Pith. sign in

REVIEW 2 major objections 1 minor 51 references

WT-UMI: Tactile-based Whole-Body Manipulation via Force-Supervised Contact-Aware Planning

T0 review · 2 major / 1 minor · reviewed 2026-06-27 · grok-4.3

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

Pith's one-line read WT-UMI combines a wearable tactile interface with a force-supervised planner to raise success in contact-rich whole-body humanoid tasks.

desk verdict WT-UMI adds a wearable whole-body tactile interface plus a force-conditioned correction and force-supervised planner that mixes human demos with teleop data for contact-rich humanoid tasks. read the letter →

arxiv 2606.13232 v1 pith:4UFV37AL submitted 2026-06-11 cs.RO

Jaehwi Jang , Zhaoyuan Gu , Alfred Cueva , Zimeng Chai , Junjie Sheng , Thong Nguyen , Himank Galundia , Yifan Wu
show 10 more authors
Huishu Xue Isaac Legene Ojas Mediratta Davin Doan Andrew Collins Sarah Sadegh KyoungMok Kim Rishita Dhalbisoi Zun Chen Ye Zhao
This is my paper
classification cs.RO
keywords whole-bodymanipulationtactilesensingforce-supervisedplanningcontact-awarecontrolhumanoidrobotsadmittancecontrollerimitationlearningwearableinterface
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper introduces WT-UMI as a system that supplies tactile images, contact forces, and poses through a wearable interface usable in both human demonstration and robot teleoperation modes. A force-conditioned module learns to adjust human poses into robot targets from teleoperation recordings. A planner then forecasts sequences of end-effector poses together with contact-force trajectories, and the force output directly references a tactile admittance controller for explicit regulation. Across five tasks involving deformable objects, bulky rigid items, and human-humanoid collaboration, the method raises success rates and lowers contact-position tracking error relative to four imitation baselines. A sympathetic reader would care because conventional policies handle forces only implicitly, which restricts reliable performance when objects must be shared or handled delicately.

What carries the argument

The force-supervised planner that outputs end-effector pose chunks and contact-force trajectories, using the predicted forces as the reference signal for the tactile-based admittance controller, together with the force-conditioned target-pose correction module trained on teleoperation data.

What would settle it

Measure whether success rates on a sixth unseen contact-rich task remain above the four baselines when the correction module receives no further training data from that task.

Watch

Extended reading notes

Core claim

WT-UMI shows that a force-supervised planner predicting both end-effector pose chunks and contact-force trajectories, with the force serving as reference for a tactile admittance controller, combined with a force-conditioned target-pose correction module, produces higher success rates and lower tracking errors than standard policies on contact-rich whole-body tasks.

Load-bearing premise

The force-conditioned target-pose correction module learned from teleoperation data will convert measured human poses into contact-aware robot targets that generalize when the force-supervised planner is applied to new tasks.

Editorial extensions

If this is right

  • Success rates improve over four policy baselines on five contact-rich tasks.
  • Contact-position tracking error decreases.
  • The approach covers deformable objects, bulky rigid objects, and human-humanoid collaboration.
  • Natural force interactions captured in human demonstrations are made usable through explicit force prediction and admittance control.

Reading between the lines

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

  • The chunked prediction structure could support longer task horizons without retraining the full policy.
  • A similar wearable interface might collect training data for non-humanoid platforms that also require distributed contact sensing.
  • Explicit force references could be extended to enforce safety limits during shared-load operations.
Share X Bluesky LinkedIn Reddit HN

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

2 major / 1 minor

Summary. The paper presents WT-UMI, a wearable whole-body tactile interface that supports both human demonstration and humanoid teleoperation modes. It introduces a force-conditioned target-pose correction module learned from teleoperation data to convert human poses into contact-aware robot targets, together with a force-supervised planner that predicts end-effector pose chunks and contact-force trajectories; the predicted forces serve as references for a tactile-based admittance controller. Across five contact-rich tasks involving deformable objects, bulky rigid objects, and human-humanoid collaboration, the method is claimed to improve success rate and reduce contact-position tracking error relative to four policy baselines.

Significance. If the empirical results hold with proper statistical support, the work would offer a concrete way to fuse complementary demonstration modalities while explicitly supervising contact forces, which is a recurring bottleneck in whole-body humanoid manipulation. The combination of a learned correction module with force-supervised planning could serve as a template for other contact-rich domains where pure teleoperation or pure human demonstration is insufficient.

major comments (2)
  1. [Abstract] Abstract: the central empirical claim that WT-UMI 'improves success rate and reduces contact-position tracking error' over four baselines is stated without any numerical values, error bars, dataset sizes, number of trials, or ablation results. This absence directly undermines verification of the load-bearing performance assertion.
  2. [Abstract] The force-conditioned target-pose correction module is described as converting human poses into robot targets by learning from teleoperation data, yet no details are supplied on the training objective, network architecture, loss terms, or regularization that would ensure the module generalizes beyond the training distribution to the five evaluation tasks.
minor comments (1)
  1. [Abstract] The project page URL is given but no supplementary video or dataset link is referenced in the abstract; adding these would aid reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on the abstract. We agree that incorporating quantitative results and additional specifics will strengthen the presentation and address the concerns raised. We will revise the abstract accordingly and provide point-by-point responses below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central empirical claim that WT-UMI 'improves success rate and reduces contact-position tracking error' over four baselines is stated without any numerical values, error bars, dataset sizes, number of trials, or ablation results. This absence directly undermines verification of the load-bearing performance assertion.

    Authors: We agree that the abstract would benefit from quantitative support to make the performance claims more verifiable. In the revised version, we will update the abstract to include the specific success rates, contact-position tracking errors (with standard deviations), number of trials per task, and references to the ablation studies and dataset sizes reported in Section 5. revision: yes

  2. Referee: [Abstract] The force-conditioned target-pose correction module is described as converting human poses into robot targets by learning corrections from teleoperation data, yet no details are supplied on the training objective, network architecture, loss terms, or regularization that would ensure the module generalizes beyond the training distribution to the five evaluation tasks.

    Authors: The training objective, network architecture, loss terms, and regularization for the force-conditioned target-pose correction module are detailed in Section 4.2 of the manuscript. To address the abstract-specific concern, we will add a concise clause noting that the module is trained via supervised learning on teleoperation data with force conditioning to support generalization across the evaluated tasks. revision: yes

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; empirical pipeline with no self-referential derivations

full rationale

The paper describes a modular pipeline (wearable interface, force-conditioned correction module learned from teleoperation data, force-supervised planner predicting pose/force trajectories, admittance controller) evaluated empirically across five tasks against four baselines. No equations, fitted parameters renamed as predictions, self-citations as load-bearing premises, or uniqueness theorems appear in the provided text. Claims rest on measured success rates and tracking errors rather than any derivation that reduces to its own inputs by construction. This is the expected non-finding for a purely empirical robotics systems paper.

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

Abstract mentions no explicit free parameters, axioms, or invented entities; all components are described as learned or supervised from data.

how reviews work

0 comments
Cite this review

Pith. "Pith review of WT-UMI: Tactile-based Whole-Body Manipulation via Force-Supervised Contact-Aware Planning." pith.science (2026). https://pith.science/paper/4UFV37AL

@misc{pith2026260613232,
  author       = {Pith},
  title        = {Pith review of: WT-UMI: Tactile-based Whole-Body Manipulation via Force-Supervised Contact-Aware Planning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4UFV37AL}},
  note         = {Machine review of arXiv:2606.13232}
}
read the original abstract

Whole-body humanoid manipulation of bulky, deformable, and shared-load objects requires distributed contact sensing and explicit force regulation, yet most imitation policies treat contact force only implicitly. On the other hand, different demonstration sources provide complementary modalities with inherent trade-offs: human demonstrations capture natural contact forces but not robot-executable actions, while teleoperation directly records robot actions but with less natural force regulation. This paper presents \textbf{WT-UMI}, a wearable whole-body tactile interface worn by human operators or mounted on humanoids, providing accurate observations of tactile images, contact forces, and end-effector poses across both human demonstration and humanoid teleoperation modes. We introduce a force-conditioned target-pose correction module that converts measured human poses into contact-aware robot targets by learning corrections from teleoperation data. To leverage the natural force interaction in human data, we propose a force-supervised planner that predicts end-effector pose chunks and contact-force trajectories. The predicted contact force serves as the reference for a tactile-based admittance controller. Across five contact-rich tasks spanning deformable objects, bulky rigid objects, and human--humanoid collaboration, WT-UMI improves success rate and reduces contact-position tracking error over four policy baselines. Our project page is available at https://wt-umi.github.io/WTUMI/.

Figures

Figures reproduced from arXiv: 2606.13232 by the authors.

Figure 1
Figure 1. (a) WT-UMI is a shared interface between human demonstrators and humanoid robots for whole-body tactile data collection. (b) A human demonstrator wears WT-UMI, or a humanoid is controlled via teleoperation using the same hardware. (c) A force-supervised planner trained from WT-UMI data executes contact-rich tactile-aware tasks, spanning whole-body manipulation of deformable and large rigid objects and human–humanoid… view at source ↗
Figure 2
Figure 2. A force-conditioned target-pose correction module creates action labels for human data. A force￾supervised planner produces a contact-force trajectory in addition to end-effector poses. The predicted forces are used for online force regulation via a tactile-based admittance controller. wears the chest plate, forearm covers, and GripTacs, with a PICO controller on each GripTac to track the bimanual pose. Human mode r… view at source ↗
Figure 3
Figure 3. WT-UMI includes GripTac end-effectors, forearm covers, and a chest plate, each equipped with a thin-film tactile sensor. 3.2 Force-Conditioned Target-Pose Correction We introduce a force-conditioned target-pose correction module that converts the measured human hand pose T s,m, contact force f s,m, and tactile observation I s,m into a robot target pose, where s ∈ {l, r} indexes the left and right hands and m denotes… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Deployment of our framework on three whole-body manipulation tasks. lag. This improved temporal consistency is further reflected in the lower force-rate RMS, which decreases to 3.74 N/s compared to the 19.80 N/s from the force head trained on teleoperation data. 5.2 Ef…
Figure 5
Figure 5. Figure 5: Force calibration of the SensX 160 palm sensor with and without a gel pad. The bare [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Target-pose correction training. Contact-mode-aligned teleoperation and human segments [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: Human–humanoid collaborative manipulation tasks. Top images show the beam transport [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Admittance controller force-tracking. Force regulation evaluation. We evaluate the closed-loop force regulation of the tactile admittance controller from Sec. 3.4. The robot holds a yoga ball between its chest and palms under a fixed pose target, with both palms ini￾ti…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

51 extracted references · 1 canonical work pages

  1. [1]

    Z. Gu, J. Li, W. Shen, W. Yu, Z. Xie, S. McCrory, X. Cheng, A. Shamsah, R. Griffin, C. K. Liu, A. Kheddar, X. B. Peng, Y . Zhu, G. Shi, Q. Nguyen, G. Cheng, H. Gao, and Y . Zhao. Humanoid locomotion and manipulation: Current progress and challenges in control, planning, and learning.IEEE/ASME Transactions on Mechatronics, 31(2):2300–2330, 2026

    2026

  2. [2]

    C. Chi, Z. Xu, S. Feng, E. Cousineau, Y . Du, B. Burchfiel, R. Tedrake, and S. Song. Diffusion policy: Visuomotor policy learning via action diffusion.The International Journal of Robotics Research, 44(10-11):1684–1704, 2025

    2025

  3. [3]

    Y . Niu, Z. Fang, B. Chen, S. Zhou, R. Senthilkumaran, H. Zhang, B. Chen, C. Qiu, H. E. Tseng, J. Francis, and D. Zhao. Learning versatile humanoid manipulation with touch dreaming, 2026

    2026

  4. [4]

    Y . Hou, Z. Liu, C. Chi, E. Cousineau, N. Kuppuswamy, S. Feng, B. Burchfiel, and S. Song. Adaptive compliance policy: Learning approximate compliance for diffusion guided control. InIEEE International Conference on Robotics and Automation, pages 4829–4836, 2025

    2025

  5. [5]

    Y . Han, K. Yu, R. Batra, N. Boyd, C. Mehta, T. Zhao, Y . She, S. Hutchinson, and Y . Zhao. Learning generalizable vision-tactile robotic grasping strategy for deformable objects via trans- former.IEEE/ASME Transactions on Mechatronics, 30(1):554–566, 2024

    2024

  6. [6]

    Helmut, L

    E. Helmut, L. Dziarski, N. Funk, B. Belousov, and J. Peters. Learning force distribution esti- mation for the gelsight mini optical tactile sensor based on finite element analysis. InIEEE/RSJ International Conference on Intelligent Robots and Systems, pages 8553–8560, 2025

    2025

  7. [7]

    Helmut, N

    E. Helmut, N. Funk, T. Schneider, C. de Farias, and J. Peters. Tactile-conditioned diffusion policy for force-aware robotic manipulation. InIEEE International Conference on Robotics and Automation, 2026

    2026

  8. [8]

    Zhang, H

    K. Zhang, H. Zhang, Z. Xu, Z. Zhang, M. R. I. Prince, X. Li, X. Han, Y . Zhou, A. Ajoudani, and Y . She. Tacvla: Contact-aware tactile fusion for robust vision-language-action manipulation, 2026

    2026

Show all 51 references
  1. [9]

    Mittendorfer, E

    P. Mittendorfer, E. Yoshida, T. Moulard, and G. Cheng. A general tactile approach for grasping unknown objects with a humanoid robot. InIEEE/RSJ International Conference on Intelligent Robots and Systems, pages 4747–4752, 2013

    2013

  2. [10]

    J. A. Barreiros, A. ¨Ozg¨un ¨Onol, M. Zhang, S. Creasey, A. Goncalves, A. Beaulieu, A. Bhat, K. M. Tsui, and A. Alspach. Learning contact-rich whole-body manipulation with example- guided reinforcement learning.Science Robotics, 10(105):eads6790, 2025

    2025

  3. [11]

    Cheng, K

    T. Cheng, K. Chen, L. Chen, L. Zhang, Y . Zhang, Y . Ling, M. Hamad, Z. Bing, F. Wu, K. Sharma, and A. Knoll. TacUMI: A multi-modal universal manipulation interface for contact-rich tasks. InExtended Abstracts of the ACM/IEEE International Conference on Human-Robot Interaction...

    2026

  4. [12]

    Armleder, F

    S. Armleder, F. Bergner, J. R. Guadarrama-Olvera, J. Nakanishi, and G. Cheng. Real-time control of a humanoid robot for whole-body tactile interaction.Advanced Intelligent Systems, 7(12):e202500149, 2025

    2025

  5. [13]

    Murooka, K

    M. Murooka, K. Fukumitsu, M. Hamze, M. Morisawa, H. Kaminaga, F. Kanehiro, and E. Yoshida. Whole-body multi-contact motion control for humanoid robots based on dis- tributed tactile sensors.IEEE Robotics and Automation Letters, 9(11):10620–10627, 2024. 9

    2024

  6. [14]

    Subburaman and O

    R. Subburaman and O. Stasse. A whole-body multi contact large object manipulation and esti- mation framework for humanoids using skin patches. InIEEE-RAS International Conference on Humanoid Robots, pages 1–8, 2025

    2025

  7. [15]

    Zheng, K

    C. Zheng, K. Chen, Z. Bi, Y . Li, L. Pan, J. Zhou, H. Li, and J. Ma. Embracing bulky ob- jects with humanoid robots: Whole-body manipulation with reinforcement learning. InIEEE International Conference on Robotics and Automation, page 16930, 2026

    2026

  8. [16]

    F. Liu, Z. Gu, Y . Cai, Z. Zhou, H. Jung, J. Jang, S. Zhao, S. Ha, Y . Chen, D. Xu, and Y . Zhao. Opt2skill: Imitating dynamically-feasible whole-body trajectories for versatile humanoid loco- manipulation.IEEE Robotics and Automation Letters, 10(11):12261–12268, 2025

    2025

  9. [17]

    Chen, Z.-a

    S. Chen, Z.-a. Cao, Z. Luo, F. Casta ˜neda, C. Li, T. Wang, Y . Yuan, L. Fan, C. K. Liu, and Y . Zhu. Chip: Learning adaptive compliance for humanoid control through hindsight pertur- bation. 2025

    2025

  10. [18]

    L. Wei, X. Peng, R.-Z. Qiu, T. Huang, X. Cheng, and X. Wang. Hmc: Learning heteroge- neous meta-control for contact-rich loco-manipulation. InIEEE International Conference on Robotics and Automation, 2026

    2026

  11. [19]

    G. B. Margolis, M. Wang, N. Fey, and P. Agrawal. SoftMimic: Learning compliant whole-body control from examples. 2025

    2025

  12. [20]

    F. Wu, X. Nal, J. Jang, W. Zhu, Z. Gu, A. Wu, and Y . Zhao. Learn to teach: Sample-efficient privileged learning for humanoid locomotion over real-world uneven terrain. InIEEE Robotics and Automation Letters, pages 9048–9055, 2025

    2025

  13. [21]

    Z. Gu, Y . Chen, Z. Chai, A. Cueva, T. Nguyen, Y . Wu, H. Xue, M. Kim, I. Legene, F. Liu, M. Kim, A. Barula, Y . Chen, and Y . Zhao. Refine-dp: Diffusion policy fine-tuning for hu- manoid loco-manipulation via reinforcement learning, 2026

    2026

  14. [22]

    Murooka, T

    M. Murooka, T. Hoshi, K. Fukumitsu, S. Masuda, M. Hamze, T. Sasaki, M. Morisawa, and E. Yoshida. Tact: Humanoid whole-body contact manipulation through deep imitation learning with tactile modality.IEEE Robotics and Automation Letters, 10(8):7819–7826, 2025

    2025

  15. [23]

    Khatib, M

    O. Khatib, M. Jorda, J. Park, L. Sentis, and S.-Y . Chung. Constraint-consistent task-oriented whole-body robot formulation: Task, posture, constraints, multiple contacts, and balance.The International Journal of Robotics Research, 41(13-14):1079–1098, 2022

    2022

  16. [24]

    Wijayarathne, Z

    L. Wijayarathne, Z. Zhou, Y . Zhao, and F. L. Hammond. Real-time deformable-contact-aware model predictive control for force-modulated manipulation.IEEE Transactions on Robotics, 39(5):3549–3566, 2023

    2023

  17. [25]

    W. Liu, J. Wang, Y . Wang, W. Wang, and C. Lu. Forcemimic: Force-centric imitation learn- ing with force-motion capture system for contact-rich manipulation. InIEEE International Conference on Robotics and Automation, pages 1105–1112, 2025

    2025

  18. [26]

    H. Choi, Y . Hou, C. Pan, S. Hong, A. Patel, X. Xu, M. R. Cutkosky, and S. Song. In-the- wild compliant manipulation with umi-ft. InIEEE International Conference on Robotics and Automation, 2026

    2026

  19. [27]

    Huang, P

    Y . Huang, P. Lin, W. Li, D. Li, J. Li, J. Jiang, C. Xiao, and Z. Jiao. Taf-vla: Tactile-force alignment in vision-language-action models for force-aware manipulation, 2026

    2026

  20. [28]

    J. Yu, H. Liu, Q. Yu, J. Ren, C. Hao, H. Ding, G. Huang, G. Huang, Y . Song, P. Cai, W. Zhang, and C. Lu. Forcevla: Enhancing vla models with a force-aware moe for contact-rich manipula- tion. InAdvances in Neural Information Processing Systems, volume 38, pages 93409–93439, 2025. 10

    2025

  21. [29]

    Li, Zhaxizhuoma, H

    Y . Li, Zhaxizhuoma, H. Jiang, J. Xia, H. Zhang, J. Du, Y . Zhou, J. Zeng, C. Hao, J. Ren, Q. Yu, C. Lu, Y . Qiao, and J. Pang. Forcevla2: Unleashing hybrid force-position control with force awareness for contact-rich manipulation, 2026

    2026

  22. [30]

    R. Zhao, W. Wang, Y . Ma, X. Li, F. E. H. Tay, M. H. Ang, Jr., and H. Zhu. Fd-vla: Force- distilled vision-language-action model for contact-rich manipulation, 2026

    2026

  23. [31]

    Zhang, Z

    T. Zhang, Z. McCarthy, O. Jow, D. Lee, X. Chen, K. Goldberg, and P. Abbeel. Deep im- itation learning for complex manipulation tasks from virtual reality teleoperation. InIEEE International Conference on Robotics and Automation, pages 5628–5635, 2018

    2018

  24. [32]

    Y . Qin, W. Yang, B. Huang, K. Wyk, H. Su, X. Wang, Y .-W. Chao, and D. Fox. AnyTeleop: A General Vision-Based Dexterous Robot Arm-Hand Teleoperation System. InRobotics: Sci- ence and Systems, 2023. ISBN 978-0-9923747-9-2

    2023

  25. [33]

    Z. Fu, T. Z. Zhao, and C. Finn. Mobile aloha: Learning bimanual mobile manipulation using low-cost whole-body teleoperation. InProceedings of The 8th Conference on Robot Learning, volume 270, pages 4066–4083, 2025

    2025

  26. [34]

    Aldaco, T

    J. Aldaco, T. Armstrong, R. Baruch, J. Bingham, S. Chan, K. Draper, D. Dwibedi, C. Finn, P. Florence, S. Goodrich, et al. Aloha 2: An enhanced low-cost hardware for bimanual teleop- eration. 2024

    2024

  27. [35]

    Y . Ze, S. Zhao, W. Wang, A. Kanazawa, R. Duan, P. Abbeel, G. Shi, J. Wu, and C. K. Liu. Twist2: Scalable, portable, and holistic humanoid data collection system. InIEEE Interna- tional Conference on Robotics and Automation, 2026

    2026

  28. [36]

    C. Chi, Z. Xu, C. Pan, E. Cousineau, B. Burchfiel, S. Feng, R. Tedrake, and S. Song. Universal manipulation interface: In-the-wild robot teaching without in-the-wild robots. InProceedings of Robotics: Science and Systems, page p045, 2024

    2024

  29. [37]

    Y . Wi, J. Yin, E. Xiang, A. Sharma, J. Malik, M. Mukadam, N. Fazeli, and T. Hellebrekers. Tactalign: Human-to-robot policy transfer via tactile alignment, 2026

    2026

  30. [38]

    PICO Immersive Pte. Ltd. PICO 4 Ultra: An All-New Mixed Reality Experience.https: //www.picoxr.com/global/products/pico4-ultra, 2023

    2023

  31. [39]

    Z. Zhao, L. Yu, K. Jing, and N. Yang. XRoboToolkit: A cross-platform framework for robot teleoperation. InIEEE/SICE International Symposium on System Integration, pages 15–20, 2026

    2026

  32. [40]

    Y . Zhou, C. Barnes, J. Lu, J. Yang, and H. Li. On the continuity of rotation representations in neural networks. Inthe IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5745–5753, 2019

    2019

  33. [41]

    Dosovitskiy, L

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. De- hghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale. InInternational Conference on Learning ...

    2021

  34. [42]

    Lipman, R

    Y . Lipman, R. T. Q. Chen, H. Ben-Hamu, M. Nickel, and M. Le. Flow matching for generative modeling. InInternational Conference on Learning Representations, 2023

    2023

  35. [43]

    Bjorck, N

    NVIDIA, J. Bjorck, N. C. Fernando Casta ˜neda, X. Da, R. Ding, L. J. Fan, Y . Fang, D. Fox, F. Hu, S. Huang, J. Jang, Z. Jiang, J. Kautz, K. Kundalia, L. Lao, Z. Li, Z. Lin, K. Lin, G. Liu, E. Llontop, L. Magne, A. Mandlekar, A. Narayan, S. Nasiriany, S. Reed, Y . L. Tan, G. W...

    2025

  36. [44]

    Caron, Y

    S. Caron, Y . De Mont-Marin, R. Budhiraja, S. H. Bang, I. Domrachev, S. Nedelchev, P. Du, A. Escande, J. Vaillant, B. Wingo, S. Patapati, and D. San Jos ´e Pro. Pink: Python inverse kinematics based on Pinocchio, 2026

    2026

  37. [45]

    Black, N

    K. Black, N. Brown, J. Darpinian, K. Dhabalia, D. Driess, A. Esmail, M. Equi, C. Finn, N. Fu- sai, M. Y . Galliker, D. Ghosh, L. Groom, K. Hausman, B. Ichter, S. Jakubczak, T. Jones, L. Ke, D. LeBlanc, S. Levine, A. Li-Bell, M. Mothukuri, S. Nair, K. Pertsch, A. Z. Ren, L. X. ...

    2025

  38. [46]

    S. Wei, H. Jing, B. Li, Z. Zhao, J. Mao, Z. Ni, S. He, J. Liu, X. Liu, K. Kang, S. Zang, W. Yuan, M. Pavone, D. Huang, and Y . Wang.ψ0: An open foundation model towards universal humanoid loco-manipulation. InProceedings of Robotics: Science and Systems, 2026

    2026

  39. [47]

    Y . Lu, Z. Liu, X. Fan, Z. Yang, J. Hou, J. Li, K. Ding, and H. Zhao. Faster: Rethinking real-time flow vlas, 2026

    2026

  40. [48]

    J. Song, C. Meng, and S. Ermon. Denoising diffusion implicit models. InInternational Con- ference on Learning Representations, 2021

    2021

  41. [49]

    Black, A

    K. Black, A. Z. Ren, M. Equi, and S. Levine. Training-time action conditioning for efficient real-time chunking, 2025

    2025

  42. [50]

    SensX thin-film tactile sensors

    TouchTronix Robotics Inc. SensX thin-film tactile sensors. https://www.touchtronix.io/

  43. [51]

    Huang, Y

    B. Huang, Y . Wang, X. Yang, Y . Luo, and Y . Li. 3D-ViTac: Learning fine-grained manipulation with visuo-tactile sensing. InProceedings of The 8th Conference on Robot Learning, volume 270, pages 2557–2578, 2025. 12 8 Supplementary 8.1 Sensor Specification and Force Calibratio...

    2025

Pith tools

Reviewed June 27, 2026 · model on record in the stance chip above.