Pith. sign in

REVIEW 4 major objections 5 minor 30 references

SViP: Sequencing Bimanual Visuomotor Policies with Object-Centric Motion Primitives

T0 review · 4 major / 5 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 SViP shows that 20 demonstrations plus task-and-motion planning can replace hundreds of demos for bimanual manipulation.

desk verdict A genuinely useful hybrid system that shows learned visuomotor policies can be stitched into TAMP with point-cloud-based switching, but the feasibility validator is a heuristic and the evaluation has holes. read the letter →

arxiv 2506.18825 v1 pith:UZQZP34C submitted 2025-06-23 cs.RO

Yizhou Chen , Hang Xu , Dongjie Yu , Zeqing Zhang , Yi Ren , Jia Pan This is my paper · ORCID
classification cs.RO
keywords bimanualmanipulationimitationlearningtaskandmotionplanningvisuomotorpolicydiffusionmodelscenegraphout-of-distributiongeneralizationprimitives
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

Visuomotor policies trained by imitation tend to fail when the scene shifts a few centimeters from the training distribution, and their errors accumulate over long horizons. SViP argues that the fix is not more data but a division of labor: let a learned policy own only the contact-rich bimanual moments, and hand everything else—picking, placing, transiting, transferring—to parameterized scripted motion primitives that a task-and-motion planner can re-sequence at run time. The paper claims that with only 20 real demonstrations, this hybrid achieves 100 percent success on in-distribution and XY out-of-distribution bimanual insertion in simulation, 88 percent when object headings also shift, and consistently beats a diffusion-policy baseline across three real-world tasks. It also reports that the planner can discover longer sequences—such as relocating an obstacle or an unreachable object—to satisfy reachability and collision constraints, and can compose several learned skills to reach novel goals like clearing the table into a bin.

What carries the argument

The scene graph is the load-bearing abstraction: it turns a raw demonstration into a symbolic description in which bimanual contact-rich segments are black-box skills with preconditions and effects, and unimanual segments are planner-controllable primitives. The switching-condition generator is a denoising diffusion model that learns the distribution of decision variables—starting and ending joint angles $q$ and object-centric trajectories $\tau_o$—conditioned on point clouds $C_o$, with the score function factorized as $\hat{\varepsilon}(V|C_{o_1},...,C_{o_M}) = \varepsilon_\theta(q,k) + \sum_i \varepsilon_\theta(\tau_{o_i},k|C_{o_i})$; SE(3)-equivariance is achieved through a Vector Neuron architecture and an equivariant point-cloud encoder, letting the same generator work for objects at arbitrary poses without pose estimators. The feasibility validator is an MLP that predicts collision probability from the initiation configuration of a bimanual skill, which lets the planner enforce the constraint $\mathrm{SafeBiOp}(a,h_l,h_r,q_l,q_r)$ without simulating the reactive policy's future trajectory. The planning integration uses PDDLStream-style stream functions and an adaptive sampling-based planner to interleave learned and scripted actions.

What would settle it

Collect a test set of cup-sleeve insertion rollouts where the overhead camera is moved to a new viewpoint or a second object partially occludes the sleeve, then run SViP; if success does not fall sharply relative to the 75 percent out-of-distribution result, then point-cloud distribution mismatch is not the limiting factor. Alternatively, log the feasibility validator's predicted collision probability for many random starting configurations and compare with actual collisions measured during execution; any large divergence would show that initiation-configuration-only prediction cannot certify SafeBiOp.

Watch

Extended reading notes

Core claim

SViP's central claim is that a learned visuomotor policy should be treated as one black-box subroutine inside a larger planning problem rather than as the whole controller. The paper segments human demonstrations into an event-driven scene graph whose nodes are grippers, objects, and table regions and whose edges carry continuous variables such as joint configurations and grip poses. Contact-rich bimanual segments become skills defined by pre- and post-scene graphs; the surrounding unimanual segments become object-centric motion primitives. A switching-condition generator, trained as a diffusion model with an SE(3)-equivariant point-cloud encoder, predicts the joint configurations and object-relative trajectories that let these primitives start and end exactly where the learned policy needs them. A feasibility validator predicts collision risk for a bimanual skill from its initiation configuration, and both are wrapped into a task-and-motion planner that computes a constraint-satisfying sequence. The result, the paper argues, is a data-efficient system that generalizes across out-of-distribution initial conditions, obeys reachability and safety constraints, and composes skills for goals not seen in training.

Load-bearing premise

The load-bearing premise is that deployment point clouds from one overhead depth camera look like those in the 20 demonstrations, so the switching-condition generator and feasibility validator trained on those clouds keep working; the paper's own tilted cup-sleeve result (40 percent) shows what happens when that premise weakens.

Editorial extensions

If this is right

  • With only 20 demonstrations per task, SViP reports 100 percent success on in-distribution and XY-shifted bimanual peg insertion, and 88 percent when object heading also shifts; the compared end-to-end imitation policies drop to 2 to 42 percent in the same shifted setups.
  • In real-world object handoff, screwdriver packing, and cup-sleeve insertion, SViP reaches 65 to 100 percent success across in-distribution, out-of-distribution, and tilted setups, while the diffusion-policy baseline ranges from 0 to 90 percent.
  • Because the planner can insert extra scripted primitives, SViP can satisfy reachability and collision constraints—for example, regrasping an out-of-reach socket or moving a pole out of the way—where end-to-end policies fail outright.
  • Novel goals such as table-to-bin clearance and a 19-step multiple-instruction task are achievable by re-sequencing learned bimanual skills with primitives, without collecting new demonstrations for the new goal.
  • The method does not require object pose estimators, which matters for symmetric objects such as pegs, cups, and sleeves where pose estimation is unstable.

Reading between the lines

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

  • The same scene-graph abstraction could likely be applied to unimanual contact-rich skills such as screwing or wiping, since the formulation does not hinge on having two arms; a natural extension is to replace some scripted primitives with a library of learned unimanual skills and let the planner choose among them.
  • If the switching-condition generator is the bottleneck, then improving point-cloud completeness—for example, fusing multiple depth views or online reconstruction, which the authors name as future work—should raise the tilted cup-sleeve success from 40 percent toward the other tasks' levels; that is a testable prediction.
  • The framework's wall-clock limits are set by the planner's exponential growth in object count; for scenes with many objects, the paper's 7 to 27 second planning times would grow, so a hierarchical or learned proposal for plan skeletons would be a natural next step.
  • Because the feasibility validator reasons only from the initiation configuration, its predictions may fail when a learned policy reacts differently to visual perturbations mid-execution; testing this would require comparing validator predictions against measured collisions in out-of-distribution rollouts.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes SViP, a hybrid imitation-learning/TAMP system for bimanual manipulation. Human demonstrations are segmented into scene graphs; contact-rich segments are treated as black-box bimanual skills executed by learned visuomotor policies, while pick/place/transit operations are handled by parameterized object-centric motion primitives. A DDPM-based switching condition generator predicts start/end configurations and object-centric trajectories, and an MLP feasibility validator is used to enforce a SafeBiOp collision-avoidance constraint. The system is evaluated in an ALOHA-style simulated peg-in-hole task and in three real-world tasks using 20 demonstrations per task, with comparisons against ACT and Diffusion Policy, plus qualitative novel-goal demonstrations.

Significance. If the claims hold, SViP makes a useful contribution to data-efficient bimanual manipulation by combining the generalization of TAMP with the contact-rich robustness of learned visuomotor policies, without requiring 6-DoF object pose estimators. The data efficiency (20 real demonstrations), use of SE(3)-equivariant encoders, and automatic extraction of symbolic descriptions from scene graphs are genuine strengths. The main significance risk is that the claimed 'certified' collision safety rests on an unvalidated structural assumption about the feasibility validator, and the quantitative evidence for novel-goal composition is thin.

major comments (4)
  1. [Section III-C] The SafeBiOp feasibility validator is trained on triples (q, p, ν) with δ as the closest distance of the robot to the obstacle during an unrolled execution. The paper does not establish that the collision outcome of a closed-loop visuomotor policy is determined by the initiation configuration and a static obstacle descriptor. Since the policy is conditioned on point-cloud observations at execution time, trajectories and collision outcomes can differ across scenes that share the same (q, p, ν) whenever the manipulated-object pose, grasp, or partial occlusion differs. Table II shows only 70% success in the unsafe setup, and the text attributes failures to inaccurate point clouds rather than to the validator's missing variables. The predicate SafeBiOp is therefore not 'certified' as stated; at best it encodes a learned heuristic. I recommend including object-state and point-cloud features in the validation dataset, validating the predictor against ground-truth collision labels across repeated rollouts, and either relaxing the 'certified' language or providing quantitative evidence that the predictor is conservative.
  2. [Table II and Section IV-A3] The unreachable and unsafe experiments are the central evidence for constraint handling (Q2), but they are based on only 20 trials per setup with no error bars, no confidence intervals, and no decomposition of failures into planning failures versus execution failures. The implementation of IsReachable is also not described ('scripted using distance-related metrics' is vague), which makes it difficult to interpret the 55% success rate in the unreachable setup. Please report per-trial outcomes, a failure-mode breakdown, and more trials or bootstrap intervals so the reader can assess the significance of the reported rates.
  3. [Section IV-B3] The novel-goal experiments (Table-to-Bin Clearance, Cup-sleeve Insertion and Reconfiguration, and Multiple Instructions) are reported only through illustrative screenshots; no success rates, completion statistics, or quantitative goal-satisfaction metrics are provided. Since Q3 ('Can SViP complete tasks with novel goals?') is a headline claim, the current evidence is insufficient. Please add quantitative results for these tasks, including number of rollouts, success criteria, and where possible a comparison against a non-compositional baseline.
  4. [Section IV-A and Section IV-B] Reproducibility is hampered by missing details. The simulation environment is deferred to the ALOHA repository, and the switching condition generator is described only at a high level without network architecture, DDPM hyperparameters, training loss, or the exact point-cloud preprocessing used in the real-world experiments. The real-world comparison also uses only Diffusion Policy as a baseline, while ACT results are reported in simulation but not on real hardware. Reporting these details would substantially strengthen confidence in the 100%/88% simulation claims and the real-world comparisons.
minor comments (5)
  1. [Abstract and Section I] There are typos: 'Seqeuncing' in the abstract, 'Furthurmore' in the introduction, and inconsistent spelling of 'bimanual' as 'bimannual' in Section III-A; please proofread the final version.
  2. [Section IV-B2] Figure 13 is captioned 'The leaky scenarios in three real-world experiments,' but the term 'leaky' is never defined or discussed in the text; please either define it or rename the figure.
  3. [Section IV-B and Section E (including Figs. 9-13)] The manuscript contains apparent layout artifacts: figure numbers jump from Fig. 5 to Figs. 9-13, and duplicate text blocks appear in the real-world experiment section and in Section E. These should be cleaned up before publication.
  4. [Table III] The real-world success rates in Table III are reported without confidence intervals or per-trial variance over 20 rollouts; for small differences (e.g., 80% vs 75%) this makes it hard to assess whether the superiority is meaningful. Please add confidence intervals or raw trial data.
  5. [Section III-B] Equation (2) introduces the notation \hat{\epsilon}(V|C_{o_1},...,C_{o_M}) without explicitly defining how it relates to the standard noise-prediction objective of the DDPM; please clarify the notation and the training loss used for the switching condition generator.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the fitted switching-condition and feasibility models are planning subroutines, not renamed predictions of the reported success metrics.

full rationale

The derivation chain is: demonstration segmentation into scene graphs; training of a contact-rich visuomotor policy pi; training of switching-condition generators phi and feasibility validator xi on the same demonstrations; and a PDDLStream/TAMP search that samples from phi and filters with xi to compose object-centric motion primitives with pi. The paper's headline claims are end-to-end success rates under ID/OOD and novel-goal conditions, measured by task completion rollouts against ACT/DP baselines. The learned phi and xi are planning subroutines; their outputs are not the evaluation metric, and the TAMP search over PDDL constraints and standard motion primitives is not forced by any fitted parameter. The feasibility validator is a learned surrogate for collision likelihood, and the paper's own Table II and conclusion acknowledge its empirical limits (e.g., 40% Tilted cup-sleeve success attributed to noisy single-view point clouds); this is a correctness/robustness concern, not an equivalence-by-construction between input and output. The only apparent self-citation ([25], BiKC) is used for the standard Kabsch algorithm in point-cloud registration and is not load-bearing for the main result. No equation in the paper reduces a claimed prediction to the training labels or to a self-citation chain, so no circularity is established.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

No new physical entities are posited. The free parameters are mostly unspecified learning hyperparameters and thresholds. The main burdens are the assumptions that learned switching conditions and feasibility predictions transfer to test scenes.

free parameters (2)
  • SafeBiOp collision threshold = Not stated
    The feasibility validator predicts collision probability from initiation configurations; the threshold that certifies SafeBiOp is not specified in the paper.
  • Diffusion policy and switching generator hyperparameters = Not stated
    Network sizes, diffusion steps, and training details are not reported, so the success rates implicitly depend on unspecified tuning choices.
assumptions (3)
  • domain assumption The learned switching condition generator successfully transitions when the scene is close enough to the training distribution.
    The entire interleaving relies on this learned classifier to decide when to start the visuomotor policy; the paper does not report its accuracy, only end-to-end success rates.
  • ad hoc to paper Collision risk during an unrolled bimanual reactive policy can be predicted from the initiation configuration alone.
    The SafeBiOp feasibility validator ignores the closed-loop trajectory of the learned policy, which is a strong modeling assumption for a reactive policy.
  • domain assumption Single-view depth point clouds are sufficient to represent the scene for switching and feasibility.
    The authors state in the conclusion that single-view depth limits performance, and the Tilted cup-sleeve result is consistent with a broken assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SViP: Sequencing Bimanual Visuomotor Policies with Object-Centric Motion Primitives." pith.science (2026). https://pith.science/paper/UZQZP34C

@misc{pith2026250618825,
  author       = {Pith},
  title        = {Pith review of: SViP: Sequencing Bimanual Visuomotor Policies with Object-Centric Motion Primitives},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UZQZP34C}},
  note         = {Machine review of arXiv:2506.18825}
}
read the original abstract

Imitation learning (IL), particularly when leveraging high-dimensional visual inputs for policy training, has proven intuitive and effective in complex bimanual manipulation tasks. Nonetheless, the generalization capability of visuomotor policies remains limited, especially when small demonstration datasets are available. Accumulated errors in visuomotor policies significantly hinder their ability to complete long-horizon tasks. To address these limitations, we propose SViP, a framework that seamlessly integrates visuomotor policies into task and motion planning (TAMP). SViP partitions human demonstrations into bimanual and unimanual operations using a semantic scene graph monitor. Continuous decision variables from the key scene graph are employed to train a switching condition generator. This generator produces parameterized scripted primitives that ensure reliable performance even when encountering out-of-the-distribution observations. Using only 20 real-world demonstrations, we show that SViP enables visuomotor policies to generalize across out-of-distribution initial conditions without requiring object pose estimators. For previously unseen tasks, SViP automatically discovers effective solutions to achieve the goal, leveraging constraint modeling in TAMP formulism. In real-world experiments, SViP outperforms state-of-the-art generative IL methods, indicating wider applicability for more complex tasks. Project website: https://sites.google.com/view/svip-bimanual

Figures

Figures reproduced from arXiv: 2506.18825 by the authors.

Figure 1
Figure 1. Provided with motion planning primitives and trained bimanual [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (a) and (c) show the start and the end of the contact-rich part, while (b) shows the contact-rich part of the Object Handoff operation. As marked in the overhead camera images, pH and pO are the center points of robot grippers and objects. The corresponding scene graphs Gpre, Gmid, Geff are sketched below, with continuous variables labeled on each edge. Here, ρ, hl , hr, o1 denote the table, the left robot, the righ… view at source ↗
Figure 3
Figure 3. Execution by SViP in the unreachable and the unsafe setup. TABLE II RESULTS OF SVIP’S ZERO-SHOT TRANSFER TO TASKS THAT INVOLVE OPERATIONS UNSEEN IN THE DEMONSTRATION. Setup Sequence Len. Success Rate Computation Time (s) ID (original) 7.0 100 7.55 unreachable 11.0 55 12.36 unsafe 11.6 70 27.39 operations. To this end, we design two distinct simulation settings, named unreachable and unsafe, by updating the simulatio… view at source ↗
Figures from the paper (2 more)
Figure 9
Figure 9. Figure 9: The demonstrated action sequence of the object handoff task. [PITH_FULL_IMAGE:figures/full_fig_p005_9.png]
Figure 5
Figure 5. Figure 5: The screenshots of rollouts in (a) Table-to-Bin Clearance task and (b) Multiple Instructions task. Insertion (OOD) scenario. • Multiple Instructions: The goal is given as η = T ai∈{a0,a1} DoneBiOp(ai) ∧ T oi∈O On(oi , table), where a0 and a1 are the screwdriver-packing…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 17 canonical work pages

  1. [1]

    arXiv preprint arXiv:2410.24164 , 2024

    Kevin Black, Noah Brown, Danny Driess, Adnan Es- mail, Michael Equi, Chelsea Finn, Niccolo Fusai, Lachy Groom, Karol Hausman, Brian Ichter, et al.π0: A vision- language-action flow model for general robot control. arXiv preprint arXiv:2410.24164 , 2024

    arXiv

  2. [2]

    Combined task and motion planning for a dual-arm robot to use a suction cup tool

    Hao Chen, Weiwei Wan, and Kensuke Harada. Combined task and motion planning for a dual-arm robot to use a suction cup tool. In 2019 IEEE-RAS 19th International Conference on Humanoid Robots (Humanoids) , pages 446–452. IEEE, 2019

    work page 2019

  3. [3]

    League: Guided skill learning and abstraction for long-horizon manipulation

    Shuo Cheng and Danfei Xu. League: Guided skill learning and abstraction for long-horizon manipulation. IEEE Robotics and Automation Letters , 2023

    work page 2023

  4. [4]

    NOD-TAMP: Multi-step manipulation planning with neural object descriptors

    Shuo Cheng, Caelan Reed Garrett, Ajay Mandlekar, and Danfei Xu. NOD-TAMP: Multi-step manipulation planning with neural object descriptors. In CoRL 2023 Workshop on Learning Effective Abstractions for Plan- ning (LEAP), 2023

    work page 2023

  5. [5]

    Diffu- sion Policy: Visuomotor Policy Learning via Action Dif- fusion

    Cheng Chi, Siyuan Feng, Yilun Du, Zhenjia Xu, Eric Cousineau, Benjamin Burchfiel, and Shuran Song. Diffu- sion Policy: Visuomotor Policy Learning via Action Dif- fusion. In Robotics: Science and Systems XIX . Robotics: Science and Systems Foundation, July 2023. ISBN 978- 0-9923747-9-2. doi: 10.15607/RSS.2023.XIX.026

    doi:10.15607/rss.2023.xix.026 2023

  6. [6]

    Plan-seq-learn: Language model guided rl for solving long horizon robotics tasks

    Murtaza Dalal, Tarun Chiruvolu, Devendra Chaplot, and Ruslan Salakhutdinov. Plan-seq-learn: Language model guided rl for solving long horizon robotics tasks. arXiv preprint arXiv:2405.01534, 2024

    arXiv 2024

  7. [7]

    Vec- tor Neurons: A general framework for SO(3)-equivariant networks

    Congyue Deng, Or Litany, Yueqi Duan, Adrien Poule- nard, Andrea Tagliasacchi, and Leonidas J Guibas. Vec- tor Neurons: A general framework for SO(3)-equivariant networks. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 12200–12209, 2021

    work page 2021

  8. [8]

    POPE: 6-DoF prompt- able pose estimation of any object in any scene with one reference

    Zhiwen Fan, Panwang Pan, Peihao Wang, Yifan Jiang, Dejia Xu, and Zhangyang Wang. POPE: 6-DoF prompt- able pose estimation of any object in any scene with one reference. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 7771–7781, 2024

    work page 2024

Show all 30 references
  1. [9]

    Skillmimicgen: Automated demonstration genera- tion for efficient skill learning and deployment

    Caelan Garrett, Ajay Mandlekar, Bowen Wen, and Dieter Fox. Skillmimicgen: Automated demonstration genera- tion for efficient skill learning and deployment. arXiv preprint arXiv:2410.18907, 2024

    2024 arXiv

  2. [10]

    Sampling-based methods for factored task and motion planning

    Caelan Reed Garrett, Tom ´as Lozano-P ´erez, and Leslie Pack Kaelbling. Sampling-based methods for factored task and motion planning. The International Journal of Robotics Research , 37(13-14):1796–1825, 2018

    2018

  3. [11]

    PDDLStream: Integrating symbolic planners and blackbox samplers via optimistic adaptive planning

    Caelan Reed Garrett, Tom ´as Lozano-P ´erez, and Leslie Pack Kaelbling. PDDLStream: Integrating symbolic planners and blackbox samplers via optimistic adaptive planning. In Proceedings of the international conference on automated planning and scheduling , volume 30, pages 440–...

    2020

  4. [12]

    Stem-OB: Generalizable visual imitation learning with stem-like convergent observation through diffusion inversion

    Kaizhe Hu, Zihang Rui, Yao He, Yuyao Liu, Pu Hua, and Huazhe Xu. Stem-OB: Generalizable visual imitation learning with stem-like convergent observation through diffusion inversion. arXiv preprint arXiv:2411.04919 , 2024

    2024 arXiv

  5. [13]

    3D Diffuser Actor: Policy diffusion with 3d scene representations, March 2024

    Tsung-Wei Ke, Nikolaos Gkanatsios, and Katerina Fragkiadaki. 3D Diffuser Actor: Policy diffusion with 3d scene representations, March 2024

    2024

  6. [14]

    A Bimanual Ma- nipulation Taxonomy

    Franziska Krebs and Tamim Asfour. A Bimanual Ma- nipulation Taxonomy. IEEE Robot. Autom. Lett. , 7(4): 11031–11038, October 2022. ISSN 2377-3766, 2377-

    2022

  7. [15]

    EFEM: Equivariant neural field expectation maximization for 3D object segmenta- tion without scene supervision

    Jiahui Lei, Congyue Deng, Karl Schmeckpeper, Leonidas Guibas, and Kostas Daniilidis. EFEM: Equivariant neural field expectation maximization for 3D object segmenta- tion without scene supervision. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni...

    2023

  8. [16]

    RDT-1B: A diffusion foundation model for bi- manual manipulation

    Songming Liu, Lingxuan Wu, Bangguo Li, Hengkai Tan, Huayu Chen, Zhengyi Wang, Ke Xu, Hang Su, and Jun Zhu. RDT-1B: A diffusion foundation model for bi- manual manipulation. arXiv preprint arXiv:2410.07864 , 2024

    2024 arXiv

  9. [17]

    Gen6D: Generalizable model-free 6-DoF object pose estimation from rgb images

    Yuan Liu, Yilin Wen, Sida Peng, Cheng Lin, Xiaoxiao Long, Taku Komura, and Wenping Wang. Gen6D: Generalizable model-free 6-DoF object pose estimation from rgb images. In European Conference on Computer Vision, pages 298–315. Springer, 2022

    2022

  10. [18]

    Human-in-the-loop task and motion planning for imitation learning

    Ajay Mandlekar, Caelan Reed Garrett, Danfei Xu, and Dieter Fox. Human-in-the-loop task and motion planning for imitation learning. In Conference on Robot Learning, pages 3030–3060. PMLR, 2023

    2023

  11. [19]

    Grounded SAM: Assembling open-world models for diverse visual tasks

    Tianhe Ren, Shilong Liu, Ailing Zeng, Jing Lin, Kun- chang 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

    2024 arXiv

  12. [20]

    Should ebms model the energy or the score? In Energy Based Models Workshop- ICLR 2021, 2021

    Tim Salimans and Jonathan Ho. Should ebms model the energy or the score? In Energy Based Models Workshop- ICLR 2021, 2021

    2021

  13. [21]

    Mu- JoCo: A physics engine for model-based control

    Emanuel Todorov, Tom Erez, and Yuval Tassa. Mu- JoCo: A physics engine for model-based control. In 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems , pages 5026–5033, 2012. doi: 10.1109/IROS.2012.6386109

    2012

  14. [22]

    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. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17868–17879, 2024

    2024

  15. [23]

    GELLO: A general, low-cost, and intuitive teleoperation framework for robot manipulators

    Philipp Wu, Yide Shentu, Zhongke Yi, Xingyu Lin, and Pieter Abbeel. GELLO: A general, low-cost, and intuitive teleoperation framework for robot manipulators. In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages 12156–12163. IEEE, 2024

    2024

  16. [24]

    Equi- bot: SIM(3)-Equivariant diffusion policy for general- izable and data efficient learning

    Jingyun Yang, Zi-ang Cao, Congyue Deng, Rika Antonova, Shuran Song, and Jeannette Bohg. Equi- bot: SIM(3)-Equivariant diffusion policy for general- izable and data efficient learning. arXiv preprint arXiv:2407.01479, 2024

    2024 arXiv

  17. [25]

    BiKC: Keypose-conditioned consistency pol- icy for bimanual robotic manipulation

    Dongjie Yu, Hang Xu, Yizhou Chen, Yi Ren, and Jia Pan. BiKC: Keypose-conditioned consistency pol- icy for bimanual robotic manipulation. arXiv preprint arXiv:2406.10093, 2024

    2024 arXiv

  18. [26]

    3D Diffusion Policy: Generalizable visuomotor policy learning via simple 3D representations

    Yanjie Ze, Gu Zhang, Kangning Zhang, Chenyuan Hu, Muhan Wang, and Huazhe Xu. 3D Diffusion Policy: Generalizable visuomotor policy learning via simple 3D representations. In ICRA 2024 Workshop on 3D Visual Representations for Robot Manipulation , 2024

    2024

  19. [27]

    Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware

    Tony Zhao, Vikash Kumar, Sergey Levine, and Chelsea Finn. Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware. In Robotics: Science and Systems XIX. Robotics: Science and Systems Foundation, July 2023. ISBN 978-0-9923747-9-2. doi: 10.15607/ RSS.2023.XIX.016

    2023

  20. [28]

    ALOHA Unleashed: A simple recipe for robot dexterity

    Tony Z Zhao, Jonathan Tompson, Danny Driess, Pete Florence, Kamyar Ghasemipour, Chelsea Finn, and Ayzaan Wahid. ALOHA Unleashed: A simple recipe for robot dexterity. arXiv preprint arXiv:2410.13126 , 2024

    2024 arXiv

  21. [29]

    On the continuity of rotation representations in neural networks

    Yi Zhou, Connelly Barnes, Jingwan Lu, Jimei Yang, and Hao Li. On the continuity of rotation representations in neural networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 5745–5753, 2019

    2019

  22. [3774]

    doi: 10.1109/LRA.2022.3196158

    2022

Pith tools

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