Pith. sign in

REVIEW 3 major objections 5 minor 63 references

DreamX-Phi 1.0: Action-Conditioned Video World Model for Robotic Manipulation

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read DreamX-Phi 1.0 injects per-arm $\mathrm{SE}(3)$ transforms into attention to make predicted videos follow commanded bimanual actions, and reports first place on WorldArena 2.0 Track 1.

desk verdict A credible, clearly-written systems paper whose leaderboard ranking is probably real, but whose component-level claims are not yet supported and whose train/eval overlap on RoboTwin 2.0 is a genuine problem. read the letter →

arxiv 2608.13489 v1 pith:DT75QUVF submitted 2026-08-13 cs.CV cs.RO

DreamX Team , Rui Chen , Xiangxiang Chu , Geng Li , Jifan Li , Qingfeng Shi , Datao Tang , Jing Tang
show 2 more authors
Jun Wang Pengfei Zhang
This is my paper · ORCID
classification cs.CVcs.RO
keywords action-conditionedvideopredictionworldmodelsbimanualroboticmanipulationSE(3)geometricencodingrelativepositionalobject-centricsupervisiondiffusiontransformermodeldistillation
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 is building and testing an action-conditioned video world model for bimanual robotic manipulation: from one observed frame, a language instruction, and a prescribed sequence of end-effector poses and gripper states, the model generates the future video. Its central claim is that a rollout can be both visually realistic and faithful to the commanded action only if the commanded motion is injected into attention as per-arm $\mathrm{SE}(3)$ transforms and if training separately supervises scene geometry and the manipulated object's temporal consistency. On the official WorldArena 2.0 leaderboard snapshot, the authors report first place among 31 Track 1 entries with an EWMScore-P of 60.65, and a policy trained using the model as its rollout environment reaching 67.19% success on the Adjust Bottle task. If this holds, video world models become usable as learned simulators for evaluating manipulation actions before they are executed on a robot.

What carries the argument

The main mechanism is arm-grouped PRoPE-style geometric encoding: a variant of relative positional encoding that inserts relative $\mathrm{SE}(3)$ transforms directly into the attention computation, so that attention weights and value aggregation depend on the relative motion between frames rather than an absolute coordinate frame. DreamX-Phi assigns each arm a fixed contiguous group of attention heads, normalizes translation by a single motion-amplitude factor, and injects the gripper opening as a per-arm bias. Around this core sit three supervisions: a lightweight depth branch that predicts depth latents from the final transformer blocks and applies a latent-space MSE, SAM3-derived binary masks that reweight the flow-matching loss toward the manipulated object, and a frozen V-JEPA teacher whose projected features are aligned through a Gram-matrix relational loss gated by mask coverage and noise level. Distribution-matching distillation (DMD2) with an adversarial term then compresses the multi-step generator into a few-step student.

What would settle it

Run DreamX-Phi on a held-out manipulation benchmark whose episodes were generated from a simulator absent from its training corpus, or retrain the model with all RoboTwin 2.0 clips removed; if the EWMScore-P advantage over the nearest baseline disappears or the Track 2 success rate drops to the level of a model with no geometric or object-centric supervision, the central claim is falsified. A second check: ablate each component (arm-grouped PRoPE, depth branch, SAM3 weighting, V-JEPA loss) and look for a large, consistent drop on interaction-quality and trajectory-accuracy metrics; if removing them leaves the leaderboard scores nearly unchanged, the claimed mechanism is not load-bearing.

Watch

Extended reading notes

Core claim

DreamX-Phi 1.0 adapts the Wan2.2-TI2V-5B video diffusion transformer into a forward dynamics model for bimanual manipulation. Its central testable claim is that preserving the rigid-body structure of each arm's commanded trajectory—by expressing every end-effector pose relative to the first arm's initial pose, grouping attention heads by arm, and inserting those relative $\mathrm{SE}(3)$ transforms into query, key, and value projections—makes the generated video actually follow the commanded motion, while the auxiliary depth branch and object-centric supervision keep scene geometry and the grasped object physically coherent. The reported results are system-level: first place on WorldArena 2.0 Track 1 with EWMScore-P 60.65 at the August 12, 2026 snapshot, and 67.19% Adjust Bottle success on Track 2, tying for second. The paper also reports an offline WorldArena 1.0 Track 1 score of 76.88.

Load-bearing premise

The load-bearing premise is that the WorldArena 2.0 evaluation is a valid, non-overlapping test of action-conditioned dynamics; if the top scores come from training on the same RoboTwin 2.0 simulator distribution from which the held-out episodes are drawn, rather than from the paper's geometric and object-centric mechanisms, the central claim of generalizable action faithfulness is not established.

Editorial extensions

If this is right

  • If the reported leaderboard result is accurate, action-conditioned video models can serve as learned simulators for bimanual manipulation, letting candidate action sequences be scored in imagination before physical execution.
  • The arm-grouped $\mathrm{SE}(3)$ conditioning generalizes to any number of arms or articulated agents, provided each has an ordered rigid-body trajectory, which would extend controllable video prediction to multi-agent scenes.
  • The combination of depth supervision, mask-weighted losses, and relational feature alignment offers a reusable recipe for making video diffusion models respect object identity through contact interactions.
  • The DMD2 post-training result implies that a few-step student can retain much of the multi-step teacher's action fidelity, making the world model cheap enough for iterative planning loops.
  • On Track 2, the result that the π0.5 policy trained inside the model succeeds on held-out simulator episodes supports the use of learned video dynamics as a rollout environment for policy optimization.

Reading between the lines

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

  • The paper does not isolate which component produces the leaderboard gain; a natural untested hypothesis is that arm-grouped PRoPE is the main driver of trajectory accuracy, with depth and object supervision improving interaction quality.
  • Because the training corpus includes 25,000 RoboTwin 2.0 clips and the WorldArena 2.0 evaluation episodes are curated from RoboTwin 2.0 trajectories, part of the reported advantage may reflect overlap with the training distribution; a sharper test would withhold all RoboTwin 2.0 data or evaluate on a different simulator or real robot.
  • The Gram-matrix relational loss could be applied outside robotics to any video generation task that needs identity preservation, such as keeping a character's appearance consistent under heavy motion or occlusion.
  • A testable extension suggested by the method: use the same per-arm $\mathrm{SE}(3)$ conditioning but replace the RGB branch with a different base video model; if the gain persists, the mechanism is model-agnostic.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. DreamX-Phi 1.0 is an action-conditioned video world model built on Wan2.2-TI2V-5B for bimanual robotic manipulation. The paper's main proposals are (i) arm-grouped PRoPE geometric attention that injects per-arm SE(3) transformations into transformer attention, (ii) a robot-only optical-flow cue, (iii) an auxiliary depth branch supervised by DA3 depth latents, (iv) SAM3-mask weighting of the RGB flow-matching loss, (v) a frozen V-JEPA relational Gram-matrix loss for object consistency, and (vi) DMD2-based few-step distillation. The system is evaluated on WorldArena 2.0 and 1.0: the paper reports first place on WorldArena 2.0 Track 1 (EWMScore-P 60.65 in the August 12, 2026 snapshot), a tie for second on Track 2 (67.19% Adjust Bottle success), and an offline WorldArena 1.0 Track 1 score of 76.88. The paper explicitly limits its claims by noting that leaderboard scores do not isolate components and that matched ablations are still needed.

Significance. If the reported leaderboard results are valid, DreamX-Phi is a top-performing action-conditioned world model, and the proposed combination of geometry-aware conditioning and object-centric supervision is well motivated. The paper is transparent about snapshot anchoring, reports full 15-metric leaderboard tables, and acknowledges the lack of ablations. However, the evidence currently supports only a system-level ranking claim. The component-level contributions (PRoPE arm faithfulness, depth geometry, SAM3/V-JEPA object consistency) are not established by whole-system scores, and the training/evaluation overlap with RoboTwin 2.0 is not analyzed. The work is therefore a solid system report whose central causal claims require additional experiments.

major comments (3)
  1. [Section 3, Section 5.1] The action-conditioned fine-tuning pool includes 25,000 RoboTwin 2.0 clips (Table 1), while Section 5.1 states that both WorldArena benchmarks are 'curated and released by the WorldArena organizers from RoboTwin 2.0 trajectories.' The paper never states that the 1,000 Track 1 episodes and the held-out Track 2 Adjust Bottle episodes are disjoint from these 25,000 clips, nor does it report any overlap check. Without this, the small leaderboard margins (60.65 vs. Alpha-World's 60.13 on Track 1; a 67.19% tie with Lute on Track 2) could be explained by training/evaluation distribution overlap rather than by the proposed geometric and object-centric mechanisms. Please report an episode-level overlap analysis (e.g., by episode ID, trajectory hash, or scene seed) and either remove overlapping episodes from training or show that the rankings are unchanged when they are withheld. This is load-bearing for the claim of generalizable action faithfulness.
  2. [Section 6, Section 7] The conclusion attributes the result to 'arm-specific SE(3) conditioning with auxiliary depth and object-centric supervision,' and the abstract states that PRoPE 'ensures' arm faithfulness. However, Section 6 concedes that leaderboard scores do not isolate individual components, and Section 7 states that 'matched ablations are still needed.' No ablation of the PRoPE branch, robot-only flow cue, depth branch, SAM3 mask weighting, or V-JEPA loss is reported. As written, the component-level claims of Sections 4.2-4.4 are unsupported by whole-system scores. Please add matched ablations (each component removed or disabled) or explicitly re-scope the abstract and conclusion to describe an architecture motivated by these signals, with system-level evaluation only.
  3. [Section 5.3, Tables 2-3] The headline rank claims rest on a single leaderboard snapshot with no repeated runs, confidence intervals, or statistical comparisons. The Track 1 margin over Alpha-World is 0.52 EWMScore-P points out of 100, and Track 2 is a tie with Lute at 67.19%; these differences are not shown to exceed evaluation noise. If the leaderboard evaluation is deterministic, state that explicitly; if it is stochastic, provide at least three independent evaluation runs (or seeds) for the submitted checkpoint and nearest competitors, and report error bars. This is needed before 'ranks first' and 'ties for second' can be treated as robust system-level findings.
minor comments (5)
  1. [Section 4.2, Eq. (4)] The notation 'D_i = I_{d_h/4} \otimes A^k_{n(i)}' should be defined more carefully; it presumably denotes an identity matrix of dimension d_h/4, but the block layout and the meaning of d_h are not stated before use.
  2. [Section 4.4, Eq. (10)] The gate r_b uses \sigma_b, but \sigma_b is not defined before Eq. (10); define the per-sample flow-matching noise level or refer to the sampling distribution explicitly.
  3. [Section 4.4, Eq. (8)] The phrase 'normalized weight' could be misleading: after the mean normalization the weights have mean 1 over valid tokens rather than summing to 1; state this convention directly.
  4. [Section 5.1] For WorldArena 2.0 Track 2, the number of held-out Adjust Bottle evaluation episodes is not reported, so the 67.19% success rate lacks a sample-size context; please report n.
  5. [Section 4.2] The acronym PRoPE is not expanded and is only loosely tied to the cited 'Cameras as Relative Positional Encoding' (Li et al., 2025a); consider defining the acronym and naming the reference explicitly in the text.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: external leaderboard evidence; only a minor non-load-bearing self-citation.

full rationale

The paper's central evidence is an external leaderboard snapshot (WorldArena 2.0 at commit cb8f9c2) plus an offline WorldArena 1.0 score, not a derived quantity that reduces to its own inputs. The method equations define losses from actions, depth targets, SAM3 masks, and a frozen V-JEPA teacher; none of these is defined in terms of the EWMScore-P it later reports, and no fitted parameter is renamed as a prediction. The only self-citation occurs in Section 4.2, where the residual-branch design is credited to DreamX-World, but the branch is fully specified by Eqs. (4)-(5) with zero initialization, so the citation is not load-bearing. The RoboTwin 2.0 training/evaluation overlap (Sections 3 and 5.1) is a legitimate data-contamination and external-validity concern, but it is not a circular derivation: the paper never claims the leaderboard result follows from that overlap, and it explicitly concedes that 'matched ablations are still needed.' The benchmark components that resemble training signals (JEPA Similarity, Depth Accuracy) are not decisive margins and are better characterized as optimization/benchmark-alignment issues than as reductions. Overall, no circular step can be exhibited; the score reflects only the minor non-load-bearing self-citation.

Assumptions & free parameters 5 free parameters · 8 assumptions · 0 invented entities

The central performance claim depends on several borrowed models (Wan2.2, V-JEPA, SAM3, DA3, DMD2) whose reliability is assumed, and on hand-chosen hyperparameters whose values are not reported. No new physical entities are introduced; the only new object is the training architecture itself.

free parameters (5)
  • lambda_m = not reported
    Object-to-background ratio in Eq. 8 reweights the flow-matching loss for masked object tokens; the paper gives no value or sensitivity analysis.
  • M_min = not reported
    Minimum mask token count for a sample to contribute to the V-JEPA loss in Eq. 10; not reported.
  • sigma_max = not reported
    Upper bound on flow-matching noise for eligibility in Eq. 10; not reported.
  • lambda_adv = not reported
    Weight of the adversarial loss in Eq. 13 during DMD distillation; not reported.
  • M depth-block count = not reported
    Number of final transformer blocks replicated for the depth branch in Section 4.3; not reported.
assumptions (8)
  • standard math PRoPE relative positional encoding from Li et al. 2025a is valid for SE(3) transformations and for the robot-arm reuse in Eq. 4.
    Adopted without re-derivation.
  • standard math Flow matching objective from Lipman et al. 2023 applies to the video latent setup of Eq. 1.
    Adopted without re-derivation.
  • standard math DMD2 distillation from Yin et al. 2024a transfers to the conditioned video setting of Eqs. 11 to 13.
    Adopted without proof of convergence in the conditioned video setting.
  • domain assumption Wan2.2-TI2V-5B pretrained weights provide a usable video prior for robot manipulation.
    Section 4.1 builds the model on this base; if the prior does not transfer, all results change.
  • domain assumption SAM3 masks and DA3 depth maps are accurate enough to serve as supervision targets.
    Section 4.3 and Section 4.4 use these as training signals; noisy targets would weaken the proposed supervision.
  • domain assumption Frozen V-JEPA features encode object identity and physical consistency relevant to manipulation.
    Section 4.4 uses V-JEPA Gram matrices without evidence that the features are task-aligned.
  • domain assumption Heterogeneous corpora from egocentric video, real robot, and simulation improve action-conditioned prediction without harmful distribution shift.
    Section 3 constructs the corpus on this premise; no controlled study supports it.
  • ad hoc to paper The fixed contiguous head-group partition H_k in Section 4.2 is sufficient to preserve arm identity.
    This is a design choice with no analysis of sensitivity to head-group size or assignment.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DreamX-Phi 1.0: Action-Conditioned Video World Model for Robotic Manipulation." pith.science (2026). https://pith.science/paper/DT75QUVF

@misc{pith2026260813489,
  author       = {Pith},
  title        = {Pith review of: DreamX-Phi 1.0: Action-Conditioned Video World Model for Robotic Manipulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DT75QUVF}},
  note         = {Machine review of arXiv:2608.13489}
}
abstract

We present \textbf{DreamX-Phi 1.0}, an action-conditioned video world model for robotic manipulation that, given an observed frame, a language instruction, and a prescribed action sequence comprising end-effector poses and gripper states, predicts the resulting future observations. Yet realism alone does not guarantee faithfulness: a convincing rollout can still move the wrong arm or lose the manipulated object. To ensure the prediction respects each arm's commanded path, we inject per-arm $\mathrm{SE}(3)$ transformations into attention via \textbf{PRoPE-style geometric encoding}, preserving arm identity and rigid-motion structure. Action control alone does not fully constrain scene geometry or the evolution of small manipulated objects. We therefore add a lightweight \textbf{depth branch} for scene-level geometry and use \textbf{SAM3 masks} with a frozen \textbf{V-JEPA teacher} to maintain object consistency throughout grasping. We further distill the multi-step generator into a few-step student via distribution-matching distillation for efficient deployment. At the time of writing, \model{} achieves first place on Track~1 and second place on Track~2 of the WorldArena~2.0 Challenge. Our model and code will be publicly available.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 29 canonical work pages

  1. [1]

    AgiBot World 2026

    AgiBot World Team . AgiBot World 2026 . Hugging Face dataset, 2026. URL https://huggingface.co/datasets/agibot-world/AgiBotWorld2026

    work page 2026

  2. [2]

    V-JEPA 2: Self-supervised video models enable understanding, prediction and planning

    Mahmoud Assran, Adrien Bardes, David Fan, Quentin Garrido, Russell Howes, Mojtaba Komeili, Matthew Muckley, Ammar Rizvi, Claire Roberts, Koustuv Sinha, et al. V-JEPA 2: Self-supervised video models enable understanding, prediction and planning. arXiv preprint arXiv:2506.09985, 2025. URL https://arxiv.org/abs/2506.09985

    arXiv 2025

  3. [3]

    SAM 3 : Segment anything with concepts

    Nicolas Carion, Laura Gustafson, Yuan-Ting Hu, Shoubhik Debnath, Ronghang Hu, Didac Suris, Chaitanya Ryali, Kalyan Vasudev Alwala, Haitham Khedr, Andrew Huang, et al. SAM 3 : Segment anything with concepts. arXiv preprint arXiv:2511.16719, 2025. URL https://arxiv.org/abs/2511.16719

    arXiv 2025

  4. [4]

    WorldVLA : Towards autoregressive action world model

    Jun Cen, Chaohui Yu, Hangjie Yuan, Yuming Jiang, Siteng Huang, Jiayan Guo, Xin Li, Yibing Song, Hao Luo, Fan Wang, Deli Zhao, and Hao Chen. WorldVLA : Towards autoregressive action world model. arXiv preprint arXiv:2506.21539, 2025. URL https://arxiv.org/abs/2506.21539

    arXiv 2025

  5. [5]

    RoboTwin 2.0 : A scalable data generator and benchmark with strong domain randomization for robust bimanual robotic manipulation

    Tianxing Chen, Zanxin Chen, Baijun Chen, Zijian Cai, Yibin Liu, Zixuan Li, Qiwei Liang, Xianliang Lin, Yiheng Ge, Zhenyu Gu, et al. RoboTwin 2.0 : A scalable data generator and benchmark with strong domain randomization for robust bimanual robotic manipulation. arXiv preprint arXiv:2506.18088, 2025. URL https://arxiv.org/abs/2506.18088

    arXiv 2025

  6. [6]

    FlowWAM : Optical flow as a unified action representation for world action models

    Yixiang Chen, Peiyan Li, Yuan Xu, Qisen Ma, Jiabing Yang, Kai Wang, Jianhua Yang, Dong An, He Guan, Gaoteng Liu, Jianlou Si, Jun Huang, Jing Liu, Nianfeng Liu, Yan Huang, and Liang Wang. FlowWAM : Optical flow as a unified action representation for world action models. arXiv preprint arXiv:2607.13017, 2026. URL https://arxiv.org/abs/2607.13017

    arXiv 2026

  7. [8]

    Ego4D : Around the world in 3,000 hours of egocentric video

    Kristen Grauman, Andrew Westbury, Eugene Byrne, Zachary Chavis, Antonino Furnari, Rohit Girdhar, Jackson Hamburger, Hao Jiang, Miao Liu, Xingyu Liu, et al. Ego4D : Around the world in 3,000 hours of egocentric video. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022. URL https://arxiv.org/abs/2110.07058

    arXiv 2022

  8. [9]

    Unified 4D world action modeling from video priors with asynchronous denoising

    Jun Guo, Qiwei Li, Peiyan Li, Zilong Chen, Nan Sun, Yifei Su, Heyun Wang, Yuan Zhang, Xinghang Li, and Huaping Liu. Unified 4D world action modeling from video priors with asynchronous denoising. arXiv preprint arXiv:2604.26694, 2026. URL https://arxiv.org/abs/2604.26694

    arXiv 2026

Show all 63 references
  1. [10]

    Vid2World : Crafting video diffusion models to interactive world models

    Siqiao Huang, Jialong Wu, Qixing Zhou, Shangchen Miao, and Mingsheng Long. Vid2World : Crafting video diffusion models to interactive world models. arXiv preprint arXiv:2505.14357, 2025. URL https://arxiv.org/abs/2505.14357

    2025

  2. [11]

    DROID : A large-scale in-the-wild robot manipulation dataset

    Alexander Khazatsky, Karl Pertsch, Suraj Nair, Ashwin Balakrishna, Sudeep Dasari, Siddharth Karamcheti, Soroush Nasiriany, Mohan Kumar Srirama, Lawrence Yunliang Chen, Kirsty Ellis, et al. DROID : A large-scale in-the-wild robot manipulation dataset. arXiv preprint arXiv:2403....

    2024 arXiv

  3. [13]

    Cosmos Policy : Fine-tuning video models for visuomotor control and planning

    Moo Jin Kim, Yihuai Gao, Tsung-Yi Lin, Yen-Chen Lin, Yunhao Ge, Grace Lam, Percy Liang, Shuran Song, Ming-Yu Liu, Chelsea Finn, and Jinwei Gu. Cosmos Policy : Fine-tuning video models for visuomotor control and planning. In International Conference on Learning Representations,...

    2026 arXiv

  4. [15]

    Cameras as relative positional encoding

    Ruilong Li, Brent Yi, Junchen Liu, Hang Gao, Yi Ma, and Angjoo Kanazawa. Cameras as relative positional encoding. Advances in Neural Information Processing Systems, 38, 2025 a . URL https://arxiv.org/abs/2507.10496

    2025

  5. [17]

    Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le. Flow matching for generative modeling. In International Conference on Learning Representations, 2023. URL https://arxiv.org/abs/2210.02747

    2023 arXiv

  6. [18]

    GTA : A geometry-aware attention mechanism for multi-view transformers

    Takeru Miyato, Bernhard Jaeger, Max Welling, and Andreas Geiger. GTA : A geometry-aware attention mechanism for multi-view transformers. In International Conference on Learning Representations, 2024. URL https://arxiv.org/abs/2310.10375

    2024 arXiv

  7. [20]

    Cosmos3-DROID

    NVIDIA Corporation . Cosmos3-DROID . Hugging Face dataset, 2026. URL https://huggingface.co/datasets/nvidia/Cosmos3-DROID

    2026

  8. [22]

    AVID : Adapting video diffusion models to world models

    Marc Rigter, Tarun Gupta, Agrin Hilmkil, and Chao Ma. AVID : Adapting video diffusion models to world models. Reinforcement Learning Journal, 6: 0 737--764, 2025. URL https://arxiv.org/abs/2410.12822

    2025 arXiv

  9. [23]

    WorldArena : A unified benchmark for evaluating perception and functional utility of embodied world models

    Yu Shang, Zhuohang Li, Yiding Ma, Weikang Su, Xin Jin, Ziyou Wang, Lei Jin, Xin Zhang, Yinzhou Tang, Haisheng Su, et al. WorldArena : A unified benchmark for evaluating perception and functional utility of embodied world models. arXiv preprint arXiv:2602.08971, 2026 a . URL ht...

    2026

  10. [24]

    WorldArena 2.0 : Extending embodied world model benchmarking on modality, functionality and platform

    Yu Shang, Yinzhou Tang, Yiding Ma, Zhuohang Li, Lei Jin, Weikang Su, Xin Jin, Zhaolu Wang, Ziyou Wang, Xin Zhang, et al. WorldArena 2.0 : Extending embodied world model benchmarking on modality, functionality and platform. arXiv preprint arXiv:2605.17912, 2026 b . URL https://...

    2026 arXiv

  11. [26]

    InternData-A1 : Pioneering high-fidelity synthetic data for pre-training generalist policy

    Yang Tian, Yuyin Yang, Yiman Xie, Zetao Cai, Xu Shi, Ning Gao, Hangxu Liu, Xuekun Jiang, Zherui Qiu, Feng Yuan, et al. InternData-A1 : Pioneering high-fidelity synthetic data for pre-training generalist policy. arXiv preprint arXiv:2511.16651, 2025. URL https://arxiv.org/abs/2...

    2025

  12. [27]

    Wan2.2-TI2V-5B

    Wan Team . Wan2.2-TI2V-5B . Model card and released model weights, 2025. URL https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B

    2025

  13. [29]

    iVideoGPT : Interactive VideoGPTs are scalable world models

    Jialong Wu, Shaofeng Yin, Ningya Feng, Xu He, Dong Li, Jianye Hao, and Mingsheng Long. iVideoGPT : Interactive VideoGPTs are scalable world models. Advances in Neural Information Processing Systems, 37, 2024. URL https://arxiv.org/abs/2405.15223

    2024 arXiv

  14. [30]

    RoboCOIN : An open-sourced bimanual robotic data collection for integrated manipulation

    Shihan Wu, Xuecheng Liu, Shaoxuan Xie, Pengwei Wang, Xinghang Li, Bowen Yang, Zhe Li, Kai Zhu, Hongyu Wu, Yiheng Liu, et al. RoboCOIN : An open-sourced bimanual robotic data collection for integrated manipulation. arXiv preprint arXiv:2511.17441, 2025. URL https://arxiv.org/ab...

    2025 arXiv

  15. [31]

    OSCAR : Omni-embodiment action-conditioned world model for robotics

    Zhuoyuan Wu and Jun Gao. OSCAR : Omni-embodiment action-conditioned world model for robotics. arXiv preprint arXiv:2606.04463, 2026. URL https://arxiv.org/abs/2606.04463

    2026 arXiv

  16. [32]

    Learning interactive real-world simulators

    Sherry Yang, Yilun Du, Kamyar Ghasemipour, Jonathan Tompson, Leslie Kaelbling, Dale Schuurmans, and Pieter Abbeel. Learning interactive real-world simulators. In International Conference on Learning Representations, 2024. URL https://arxiv.org/abs/2310.06114

    2024 arXiv

  17. [34]

    Tianwei Yin, Micha \"e l Gharbi, Taesung Park, Richard Zhang, Eli Shechtman, Fr \'e do Durand, and William T. Freeman. Improved distribution matching distillation for fast image synthesis. In Advances in Neural Information Processing Systems, volume 37, pages 47455--47487, 202...

    2024 arXiv

  18. [35]

    Freeman, and Taesung Park

    Tianwei Yin, Micha \"e l Gharbi, Richard Zhang, Eli Shechtman, Fr \'e do Durand, William T. Freeman, and Taesung Park. One-step diffusion with distribution matching distillation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6613--...

    2024 arXiv

  19. [36]

    IRASim : A fine-grained world model for robot manipulation

    Fangqi Zhu, Hongtao Wu, Song Guo, Yuxiao Liu, Chilam Cheang, and Tao Kong. IRASim : A fine-grained world model for robot manipulation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 9834--9844, 2025. URL https://arxiv.org/abs/2406.14540

    2025 arXiv

  20. [37]

    arXiv preprint arXiv:2503.20314 , year=

    arXiv

  21. [38]

    2025 , howpublished=

    2025

  22. [39]

    International Conference on Learning Representations , year=

    Flow Matching for Generative Modeling , author=. International Conference on Learning Representations , year=

  23. [40]

    2024 , url=

    Miyato, Takeru and Jaeger, Bernhard and Welling, Max and Geiger, Andreas , booktitle=. 2024 , url=

    2024

  24. [41]

    Advances in Neural Information Processing Systems , volume=

    Cameras as Relative Positional Encoding , author=. Advances in Neural Information Processing Systems , volume=. 2025 , url=

    2025

  25. [42]

    arXiv preprint arXiv:2606.16993 , year=

  26. [43]

    2026 , url=

    Chen, Yixiang and Li, Peiyan and Xu, Yuan and Ma, Qisen and Yang, Jiabing and Wang, Kai and Yang, Jianhua and An, Dong and Guan, He and Liu, Gaoteng and Si, Jianlou and Huang, Jun and Liu, Jing and Liu, Nianfeng and Huang, Yan and Wang, Liang , journal=. 2026 , url=

    2026

  27. [44]

    Guo, Jun and Li, Qiwei and Li, Peiyan and Chen, Zilong and Sun, Nan and Su, Yifei and Wang, Heyun and Zhang, Yuan and Li, Xinghang and Liu, Huaping , journal=. Unified. 2026 , url=

    2026

  28. [45]

    2025 , url=

    Carion, Nicolas and Gustafson, Laura and Hu, Yuan-Ting and Debnath, Shoubhik and Hu, Ronghang and Suris, Didac and Ryali, Chaitanya and Alwala, Kalyan Vasudev and Khedr, Haitham and Huang, Andrew and others , journal=. 2025 , url=

    2025

  29. [46]

    2025 , url=

    Assran, Mahmoud and Bardes, Adrien and Fan, David and Garrido, Quentin and Howes, Russell and Komeili, Mojtaba and Muckley, Matthew and Rizvi, Ammar and Roberts, Claire and Sinha, Koustuv and others , journal=. 2025 , url=

    2025

  30. [47]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    One-step Diffusion with Distribution Matching Distillation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=. 2024 , url=

    2024

  31. [48]

    Advances in Neural Information Processing Systems , volume=

    Improved Distribution Matching Distillation for Fast Image Synthesis , author=. Advances in Neural Information Processing Systems , volume=. 2024 , url=

    2024

  32. [49]

    2025 , url=

    Chen, Tianxing and Chen, Zanxin and Chen, Baijun and Cai, Zijian and Liu, Yibin and Li, Zixuan and Liang, Qiwei and Lin, Xianliang and Ge, Yiheng and Gu, Zhenyu and others , journal=. 2025 , url=

    2025

  33. [50]

    2022 , url=

    Grauman, Kristen and Westbury, Andrew and Byrne, Eugene and Chavis, Zachary and Furnari, Antonino and Girdhar, Rohit and Hamburger, Jackson and Jiang, Hao and Liu, Miao and Liu, Xingyu and others , booktitle=. 2022 , url=

    2022

  34. [51]

    2025 , url=

    Tian, Yang and Yang, Yuyin and Xie, Yiman and Cai, Zetao and Shi, Xu and Gao, Ning and Liu, Hangxu and Jiang, Xuekun and Qiu, Zherui and Yuan, Feng and others , journal=. 2025 , url=

    2025

  35. [52]

    2024 , url=

    Khazatsky, Alexander and Pertsch, Karl and Nair, Suraj and Balakrishna, Ashwin and Dasari, Sudeep and Karamcheti, Siddharth and Nasiriany, Soroush and Srirama, Mohan Kumar and Chen, Lawrence Yunliang and Ellis, Kirsty and others , journal=. 2024 , url=

    2024

  36. [53]

    2025 , url=

    Wu, Shihan and Liu, Xuecheng and Xie, Shaoxuan and Wang, Pengwei and Li, Xinghang and Yang, Bowen and Li, Zhe and Zhu, Kai and Wu, Hongyu and Liu, Yiheng and others , journal=. 2025 , url=

    2025

  37. [54]

    2026 , url=

    Wang, Jianyi and Lin, Shanchuan and Lin, Zhijie and Ren, Yuxi and Wei, Meng and Yue, Zongsheng and Zhou, Shangchen and Chen, Hao and Zhao, Yang and Yang, Ceyuan and Xiao, Xuefeng and Loy, Chen Change and Jiang, Lu , booktitle=. 2026 , url=

    2026

  38. [55]

    2026 , url=

    Shang, Yu and Li, Zhuohang and Ma, Yiding and Su, Weikang and Jin, Xin and Wang, Ziyou and Jin, Lei and Zhang, Xin and Tang, Yinzhou and Su, Haisheng and others , journal=. 2026 , url=

    2026

  39. [56]

    2026 , url=

    Shang, Yu and Tang, Yinzhou and Ma, Yiding and Li, Zhuohang and Jin, Lei and Su, Weikang and Jin, Xin and Wang, Zhaolu and Wang, Ziyou and Zhang, Xin and others , journal=. 2026 , url=

    2026

  40. [57]

    2024 , url=

    Wu, Jialong and Yin, Shaofeng and Feng, Ningya and He, Xu and Li, Dong and Hao, Jianye and Long, Mingsheng , journal=. 2024 , url=

    2024

  41. [58]

    2025 , url=

    Zhu, Fangqi and Wu, Hongtao and Guo, Song and Liu, Yuxiao and Cheang, Chilam and Kong, Tao , booktitle=. 2025 , url=

    2025

  42. [59]

    2025 , url=

    Huang, Siqiao and Wu, Jialong and Zhou, Qixing and Miao, Shangchen and Long, Mingsheng , journal=. 2025 , url=

    2025

  43. [60]

    arXiv preprint arXiv:2606.02800 , year=

    arXiv

  44. [61]

    2025 , url=

    Rigter, Marc and Gupta, Tarun and Hilmkil, Agrin and Ma, Chao , journal=. 2025 , url=

    2025

  45. [62]

    International Conference on Learning Representations , year=

    Learning Interactive Real-World Simulators , author=. International Conference on Learning Representations , year=

  46. [63]

    2025 , url=

    Jang, Joel and Ye, Seonghyeon and Lin, Zongyu and Xiang, Jiannan and Bjorck, Johan and Fang, Yu and Hu, Fengyuan and Huang, Spencer and Kundalia, Kaushil and Lin, Yen-Chen and others , booktitle=. 2025 , url=

    2025

  47. [64]

    arXiv preprint arXiv:2502.04296 , year=

    Learning Real-World Action-Video Dynamics with Heterogeneous Masked Autoregression , author=. arXiv preprint arXiv:2502.04296 , year=

    arXiv

  48. [65]

    2026 , url=

    Wu, Zhuoyuan and Gao, Jun , journal=. 2026 , url=

    2026

  49. [66]

    arXiv preprint arXiv:2607.22535 , year=

    Robot-Factored World Models via Robot Rendering , author=. arXiv preprint arXiv:2607.22535 , year=

    arXiv

  50. [67]

    2026 , url=

    Kim, Moo Jin and Gao, Yihuai and Lin, Tsung-Yi and Lin, Yen-Chen and Ge, Yunhao and Lam, Grace and Liang, Percy and Song, Shuran and Liu, Ming-Yu and Finn, Chelsea and Gu, Jinwei , booktitle=. 2026 , url=

    2026

  51. [68]

    arXiv preprint arXiv:2503.00200 , year=

    Unified Video Action Model , author=. arXiv preprint arXiv:2503.00200 , year=

    arXiv

  52. [69]

    doi:10.48550/arXiv.2504.16054 , url=

    arXiv preprint arXiv:2504.16054 , year=. doi:10.48550/arXiv.2504.16054 , url=

    doi arXiv

  53. [70]

    2025 , url=

    Cen, Jun and Yu, Chaohui and Yuan, Hangjie and Jiang, Yuming and Huang, Siteng and Guo, Jiayan and Li, Xin and Song, Yibing and Luo, Hao and Wang, Fan and Zhao, Deli and Chen, Hao , journal=. 2025 , url=

    2025

  54. [71]

    arXiv preprint arXiv:2601.21998 , year=

    Causal World Modeling for Robot Control , author=. arXiv preprint arXiv:2601.21998 , year=

    arXiv

  55. [72]

    arXiv preprint arXiv:2602.15922 , year=

    World Action Models Are Zero-Shot Policies , author=. arXiv preprint arXiv:2602.15922 , year=

    arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.