Mamba achieves competitive performance on standard short-term benchmarks while significantly outperforming transformer-based approaches on longer sequences. Additionally, we propose a new benchmark for evaluating long-term motion synthesis quality, providing a standardized framework for future research. Our results demonstrate that SSM-based architectures offer a promising direction for addressing the challenging task of long-term dyadic human motion synthesis from text descriptions.

Figures

Figures reproduced from arXiv: 2505.09827 by the authors.

Figure 1
Figure 1. Given a textual description of a dyadic interaction, our model produces the dyadic human motion for the two persons. Our [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Model Overview: The Dyadic Mamba performs single-step denoising of a t-noised dyadic signal {x (t) a , x (t) b } to produce the denoised dyadic signal {ˆx (t−1) a , ˆx (t−1) b }. Our architecture comprises N cooperative blocks that process signals iteratively, culminating in the final motion prediction. Each block integrates text embedding c and diffusion step embedding t through projection and summation, yielding a… view at source ↗
Figure 3
Figure 3. Qualitative Results: Dyadic motion generation results for various text descriptions. Dataset Method R Precision↑ FID ↓ MM Dist.↓ Diversity→ MModality ↑ Arbitrary Top 1 Top 2 Top 3 length InterHuman [18] Real 0.452±.008 0.610±.009 0.701±.008 0.273±.007 3.755±.008 7.948±.064 - - TEMOS [26] 0.224±.010 0.316±.013 0.450±.018 17.375±.043 6.342±.015 6.939±.071 0.535±.014 ✗ T2M [9] 0.238±.012 0.325±.010 0.464±.014 13.769±.0… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Per-frame long-term per-person motion quality [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Conditioning: We experiment with two variants for conditioning the Dyadic Mamba on diffusion step t and text em￾bedding c: (a) Prepending approach: The conditioning embed￾dings are simply prepended to the motion sequence before passing it to the Mamba module. (b) Adapt…
Figure 7
Figure 7. Figure 7: Failure Cases: First row: if one person is static (i.e. sitting) and the other person is instructed to interact with them, they might clip through each other and/or change order. Second + third row: out-of-distribution text: the persons are not back-to￾back but face th…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 43 canonical work pages

  1. [1]

    Seamless human motion composition with blended posi- tional encodings

    German Barquero, Sergio Escalera, and Cristina Palmero. Seamless human motion composition with blended posi- tional encodings. In Conference on Computer Vision and Pattern Recognition, 2024. 2

    work page 2024

  2. [2]

    Executing your commands via motion diffusion in latent space

    Xin Chen, Biao Jiang, Wen Liu, Zilong Huang, Bin Fu, Tao Chen, and Gang Yu. Executing your commands via motion diffusion in latent space. In Conference on Computer Vision and Pattern Recognition, 2023. 2

    work page 2023

  3. [3]

    Flexible motion in-betweening with diffusion models

    Setareh Cohan, Guy Tevet, Daniele Reda, Xue Bin Peng, and Michiel van de Panne. Flexible motion in-betweening with diffusion models. In ACM SIGGRAPH, 2024. 2

    work page 2024

  4. [4]

    Bert: Pre-training of deep bidirectional trans- formers for language understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional trans- formers for language understanding. In Conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers), 2019. 2, 3

    work page 2019

  5. [5]

    Lon- grope: Extending llm context window beyond 2 million to- kens, 2024

    Yiran Ding, Li Lyna Zhang, Chengruidong Zhang, Yuanyuan Xu, Ning Shang, Jiahang Xu, Fan Yang, and Mao Yang. Lon- grope: Extending llm context window beyond 2 million to- kens, 2024. 2

    work page 2024

  6. [6]

    Freemotion: A unified framework for number- free text-to-motion synthesis

    Ke Fan, Junshu Tang, Weijian Cao, Ran Yi, Moran Li, Jingyu Gong, Jiangning Zhang, Yabiao Wang, Chengjie Wang, and Lizhuang Ma. Freemotion: A unified framework for number- free text-to-motion synthesis. In European Conference on Computer Vision, 2025. 3, 5

    work page 2025

  7. [7]

    Tm2d: Bimodality driven 3d dance generation via music- text integration

    Kehong Gong, Dongze Lian, Heng Chang, Chuan Guo, Zi- hang Jiang, Xinxin Zuo, Michael Bi Mi, and Xinchao Wang. Tm2d: Bimodality driven 3d dance generation via music- text integration. In International Conference on Computer Vision, 2023. 2

    work page 2023

  8. [8]

    Mamba: Linear-time sequence modeling with selective state spaces

    Albert Gu and Tri Dao. Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752, 2023. 2, 3

    arXiv 2023

Show all 52 references
  1. [9]

    Generating diverse and natural 3d human motions from text

    Chuan Guo, Shihao Zou, Xinxin Zuo, Sen Wang, Wei Ji, Xingyu Li, and Li Cheng. Generating diverse and natural 3d human motions from text. In Conference on Computer Vision and Pattern Recognition, 2022. 2, 5, 6

    2022

  2. [10]

    Momask: Generative masked model- ing of 3d human motions

    Chuan Guo, Yuxuan Mu, Muhammad Gohar Javed, Sen Wang, and Li Cheng. Momask: Generative masked model- ing of 3d human motions. InConference on Computer Vision and Pattern Recognition, 2024. 2

    2024

  3. [11]

    Gans trained by a two time-scale update rule converge to a local nash equilib- rium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilib- rium. Advances in neural information processing systems ,

  4. [12]

    Denoising diffu- sion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffu- sion probabilistic models. Advances in Neural Information Processing Systems, 2020. 2, 4

    2020

  5. [13]

    Intermask: 3d human interaction generation via collaborative masked modelling

    Muhammad Gohar Javed, Chuan Guo, Li Cheng, and Xingyu Li. Intermask: 3d human interaction generation via collaborative masked modelling. Internantional Conference on Learning Representation, 2025. 2, 3, 5

    2025

  6. [14]

    Motiongpt: Human motion as a foreign language

    Biao Jiang, Xin Chen, Wen Liu, Jingyi Yu, Gang Yu, and Tao Chen. Motiongpt: Human motion as a foreign language. Advances in Neural Information Processing Systems , 2024. 2

    2024

  7. [15]

    The impact of positional encoding on length generalization in transformers

    Amirhossein Kazemnejad, Inkit Padhi, Karthikeyan Nate- san Ramamurthy, Payel Das, and Siva Reddy. The impact of positional encoding on length generalization in transformers. Advances in Neural Information Processing Systems , 2023. 2, 7

    2023

  8. [16]

    The use of intonation for turn anticipation in observed conversations without visual signals as source of information

    Anne Keitel, Moritz M Daum, et al. The use of intonation for turn anticipation in observed conversations without visual signals as source of information. Frontiers in psychology,

  9. [17]

    Dancing to music

    Hsin-Ying Lee, Xiaodong Yang, Ming-Yu Liu, Ting-Chun Wang, Yu-Ding Lu, Ming-Hsuan Yang, and Jan Kautz. Dancing to music. Advances in Neural Information Process- ing Systems, 2019. 2

    2019

  10. [18]

    Intergen: Diffusion-based multi-human motion genera- tion under complex interactions

    Han Liang, Wenqian Zhang, Wenxuan Li, Jingyi Yu, and Lan Xu. Intergen: Diffusion-based multi-human motion genera- tion under complex interactions. International Journal of Computer Vision, 2024. 1, 2, 3, 4, 5, 6, 7, 8

    2024

  11. [19]

    Smpl: A skinned multi- person linear model

    Matthew Loper, Naureen Mahmood, Javier Romero, Gerard Pons-Moll, and Michael J Black. Smpl: A skinned multi- person linear model. Transactions on Graphics, 2015. 3, 4, 5

    2015

  12. [20]

    Improved denoising diffusion probabilistic models

    Alexander Quinn Nichol and Prafulla Dhariwal. Improved denoising diffusion probabilistic models. In International conference on machine learning, 2021. 4

    2021

  13. [21]

    Georgios Pavlakos, Vasileios Choutas, Nima Ghorbani, Timo Bolkart, Ahmed A. A. Osman, Dimitrios Tzionas, and Michael J. Black. Expressive body capture: 3D hands, face, and body from a single image. In Conference on Computer Vision and Pattern Recognition, 2019. 3, 4, 5

    2019

  14. [22]

    Scalable diffusion models with transformers

    William Peebles and Saining Xie. Scalable diffusion models with transformers. In International Conference on Computer Vision, 2023. 3, 4, 8

    2023

  15. [23]

    Rwkv: Reinventing rnns for the transformer era

    Bo Peng, Eric Alcaide, Quentin Anthony, Alon Albalak, Samuel Arcadinho, Stella Biderman, Huanqi Cao, Xin Cheng, Michael Chung, Matteo Grella, et al. Rwkv: Reinventing rnns for the transformer era. arXiv preprint arXiv:2305.13048, 2023. 3

    2023 arXiv

  16. [24]

    YaRN: Efficient context window extension of large language models

    Bowen Peng, Jeffrey Quesnelle, Honglu Fan, and Enrico Shippole. YaRN: Efficient context window extension of large language models. In The Twelfth International Conference on Learning Representations, 2024. 2, 7

    2024

  17. [25]

    Film: Visual reasoning with a general conditioning layer

    Ethan Perez, Florian Strub, Harm De Vries, Vincent Du- moulin, and Aaron Courville. Film: Visual reasoning with a general conditioning layer. In Conference on Artificial Intel- ligence, 2018. 3, 4, 8

    2018

  18. [26]

    Temos: Generating diverse human motions from textual descriptions

    Mathis Petrovich, Michael J Black, and G ¨ul Varol. Temos: Generating diverse human motions from textual descriptions. European Conference on Computer Vision, 2022. 2, 5

    2022

  19. [27]

    Bamm: bidirectional autoregressive motion model

    Ekkasit Pinyoanuntapong, Muhammad Usama Saleem, Pu Wang, Minwoo Lee, Srijan Das, and Chen Chen. Bamm: bidirectional autoregressive motion model. In European Conference on Computer Vision, 2024. 2

    2024

  20. [28]

    Mmm: Generative masked motion model

    Ekkasit Pinyoanuntapong, Pu Wang, Minwoo Lee, and Chen Chen. Mmm: Generative masked motion model. In Con- ference on Computer Vision and Pattern Recognition, 2024. 2 9

    2024

  21. [29]

    Train short, test long: Attention with linear biases enables input length extrapolation

    Ofir Press, Noah A Smith, and Mike Lewis. Train short, test long: Attention with linear biases enables input length extrapolation. International Conference on Learning Repre- sentations, 2022. 2, 7

    2022

  22. [30]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  23. [31]

    in2in: Leverag- ing individual information to generate human interactions

    Pablo Ruiz-Ponce, German Barquero, Cristina Palmero, Ser- gio Escalera, and Jos ´e Garc´ıa-Rodr´ıguez. in2in: Leverag- ing individual information to generate human interactions. In Conference on Computer Vision and Pattern Recognition Workshops, 2024. 3

    2024

  24. [32]

    Human motion diffusion as a generative prior

    Yonatan Shafir, Guy Tevet, Roy Kapon, and Amit H Bermano. Human motion diffusion as a generative prior. arXiv preprint arXiv:2303.01418, 2023. 5

    2023 arXiv

  25. [33]

    Human motion diffusion as a generative prior

    Yoni Shafir, Guy Tevet, Roy Kapon, and Amit Haim Bermano. Human motion diffusion as a generative prior. In International Conference on Learning Representations ,

  26. [34]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502, 2020. 4

    2010 arXiv

  27. [35]

    Roformer: Enhanced transformer with rotary position embedding

    Jianlin Su, Murtadha Ahmed, Yu Lu, Shengfeng Pan, Wen Bo, and Yunfeng Liu. Roformer: Enhanced transformer with rotary position embedding. Neurocomputing, 2024. 2, 6, 7

    2024

  28. [36]

    Role-aware interaction generation from textual description

    Mikihiro Tanaka and Kent Fujiwara. Role-aware interaction generation from textual description. In International Confer- ence on Computer Vision, 2023. 3

    2023

  29. [37]

    Intention- based long-term human motion anticipation

    Julian Tanke, Chintan Zaveri, and Juergen Gall. Intention- based long-term human motion anticipation. InInternational Conference on 3D Vision, 2021. 6

    2021

  30. [38]

    Social diffusion: Long-term multiple human motion anticipation

    Julian Tanke, Linguang Zhang, Amy Zhao, Chengcheng Tang, Yujun Cai, Lezi Wang, Po-Chen Wu, Juergen Gall, and Cem Keskin. Social diffusion: Long-term multiple human motion anticipation. In International Conference on Com- puter Vision, 2023. 3, 5

    2023

  31. [39]

    Motionclip: Exposing human motion generation to clip space

    Guy Tevet, Brian Gordon, Amir Hertz, Amit H Bermano, and Daniel Cohen-Or. Motionclip: Exposing human motion generation to clip space. In European Conference on Com- puter Vision, 2022. 2

    2022

  32. [40]

    Human motion dif- fusion model

    Guy Tevet, Sigal Raab, Brian Gordon, Yonatan Shafir, Daniel Cohen-Or, and Amit H Bermano. Human motion dif- fusion model. arXiv preprint arXiv:2209.14916, 2022. 5

    2022 arXiv

  33. [41]

    Human motion diffu- sion model

    Guy Tevet, Sigal Raab, Brian Gordon, Yoni Shafir, Daniel Cohen-or, and Amit Haim Bermano. Human motion diffu- sion model. In International Conference on Learning Repre- sentations, 2023. 2, 5

    2023

  34. [42]

    Neural discrete representation learning

    Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning. Advances in neural information pro- cessing systems, 2017. 2

    2017

  35. [43]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017. 2, 3

    2017

  36. [44]

    Multi-person 3d motion prediction with multi- range transformers

    Jiashun Wang, Huazhe Xu, Medhini Narasimhan, and Xiao- long Wang. Multi-person 3d motion prediction with multi- range transformers. Advances in Neural Information Pro- cessing Systems, 2021. 3

    2021

  37. [45]

    Temporal and interac- tive modeling for efficient human-human motion generation

    Yabiao Wang, Shuo Wang, Jiangning Zhang, Ke Fan, Jiafu Wu, Zhengkai Jiang, and Yong Liu. Temporal and interac- tive modeling for efficient human-human motion generation. Conference on Computer Vision and Pattern Recognition ,

  38. [46]

    Inter-x: Towards versatile human- human interaction analysis

    Liang Xu, Xintao Lv, Yichao Yan, Xin Jin, Shuwen Wu, Congsheng Xu, Yifan Liu, Yizhou Zhou, Fengyun Rao, Xingdong Sheng, et al. Inter-x: Towards versatile human- human interaction analysis. In Conference on Computer Vi- sion and Pattern Recognition, 2024. 2, 3, 4, 5, 6

    2024

  39. [47]

    Generating human motion from textual descriptions with discrete representations

    Jianrong Zhang, Yangsong Zhang, Xiaodong Cun, Yong Zhang, Hongwei Zhao, Hongtao Lu, Xi Shen, and Ying Shan. Generating human motion from textual descriptions with discrete representations. In Conference on computer vi- sion and pattern recognition, 2023. 2

    2023

  40. [48]

    Motiondif- fuse: Text-driven human motion generation with diffusion model

    Mingyuan Zhang, Zhongang Cai, Liang Pan, Fangzhou Hong, Xinying Guo, Lei Yang, and Ziwei Liu. Motiondif- fuse: Text-driven human motion generation with diffusion model. arXiv preprint arXiv:2208.15001, 2022. 2

    2022 arXiv

  41. [49]

    Re- modiffuse: Retrieval-augmented motion diffusion model

    Mingyuan Zhang, Xinying Guo, Liang Pan, Zhongang Cai, Fangzhou Hong, Huirong Li, Lei Yang, and Ziwei Liu. Re- modiffuse: Retrieval-augmented motion diffusion model. In International Conference on Computer Vision, 2023. 2

    2023

  42. [50]

    Motion mamba: Efficient and long sequence motion generation

    Zeyu Zhang, Akide Liu, Ian Reid, Richard Hartley, Bohan Zhuang, and Hao Tang. Motion mamba: Efficient and long sequence motion generation. In European Conference on Computer Vision, 2024. 2

    2024

  43. [51]

    Attt2m: Text-driven human motion generation with multi- perspective attention mechanism

    Chongyang Zhong, Lei Hu, Zihao Zhang, and Shihong Xia. Attt2m: Text-driven human motion generation with multi- perspective attention mechanism. In International Confer- ence on Computer Vision, 2023. 2

    2023

  44. [52]

    Parco: Part- coordinating text-to-motion synthesis

    Qiran Zou, Shangyuan Yuan, Shian Du, Yu Wang, Chang Liu, Yi Xu, Jie Chen, and Xiangyang Ji. Parco: Part- coordinating text-to-motion synthesis. In European Confer- ence on Computer Vision, 2024. 2 10

    2024

Pith tools

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