Pith. sign in

REVIEW 3 major objections 5 minor 49 references

The paper claims that propagating a retrieved contact point through a predicted object-pose trajectory gives state-consistent post-contact guidance, raising six-task simulation success from 52.2% (DP3) and 57.8% (AffordDP) to 70.0% on ManiS

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-05 00:11 UTC pith:VWI634V2

load-bearing objection Genuinely new synthesis with a clean ablation; the propagation mechanism is plausible, but the frame definition in Eq. (7) and the untested non-accumulation claim keep the 'state-consistent' story from being fully sold. the 3 major comments →

arxiv 2608.01603 v1 pith:VWI634V2 submitted 2026-08-03 cs.RO

AffordTrajDP: Dynamic Affordance-Guided Visuomotor Policy Learning for Robotic Manipulation

Gaoyuan Wu , Ziyu Shan , Haoyang Du , Yuyao Jiang , Ziwei Wang This is my paper
classification cs.RO
keywords affordance-guided manipulationdynamic affordance trajectorydiffusion policyobject-centric representationsSE(3) pose predictionimitation learningrobotic manipulationManiSkill3
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Robotic policies that imitate demonstrations often use a static contact point to tell the policy where to touch an object. The paper argues that this prior goes stale as soon as contact nudges the object, so precision tasks like inserting a plug fail. It proposes AffordTrajDP, which predicts how the object's pose will evolve and rigidly carries the contact point along that predicted trajectory, producing contact guidance that tracks the object through the interaction. On six ManiSkill3 tasks the method reaches 70.0% average success versus 52.2% for the point-cloud baseline and 57.8% for the static-affordance baseline, with its largest margins on insertion tasks; seven real-world tasks across two arms reproduce the pattern. The point is that a learned policy should be conditioned not on where contact was, but on where contact will be.

Core claim

AffordTrajDP's central claim is that the missing ingredient in affordance-guided policies is a conditioning signal that jointly preserves where to contact and how that contact evolves as the object moves. It retrieves a single anchor contact point on a semantically similar source object, transfers it to the target through dense visual correspondence, predicts the target object's future SE(3) poses with a DDIM-based diffusion model conditioned on tracked pose history, and then propagates the anchor through those predicted poses by rigid transformation to form a dynamic affordance trajectory. This time-varying contact prior is fed, alongside the point cloud and proprioception, into a 3D diffus

What carries the argument

The central mechanism is the dynamic affordance trajectory τ_{t:t+H} = (p_t, ..., p_{t+H}), with each future contact point obtained as p_{t+k} = T_{t+k} c_target, where c_target is the retrieved anchor contact point and T_{t+k} are predicted SE(3) object poses. The pose trajectory is produced by a DDIM denoiser conditioned on tracked pose history; DDIM's deterministic reverse process keeps the propagated affordance stable across denoising steps. This mechanism turns a once-transferred static contact prior into a moving constraint that stays attached to the object as its pose changes.

Load-bearing premise

The load-bearing premise is that the predicted future object poses are accurate enough that rigidly carrying the contact point through them keeps it on the true contact region; if pose tracking or trajectory prediction drifts, the dynamic guidance drifts along with it, just on a moving target.

What would settle it

Compare AffordTrajDP against AffordDP on an insertion task with injected pose noise of known magnitude added to the tracked history, e.g., 1-4 mm translation and 1-2 degrees rotation, matching the post-contact tracking error the paper measures. If the dynamic-affordance advantage over the static prior disappears or inverts under matched perturbation, then state-consistent propagation is not what is carrying the result.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • Static affordance priors become stale once contact moves the object; propagating the anchor through predicted poses prevents the post-contact drift that dominates failures in tasks such as PegInsertion and PlugCharger.
  • Semantic contact retrieval and pose-trajectory prediction are each necessary but individually insufficient; removing either component drops performance, and pose-only conditioning can even hurt on unseen instances when the anchor is not semantically grounded.
  • The dynamic affordance signal transfers across embodiments: the same approach improves success over DP3, AffordDP, and SPOT on both a Galaxea A1 arm and a UR7e arm.
  • Expressing poses in the target object frame makes the guidance layout-invariant, so object placement perturbations and unseen object instances are handled without retraining.
  • The largest gains appear on the most precision-critical tasks (AdapterInsertion, USB Insertion), suggesting the signal addresses systematic contact-stage failures rather than generic perception noise.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the paper's observed non-accumulation of pose-prediction error holds under injected noise and heavier occlusion, the same propagation mechanism could keep affordances valid for much longer horizons; the paper itself flags this as future work.
  • Because propagation is a rigid transform of a single anchor, the method does not require a learned contact model: any future pose estimator with calibrated uncertainty could replace the diffusion trajectory model, making the design modular.
  • The principle of conditioning on where contact will be rather than where contact was could extend beyond single contact points to oriented affordance frames or grasp poses, and to tasks with actively moving targets.
  • A sharper test of the paper's necessity claim would be an ablated variant that propagates a semantically ungrounded anchor: the StackCube unseen result already hints such a variant underperforms, but measuring it explicitly would isolate the role of semantic grounding.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes AffordTrajDP, a visuomotor policy that augments diffusion-policy imitation learning with a time-varying affordance signal. A static contact point is retrieved from an object memory and transferred to the target via dense visual correspondence; a diffusion model predicts the target object's future SE(3) pose trajectory; the anchor contact point is then propagated through the predicted poses to form an affordance trajectory that conditions a DP3-style action generator. The method is evaluated on six ManiSkill3 tasks and seven real-world tasks on two robot arms, with ablations isolating the contribution of the semantic contact point and the pose-trajectory propagation. The reported results are consistently positive, with the largest gains on precision-critical insertion tasks.

Significance. If the central claim holds, the paper makes a useful conceptual advance: static affordance priors become inconsistent after contact, and jointly preserving 'where to contact' and 'how the contact evolves' is a meaningful, under-specified design axis in affordance-guided manipulation. The paper's strengths include (i) a clean ablation design that separately enables the semantic anchor and the pose-trajectory propagation, (ii) evaluation across two embodiments and seen/unseen object instances, and (iii) an unusually candid discussion in Sec. 5 that labels the non-accumulation of pose-prediction error as a hypothesis rather than an established result. The main weaknesses are the unresolved coordinate-frame definition of the propagation step and the fact that the central 'state-consistent' mechanism is supported only by a single-task correlational error analysis.

major comments (3)
  1. [§3.4, Eq. (7), and §4.1 implementation details] The propagation equation p_{t+k} = T_{t+k} c_tgt is underspecified in a load-bearing way. Eq. (4) defines the training target as the source-object pose expressed in the target-object frame, T_src_tgt(t) = (T_tgt_W(t))^{-1} T_src_W(t), which is a relative pose between two objects, not the target object's pose in a fixed frame. The implementation notes then say that for single-interaction-target tasks 'the source and target objects in Eq. (4) both refer to the same object category' and that 'the target object coordinate frame at time t0 is defined by the pose estimate from FoundationPose at the first observation frame.' This does not resolve the frame of T_{t+k} in Eq. (7): is T_{t+k} the object's pose relative to the t0 frame, or its absolute world/camera-frame pose? Is c_tgt expressed in the object frame, the t0 frame, or the camera frame? If T_{t+k} is relative to t0, then p_{t+k} is al
  2. [§4.2, Fig. 6, and §5] The paper's own Sec. 5 states that the non-accumulation of pose-prediction error is 'a hypothesis for future work rather than an established result' and that the supporting analysis is 'a correlational, single-task comparison rather than a controlled ablation.' This limitation directly bears on the central claim of state-consistent post-contact guidance. Under estimated history, the translation error at k=4 is 3.77 mm, and FoundationPose tracking error rises after contact to a magnitude comparable to the prediction error. Since the propagation mechanism in Eq. (7) inherits exactly these errors, the claim that the affordance trajectory 'tracks the target object's evolving pose' is not yet established for adverse conditions. The authors should either add the noise-injection/occlusion experiments they propose in Sec. 5, or explicitly restrict the central claim to the tested settings. As wri
  3. [§4.3, Tables 2 and 3] The real-world evaluation reports 30 trials per condition with point counts (e.g., 12/30, 6/30) and no confidence intervals, significance tests, or repeated trials. The paper concludes 'consistently superior success rates' and 'consistently outperforms' from these counts. For example, in Table 3 the difference between AffordDP (25/30) and Ours (29/30) on PickCup is 4 trials; without a binomial confidence interval or a paired protocol this is not statistically distinguishable. The aggregate 'Avg. SR' across 180 or 120 trials mixes tasks and seen/unseen conditions, which further complicates interpretation. I recommend adding confidence intervals or exact tests per condition, or tempering the per-task comparative claims. This is not a fatal flaw, but it is load-bearing for the empirical validation, especially because the real-world results are presented as a headline contribution.
minor comments (5)
  1. [Abstract and §1] 'Outperforming strong baselines by up to 17.8%' should be '17.8 percentage points' (from 52.2% to 70.0%), not '17.8%'.
  2. [Eq. (2)] The tilde notation c_tgt = T c_src is inconsistent with Eq. (7), where c_tgt is used without a tilde for the homogeneous coordinate. Please unify the notation.
  3. [References] References [3] and [4] appear to be the same paper (3D AffordanceNet, CVPR 2021). Please deduplicate.
  4. [Figures and captions] The typesetting has missing spaces in several captions (e.g., 'OverviewofAffordTrajDPandcomparisonwithpreviousmethods' in Fig. 1, 'Overview' in Fig. 2). These should be fixed.
  5. [§5] The discussion phrase 'particularly valuable under limited demonstration data' is not directly evidenced by a data-efficiency plot; consider either adding such an experiment or softening this wording.

Circularity Check

0 steps flagged

No significant circularity: the affordance trajectory is generated from external pose predictions and retrieval, and all central results are evaluated on held-out task success.

full rationale

The derivation chain is self-contained. The anchor contact point is retrieved from an object memory and transferred via dense correspondence (Sec. 3.3); the affordance trajectory is then computed by propagating this point with the predicted SE(3) pose trajectory (Eq. 7), not by fitting the trajectory to the policy output. The pose trajectory model is trained on demonstration pose tracks (Eqs. 4-6), and the action policy is trained with the standard denoising objective (Eq. 9), with success evaluated on held-out ManiSkill3 episodes and real-world seen/unseen instances (Tables 1-3). The ablations (+Center, +Contact, +Traj vs. Ours) are independent conditioning variants and are not constructed to force the central comparison. The only author-overlapping citations (Refs. 27, 28) appear in related-work context and are not used to justify any premise or to rule out alternatives. The paper's Sec. 5 explicitly labels the non-accumulation of pose-prediction error as a hypothesis for future work based on a correlational single-task comparison; that is an evidentiary limitation, not a circular step. No equation or construction reduces a claimed prediction to its own input.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 0 invented entities

The paper introduces no new physical entities; the "dynamic affordance trajectory" is a computed conditioning signal derived from existing pose predictions and a retrieved point. The central assumptions are about pose tracking and retrieval reliability.

free parameters (3)
  • Pose forecast horizon H = 4
    The trajectory diffusion model predicts H=4 future poses, which also fixes the length of the propagated affordance trajectory; chosen by hand, not tuned or derived.
  • Action chunk length = 8
    The action policy predicts 16 steps but executes 8; a design choice that affects closed-loop behavior.
  • Number of demonstrations = 30 (or 50 for insertion)
    Demonstration count is a data-scale choice; the same for all methods so it does not favor the proposed method, but it is a free choice.
axioms (4)
  • domain assumption The manipulated object is rigid and its 6D pose can be tracked by FoundationPose with bounded error.
    Sec. 3.4 trains the pose model on FoundationPose-tracked trajectories; if tracking fails or the object is non-rigid, the propagation in Eq. (7) is invalid.
  • domain assumption CLIP-based instance retrieval and dense visual correspondence transfer the contact point to novel instances correctly.
    Sec. 3.3 retrieves the anchor by cosine similarity and back-projects the matched pixel; errors here shift the anchor and propagate into the trajectory.
  • standard math DDIM denoising with quaternion renormalization yields valid SE(3) samples.
    Used in Eq. (5)-(7) for pose prediction; standard in the diffusion literature.
  • domain assumption The source-object pose expressed in the target frame (Eq. 4) provides a layout-invariant training signal that generalizes across scenes.
    This normalization assumes that relative pose, not absolute layout, determines the affordance trajectory.

pith-pipeline@v1.3.0-daily-deepseek · 12658 in / 12033 out tokens · 118094 ms · 2026-08-05T00:11:04.377412+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of AffordTrajDP: Dynamic Affordance-Guided Visuomotor Policy Learning for Robotic Manipulation." pith.science (2026). https://pith.science/paper/VWI634V2

@misc{pith2026260801603,
  author       = {Pith},
  title        = {Pith review of: AffordTrajDP: Dynamic Affordance-Guided Visuomotor Policy Learning for Robotic Manipulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VWI634V2}},
  note         = {Machine review of arXiv:2608.01603}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Affordance-guided imitation learning has shown impressive performance in robotic manipulation tasks by compressing visual perception into task-specific geometric constraints (e.g., fixed contact points). However, the commonly used static affordances can become inconsistent in precision-critical tasks or under object location perturbations, leading to post-contact trajectory drift. To address this issue, we propose AffordTrajDP, a dynamic framework that constructs affordance trajectories via object-centric temporal propagation to guide the progressive manipulation process. Specifically, given an RGB-D observation, our core insight is that a retrieved anchor affordance, which captures the desired contact point between the end-effector and the target object, can be propagated forward via affordance propagation, using the object's SE(3) pose as a natural propagation medium, to yield an affordance trajectory that provides temporally consistent, state-aware guidance throughout execution. AffordTrajDP achieves 70.0% average success rate on ManiSkill3, outperforming strong baselines by up to 17.8%. Real-world experiments on Galaxea A1 and UR7e robotic arms, covering StackCube, PickCup, AdapterInsertion, Ring-on-Peg, Put-in-Bowl, and USB Insertion, further validate robustness under object placement variations and appearance changes, with seen and unseen object instances evaluated on Galaxea A1, and ablations confirm the contribution of each proposed component.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

49 extracted references · 34 canonical work pages

  1. [1]

    Mamba policy: Towards efficient 3d diffusion policy with hybrid selective state models, 2025

    Jiahang Cao, Qiang Zhang, Jingkai Sun, Jiaxu Wang, Hao Cheng, Yulin Li, Jun Ma, Kun Wu, Zhiyuan Xu, Yecheng Shao, Wen Zhao, Gang Han, Yijie Guo, and Renjing Xu. Mamba policy: Towards efficient 3d diffusion policy with hybrid selective state models, 2025

    work page 2025

  2. [2]

    Cheng Chi, Zhenjia Xu, Siyuan Feng, Eric Cousineau, Yilun Du, Benjamin Burchfiel, Russ Tedrake, andShuranSong.Diffusionpolicy: Visuomotorpolicylearningviaactiondiffusion.The International Journal of Robotics Research, 44(10-11):1684–1704, 2025

    work page 2025

  3. [3]

    3d affordancenet: A benchmark for visual object affordance understanding

    Shengheng Deng, Xun Xu, Chaozheng Wu, Ke Chen, and Kui Jia. 3d affordancenet: A benchmark for visual object affordance understanding. Inproceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1778–1787, 2021

    work page 2021

  4. [4]

    3d affordancenet: A benchmark for visual object affordance understanding, 2021

    Shengheng Deng, Xun Xu, Chaozheng Wu, Ke Chen, and Kui Jia. 3d affordancenet: A benchmark for visual object affordance understanding, 2021

    work page 2021

  5. [5]

    Self-supervised 6d object pose estimation for robot manipulation, 2020

    Xinke Deng, Yu Xiang, Arsalan Mousavian, Clemens Eppner, Timothy Bretl, and Dieter Fox. Self-supervised 6d object pose estimation for robot manipulation, 2020

    work page 2020

  6. [6]

    Deep object-centric representations forgeneralizablerobotlearning

    Coline Devin, Pieter Abbeel, Trevor Darrell, and Sergey Levine. Deep object-centric representations forgeneralizablerobotlearning. In2018 IEEE International Conference on Robotics and Automation (ICRA), pages 7111–7118. IEEE, 2018

    work page 2018

  7. [7]

    End-to-end affordance learning for robotic manipulation, 2022

    Yiran Geng, Boshi An, Haoran Geng, Yuanpei Chen, Yaodong Yang, and Hao Dong. End-to-end affordance learning for robotic manipulation, 2022

    work page 2022

  8. [8]

    Act3d: 3d feature field transformers for multi-task robotic manipulation, 2023

    Theophile Gervet, Zhou Xian, Nikolaos Gkanatsios, and Katerina Fragkiadaki. Act3d: 3d feature field transformers for multi-task robotic manipulation, 2023

    2023

  9. [9]

    Fsag: Enhancing human-to-dexterous-hand finger-specific affordance grounding via diffusion models.arXiv preprint arXiv:2601.08246, 2026

    Yifan Han, Pengfei Yi, Junyan Li, Hanqing Wang, Gaojing Zhang, Qi Peng Liu, and Wenzhao Lian. Fsag: Enhancing human-to-dexterous-hand finger-specific affordance grounding via diffusion models.arXiv preprint arXiv:2601.08246, 2026. 12

    arXiv 2026

  10. [10]

    Spot: Se(3) pose trajectory diffusion for object-centric manipulation, 2025

    Cheng-Chun Hsu, Bowen Wen, Jie Xu, Yashraj Narang, Xiaolong Wang, Yuke Zhu, Joydeep Biswas, and Stan Birchfield. Spot: Se(3) pose trajectory diffusion for object-centric manipulation, 2025

    work page 2025

  11. [11]

    3d diffuser actor: Policy diffusion with 3d scene representations, 2024

    Tsung-Wei Ke, Nikolaos Gkanatsios, and Katerina Fragkiadaki. 3d diffuser actor: Policy diffusion with 3d scene representations, 2024

    2024

  12. [12]

    Ram: Retrieval-based affordance transfer for generalizable zero-shot robotic manipulation, 2024

    Yuxuan Kuang, Junjie Ye, Haoran Geng, Jiageng Mao, Congyue Deng, Leonidas Guibas, He Wang, and Yue Wang. Ram: Retrieval-based affordance transfer for generalizable zero-shot robotic manipulation, 2024

    work page 2024

  13. [13]

    Any6d: Model-free 6d pose estimation of novel objects

    Taeyeop Lee, Bowen Wen, Minjun Kang, Gyuree Kang, In So Kweon, and Kuk-Jin Yoon. Any6d: Model-free 6d pose estimation of novel objects. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11633–11643, 2025

    work page 2025

  14. [14]

    Learning precise affordances from egocentric videos for robotic manipulation, 2025

    Gen Li, Nikolaos Tsagkas, Jifei Song, Ruaridh Mon-Williams, Sethu Vijayakumar, Kun Shao, and Laura Sevilla-Lara. Learning precise affordances from egocentric videos for robotic manipulation, 2025

    work page 2025

  15. [15]

    Information-driven affordance discovery for efficient robotic manipulation, 2024

    Pietro Mazzaglia, Taco Cohen, and Daniel Dijkman. Information-driven affordance discovery for efficient robotic manipulation, 2024

    work page 2024

  16. [16]

    Where2act: From pixels to actions for articulated 3d objects, 2021

    Kaichun Mo, Leonidas Guibas, Mustafa Mukadam, Abhinav Gupta, and Shubham Tulsiani. Where2act: From pixels to actions for articulated 3d objects, 2021

    work page 2021

  17. [17]

    Rt-affordance: Affordances are versatile intermediate representations for robot manipulation, 2024

    SoroushNasiriany, SeanKirmani, TianliDing, LauraSmith, YukeZhu, DannyDriess, DorsaSadigh, and Ted Xiao. Rt-affordance: Affordances are versatile intermediate representations for robot manipulation, 2024

    work page 2024

  18. [18]

    Language-conditioned affordance-pose detection in 3d point clouds, 2023

    Toan Nguyen, Minh Nhat Vu, Baoru Huang, Tuan Van Vo, Vy Truong, Ngan Le, Thieu Vo, Bac Le, and Anh Nguyen. Language-conditioned affordance-pose detection in 3d point clouds, 2023

    work page 2023

  19. [19]

    Open-vocabularyaffordancedetectionin3dpointclouds

    Toan Nguyen, Minh Nhat Vu, An Vuong, Dzung Nguyen, Thieu Vo, Ngan Le, and Anh Nguyen. Open-vocabularyaffordancedetectionin3dpointclouds. In2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 5692–5698. IEEE, 2023

    work page 2023

  20. [20]

    Where2explore: Few-shot affordancelearningforunseennovelcategoriesofarticulatedobjects.Advances in Neural Information Processing Systems, 36:4585–4596, 2023

    Chuanruo Ning, Ruihai Wu, Haoran Lu, Kaichun Mo, and Hao Dong. Where2explore: Few-shot affordancelearningforunseennovelcategoriesofarticulatedobjects.Advances in Neural Information Processing Systems, 36:4585–4596, 2023

    work page 2023

  21. [21]

    3d flow diffusion policy: Visuomotor policy learning via generating flow in 3d space, 2025

    Sangjun Noh, Dongwoo Nam, Kangmin Kim, Geonhyup Lee, Yeonguk Yu, Raeyoung Kang, and Kyoobin Lee. 3d flow diffusion policy: Visuomotor policy learning via generating flow in 3d space, 2025

    work page 2025

  22. [22]

    Multilayer perceptron and neural networks.WSEAS transactions on circuits and systems, 8(7):579– 588, 2009

    Marius-Constantin Popescu, Valentina E Balas, Liliana Perescu-Popescu, and Nikos Mastorakis. Multilayer perceptron and neural networks.WSEAS transactions on circuits and systems, 8(7):579– 588, 2009

    work page 2009

  23. [23]

    Learning transferable visual models from natural language supervision

    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 supervision. InInternational conference on machine learning, pages 8748–8763. PmLR, 2021

    2021

  24. [24]

    Learning from 10 demos: Generalisable and sample-efficient policy learning with oriented affordance frames, 2025

    KrishanRana,JadAbou-Chakra,SouravGarg,RobertLee,IanReid,andNikoSuenderhauf. Learning from 10 demos: Generalisable and sample-efficient policy learning with oriented affordance frames, 2025. 13

    work page 2025

  25. [25]

    Grounded sam: Assembling open-world models for diverse visual tasks.arXiv preprint arXiv:2401.14159, 2024

    Tianhe Ren, Shilong Liu, Ailing Zeng, Jing Lin, Kunchang Li, He Cao, Jiayu Chen, Xinyu Huang, Yukang Chen, Feng Yan, et al. Grounded sam: Assembling open-world models for diverse visual tasks.arXiv preprint arXiv:2401.14159, 2024

    Pith/arXiv arXiv 2024

  26. [26]

    Goal-conditioned imitation learning using score-based diffusion policies, 2023

    Moritz Reuss, Maximilian Li, Xiaogang Jia, and Rudolf Lioutikov. Goal-conditioned imitation learning using score-based diffusion policies, 2023

    work page 2023

  27. [27]

    Dvg-wm: Disentangled video generation enables efficient embodied world model for robotic manipulation, 2026

    Ziyu Shan, Zhenyu Wu, Xiaofeng Wang, Zheng Zhu, and Ziwei Wang. Dvg-wm: Disentangled video generation enables efficient embodied world model for robotic manipulation, 2026

    work page 2026

  28. [28]

    Dockanywhere: Data-efficientvisuomotorpolicylearningformobilemanipulationvianoveldemonstrationgeneration, 2026

    Ziyu Shan, Yuheng Zhou, Gaoyuan Wu, Ziheng Ji, Zhenyu Wu, and Ziwei Wang. Dockanywhere: Data-efficientvisuomotorpolicylearningformobilemanipulationvianoveldemonstrationgeneration, 2026

    work page 2026

  29. [29]

    Denoising diffusion implicit models.arXiv preprint arXiv:2010.02502, 2020

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

    Pith/arXiv arXiv 2010

  30. [30]

    Prism-dp: Spatial pose-based observations for diffusion-policies via segmentation, mesh generation, and pose tracking.arXiv preprint arXiv:2504.20359, 2025

    Xiatao Sun, Yinxing Chen, and Daniel Rakita. Prism-dp: Spatial pose-based observations for diffusion-policies via segmentation, mesh generation, and pose tracking.arXiv preprint arXiv:2504.20359, 2025

    work page arXiv 2025

  31. [31]

    Maniskill3: Gpu parallelized robotics simulation and rendering for generalizable embodied ai.arXiv preprint arXiv:2410.00425, 2024

    Stone Tao, Fanbo Xiang, Arth Shukla, Yuzhe Qin, Xander Hinrichsen, Xiaodi Yuan, Chen Bao, Xinsong Lin, Yulin Liu, Tse-kai Chan, et al. Maniskill3: Gpu parallelized robotics simulation and rendering for generalizable embodied ai.arXiv preprint arXiv:2410.00425, 2024

    Pith/arXiv arXiv 2024

  32. [32]

    O3afford: One-shot 3d object-to-object affordance grounding for generalizable robotic manipulation, 2025

    Tongxuan Tian, Xuhui Kang, and Yen-Ling Kuo. O3afford: One-shot 3d object-to-object affordance grounding for generalizable robotic manipulation, 2025

    work page 2025

  33. [33]

    Attention is all you need.Advances in neural information processing systems, 30, 2017

    AshishVaswani,NoamShazeer,NikiParmar,JakobUszkoreit,LlionJones,AidanNGomez,Łukasz Kaiser, and Illia Polosukhin. Attention is all you need.Advances in neural information processing systems, 30, 2017

    work page 2017

  34. [34]

    Gendp: 3d semantic fields for category-level generalizable diffusion policy

    YixuanWang,GuangYin,BinghaoHuang,TarikKelestemur,JiuguangWang,andYunzhuLi. Gendp: 3d semantic fields for category-level generalizable diffusion policy. InCoRL, pages 4866–4878, 2024

    work page 2024

  35. [35]

    You only demonstrate once: Category-level manipulation from single visual demonstration.arXiv preprint arXiv:2201.12716, 2022

    Bowen Wen, Wenzhao Lian, Kostas Bekris, and Stefan Schaal. You only demonstrate once: Category-level manipulation from single visual demonstration.arXiv preprint arXiv:2201.12716, 2022

    Pith/arXiv arXiv 2022

  36. [36]

    Foundationpose: Unified 6d pose estimation and tracking of novel objects

    Bowen Wen, Wei Yang, Jan Kautz, and Stan Birchfield. Foundationpose: Unified 6d pose estimation and tracking of novel objects. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 17868–17879, 2024

    2024

  37. [37]

    Ragnet: Large-scalereasoning-basedaffordance segmentation benchmark towards general grasping

    Dongming Wu, Yanping Fu, Saike Huang, Yingfei Liu, Fan Jia, Nian Liu, Feng Dai, Tiancai Wang, RaoMuhammadAnwer,FahadShahbazKhan,etal. Ragnet: Large-scalereasoning-basedaffordance segmentation benchmark towards general grasping. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 11980–11990, 2025

    work page 2025

  38. [38]

    Learning environment-aware affordance for 3d articulated object manipulation under occlusions.Advances in Neural Information Processing Systems, 36:60966–60983, 2023

    Ruihai Wu, Kai Cheng, Yan Zhao, Chuanruo Ning, Guanqi Zhan, and Hao Dong. Learning environment-aware affordance for 3d articulated object manipulation under occlusions.Advances in Neural Information Processing Systems, 36:60966–60983, 2023

    work page 2023

  39. [39]

    Vat-mart: Learning visual action trajectory proposals for manipulating 3d articulated objects, 2022

    Ruihai Wu, Yan Zhao, Kaichun Mo, Zizheng Guo, Yian Wang, Tianhao Wu, Qingnan Fan, Xuelin Chen, Leonidas Guibas, and Hao Dong. Vat-mart: Learning visual action trajectory proposals for manipulating 3d articulated objects, 2022. 14

    work page 2022

  40. [40]

    Afforddp: Generalizable diffusion policy with transferable affordance

    Shijie Wu, Yihang Zhu, Yunao Huang, Kaizhen Zhu, Jiayuan Gu, Jingyi Yu, Ye Shi, and Jingya Wang. Afforddp: Generalizable diffusion policy with transferable affordance. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 6971–6980, 2025

    2025

  41. [41]

    Slotdiffusion: Object-centric generative modeling with diffusion models.Advances in Neural Information Processing Systems, 36:50932–50958, 2023

    ZiyiWu,JingyuHu,WuyueLu,IgorGilitschenski,andAnimeshGarg. Slotdiffusion: Object-centric generative modeling with diffusion models.Advances in Neural Information Processing Systems, 36:50932–50958, 2023

    work page 2023

  42. [42]

    A0: An affordance-aware hierarchical model for general robotic manipulation, 2026

    Rongtao Xu, Jian Zhang, Minghao Guo, Youpeng Wen, Haoting Yang, Min Lin, Jianzheng Huang, Zhe Li, Kaidong Zhang, Liqiong Wang, Yuxuan Kuang, MengCao, Feng Zheng, and Xiaodan Liang. A0: An affordance-aware hierarchical model for general robotic manipulation, 2026

    work page 2026

  43. [43]

    Equibot: Sim (3)-equivariant diffusion policy for generalizable and data efficient learning.arXiv preprint arXiv:2407.01479, 2024

    Jingyun Yang, Zi-ang Cao, Congyue Deng, Rika Antonova, Shuran Song, and Jeannette Bohg. Equibot: Sim (3)-equivariant diffusion policy for generalizable and data efficient learning.arXiv preprint arXiv:2407.01479, 2024

    Pith/arXiv arXiv 2024

  44. [44]

    Seqafford: Sequential 3d affordance reasoning via multimodal large language model, 2025

    Chunlin Yu, Hanqing Wang, Ye Shi, Haoyang Luo, Sibei Yang, Jingyi Yu, and Jingya Wang. Seqafford: Sequential 3d affordance reasoning via multimodal large language model, 2025

    work page 2025

  45. [45]

    General flow as foundation affordance for scalable robot learning.arXiv preprint arXiv:2401.11439, 2024

    Chengbo Yuan, Chuan Wen, Tong Zhang, and Yang Gao. General flow as foundation affordance for scalable robot learning.arXiv preprint arXiv:2401.11439, 2024

    Pith/arXiv arXiv 2024

  46. [46]

    3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations.arXiv preprint arXiv:2403.03954, 2024

    Yanjie Ze, Gu Zhang, Kangning Zhang, Chenyuan Hu, Muhan Wang, and Huazhe Xu. 3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations.arXiv preprint arXiv:2403.03954, 2024

    Pith/arXiv arXiv 2024

  47. [47]

    Pocketdp3: Efficient pocket-scale 3d visuomotor policy, 2026

    Jinhao Zhang, Zhexuan Zhou, Huizhe Li, Yichen Lai, Wenlong Xia, Haoming Song, Youmin Gong, and Jie Mei. Pocketdp3: Efficient pocket-scale 3d visuomotor policy, 2026

    work page 2026

  48. [48]

    Anchordp3: 3d affordance guided sparse diffusion policy for robotic manipulation.arXiv preprint arXiv:2506.19269, 2025

    Ziyan Zhao, Ke Fan, He-Yang Xu, Ning Qiao, Bo Peng, Wenlong Gao, Dongjiang Li, and Hui Shen. Anchordp3: 3d affordance guided sparse diffusion policy for robotic manipulation.arXiv preprint arXiv:2506.19269, 2025

    Pith/arXiv arXiv 2025

  49. [49]

    Learning generalizable manipulation policies with object-centric 3d representations.arXiv preprint arXiv:2310.14386, 2023

    Yifeng Zhu, Zhenyu Jiang, Peter Stone, and Yuke Zhu. Learning generalizable manipulation policies with object-centric 3d representations.arXiv preprint arXiv:2310.14386, 2023. 15

    Pith/arXiv arXiv 2023