Pith. sign in

REVIEW 3 major objections 6 minor 39 references

Improving Low-Cost Teleoperation: Augmenting GELLO with Force

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read By adding force reflection and a force-sensing learning channel to the low-cost GELLO leader, the paper reports improved task success on three of four contact-rich manipulation tasks and preference for haptics among experienced…

desk verdict A useful engineering extension of GELLO with force feedback and force-conditioned ACT, but the IL comparison is statistically underpowered and the quantitative central claim is not established beyond descriptive means. read the letter →

arxiv 2507.13602 v1 pith:QB5KGMHQ submitted 2025-07-18 cs.RO cs.HCcs.LG

Shivakanth Sujit , Luca Nunziante , Dan Ogawa Lillrank , Rousslan Fernand Julien Dossa , Kai Arulkumaran This is my paper · ORCID
classification cs.ROcs.HCcs.LG
keywords teleoperationforcefeedbackhapticimitationlearningbehavioralcloningactionchunkingtransformercontact-richmanipulationexternaltorqueestimation
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 takes the low-cost GELLO teleoperation handle—a 3D-printed, kinematically scaled leader arm that normally sends only joint positions—and shows that a force channel can be added without making the system expensive or complex. On the human side, the leader's servo motors are driven by a scaled follower-side external torque, so the user feels resistance exactly when the robot contacts the environment. On the learning side, the same torque estimate is fed into an action-chunking transformer policy alongside images and leader joint positions. Reported success rises from 0.62 to 0.93 on drawer opening and from 0.24 to 0.36 on whiteboard erasing, stays near perfect on door opening, and drops from 0.60 to 0.42 on nut assembly. A 20-person user study found no significant workload difference on the standard Task Load Index, but users with robotics experience qualitatively preferred the force-feedback controller.

What carries the argument

The load-bearing mechanism is the force-position (FP) control loop $\tau^{\mathrm{ref}}_l = -k_f\,\tau^{\mathrm{ext}}_f$, in which the follower tracks the leader's joint positions while the estimated external torque on the follower is scaled by $k_f$ and applied to the leader's servo motors as a current command. Because GELLO's Denavit-Hartenberg parameters are a kinematically scaled version of the follower's, the reflected joint torque carries the contact information despite the leader's lack of force sensors. The second mechanism is the ACT policy, an action-chunking transformer that predicts short sequences of joint positions; its observation space is extended from images plus leader joint positions to include the external torque vector, while its targets remain the GELLO joint positions. Together the two mechanisms turn a contact event on the follower into a felt resistance for the operator and a training feature that lets a behavior-cloned policy detect and recover from failed contacts.

What would settle it

Log the external-torque feature $\tau^{\mathrm{ext}}_f$ during the Drawer Opening evaluations and check whether it cleanly separates successful handle grasps from misses; if the free-motion noise floor overlaps the contact-induced values, the claimed 0.62-to-0.93 recovery benefit cannot be attributed to the policy detecting a missed grasp from absent resistance.

Watch

Extended reading notes

Core claim

The central claim is that a low-cost leader robot with no force sensors can provide meaningful haptic feedback and, more importantly, that the same contact signal improves imitation learning on contact-rich manipulation tasks. The control law is force-position reflection, $\tau^{\mathrm{ref}}_l = -k_f\,\tau^{\mathrm{ext}}_f$, where $\tau^{\mathrm{ext}}_f$ is the follower's estimated external joint torque and $k_f$ is a scalar gain; the leader feels a scaled version of what the follower feels. When this force vector is added to the observation of an ACT policy, the trained policy can notice the absence of expected resistance—for example, a missed drawer-grip—and re-grasp, which the position-only policy cannot do. The paper reports three of four tasks improving with force, with the only regression being nut assembly, where force sensitivity is not needed and the bottleneck is initial pick-up.

Load-bearing premise

Everything rests on the follower arm's internally computed external-torque signal being a faithful measure of real contact, when the paper's own figure shows non-zero, noisy torque during free motion.

Editorial extensions

If this is right

  • Force-conditioned ACT policies can recover from failed grasps on their own: drawer opening success rose from 0.62 to 0.93 because the policy recognized a missed grip from absent resistance and re-grasped.
  • The force channel is not universally helpful: nut assembly success fell from 0.60 to 0.42, so contact-force inputs should be enabled selectively by task.
  • A low-cost, sensor-free leader can produce usable haptic feedback if its power supply is upgraded and the reflection gain is scaled down; the original hardware disconnects its motors under torque.
  • User workload ratings were not significantly different between controllers, while experienced teleoperators reported that force feedback gave complementary information to vision.
  • Seed-to-seed success was consistent in the reported evaluations, indicating the ACT training procedure is stable across random initializations in these settings.

Reading between the lines

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

  • The same force-reflection recipe should transfer to any follower arm that exposes an external-torque estimate, potentially making haptic teleoperation a standard add-on rather than a custom build; the paper only demonstrates one follower platform.
  • A deployment-time ablation—zeroing the force input after training—would separate whether the policy needs force online or only benefits from learning with force; the paper does not run this ablation.
  • The grasp-recovery behavior suggests force-conditioned imitation may need fewer demonstrations on contact-rich tasks than position-only imitation, but the paper keeps demonstration count fixed.
  • Because the paper tests a single policy family (ACT), the conclusion that force inputs improve imitation learning is so far specific to action-chunking transformers; diffusion-policy variants remain untested.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper extends the GELLO low-cost teleoperation system with force feedback (force-position control) and with force/torque inputs to ACT imitation learning policies. A 20-participant user study finds no significant NASA-TLX differences but claims qualitative preference from experienced users. IL experiments on four tasks (two simulation, two real-world) report success-rate improvements in three of four tasks when force inputs are added, with the caveat that the baseline and force-conditioned policies are compared on the same demonstrations. The central contribution is a low-cost force augmentation for teleoperation and IL on contact-rich tasks.

Significance. If the reported improvements hold, the paper would provide a low-cost, accessible way to add haptic feedback and force-conditioned policies to GELLO-style teleoperation, with implications for contact-rich imitation learning. The implementation is straightforward and uses standard components (Franka FCI torque estimates, Dynamixel current control, ACT). The paper is honest about non-significant TLX results and includes ethical approval, task randomization, and a clear description of the control architecture. However, the evidence for the main IL claim is statistically weak, and the user-study preference claim is anecdotal; the significance of the contribution is therefore contingent on added statistical support and reproducibility details.

major comments (3)
  1. [III-C, Table II] The central claim that force inputs improve task success rests entirely on the differences between the 'Position' and 'Position + Force' rows of Table II, each entry being the mean over 3 seeds with 15 rollouts. No significance test, per-seed breakdown, or confidence interval is reported. For Whiteboard Erasing the improvement is 0.24 vs 0.36 with SDs of 0.43 and 0.48; for Drawer Opening it is 0.62 vs 0.93 with SDs 0.48 and 0.25. With n=3, these differences are well within seed-level sampling noise; the statement in Section III-C that 'performance across seeds was relatively consistent' is contradicted by these large SDs. The authors should report per-seed success counts and a paired test (e.g., Wilcoxon signed-rank or permutation test) or temper the abstract's 'improved task success on the majority of tasks' accordingly.
  2. [III-B] The user-study section reports no significant TLX differences and then states that 'the few users who were familiar with robots' preferred the force-feedback controller. The abstract elevates this to a claim that 'users with robotics experience preferred our controller.' The number of experienced users, how experience was determined, and the preference distribution are not reported, so this claim is unsupported by the presented data. It should be either quantified (N, proportion, inter-rater agreement) or explicitly labeled as an informal observation.
  3. [II-B, Eq. (2)] The force-feedback gain kf is a critical free parameter for stability and transparency, but its value, tuning procedure, and the 'additional conversion factor' for the Dynamixel current control are not given. Without these, the force feedback implementation is not reproducible, and the force signals used as IL inputs (tau_ext_f) are not characterized. Fig. 2 shows non-zero current in free motion due to noise in the Franka's external-torque estimate; the paper does not state whether this signal is filtered before being used in (2) or as an ACT input. Please provide the gain values, filtering, and a characterization of the torque signal under contact.
minor comments (6)
  1. [IV] In the Conclusion, 'teleportation' should be 'teleoperation'.
  2. [II-C] The maximum-likelihood objective for BC is written as 'max π(at|st)', but the likelihood of a trajectory is a product over timesteps; please clarify the notation.
  3. [II-B] The statement that 'during free motion the leader will not have any force feedback' is only true if tau_ext_f is exactly zero; Fig. 2 shows noise, so the claim should be qualified.
  4. [III-A] The Whiteboard Erasing success criterion 'majority of the mark is erased' is vague; a more objective metric (e.g., pixel coverage) would strengthen the evaluation.
  5. [Table I] The text reports only the largest Wilcoxon signed-rank result (Physical Demand, W=19.0, p=0.06); please report the W and p values for all six TLX items for completeness.
  6. [III-C] Please specify whether the position-only and position+force policies are trained on the same demonstration set or on separate collections; this affects the interpretation of the comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: force feedback and force-conditioned IL are evaluated against independent success criteria, with no fitted term that defines the outcome.

full rationale

The paper's derivation chain contains no step where a claimed prediction or first-principles result reduces to its own input. Force feedback is implemented as FP control through Eq. (2), tau_ref_l = -k_f tau_ext_f, using the Franka Control Interface's external-torque estimate; this is an independent quantity supplied by the robot's internal state estimator, not a term derived from the downstream success metric. The IL comparison in Table II trains ACT policies with and without force inputs while holding the task distribution, demonstrations, and evaluation protocol fixed, and success is defined by task-specific geometric criteria (e.g., drawer opened over 15 cm, door angle at least 0.3 rad, majority of mark erased). The force gain k_f is a hardware tuning constant from bilateral-control design, not a parameter fitted to the reported success rates. Citations to GELLO, ACT, Bi-ACT, and control-theory references provide context and baselines but are not used to justify the paper's central empirical claim; in particular, no uniqueness theorem or prior work by these authors is invoked to forbid alternative control choices. The reviewer's concern about large per-seed variance and absence of significance tests is a statistical-correctness issue, not circularity: even if the 3/4 improvement were sampling noise, the comparison itself is not constructed so that force input logically guarantees improved success. Every stated limitation (e.g., whiteboard occlusion, force-estimate noise in the Fig. 2 caption, Nut Assembly being insensitive to force) is an empirical caveat, not a circular redefinition. Accordingly, the analysis finds no significant circularity and assigns the lowest score.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claims rest on standard domain assumptions about the Franka external-torque estimate, GELLO's kinematic equivalence, and Dynamixel current control, plus the untested transfer of ACT hyperparameters. Two hardware-related quantities, kf and the current conversion factor, are free parameters whose values are not reported. No new theoretical entities are introduced.

free parameters (2)
  • k_f, force-feedback gain in Eq. (2) = Not reported
    Scalar gain scaling follower external torque to leader reference torque. The paper says it is critical for stability and transparency but does not give the value used in the experiments.
  • Dynamixel current-to-torque conversion factor = Not reported
    Section III mentions 'an additional conversion factor' to implement the torque reference on GELLO's DYNAMIXEL motors; the exact calibration is not given.
assumptions (4)
  • domain assumption The Franka Control Interface external torque estimate tau_ext_f is a reliable measure of interaction forces at the follower end-effector.
    Used both for haptic reflection in Eq. (2) and as the force input to ACT; Fig. 2's caption acknowledges noise in this estimate during free motion, so accuracy under contact is assumed.
  • domain assumption GELLO's Denavit-Hartenberg parameters are a scaled version of the follower's, so force reflection can be simplified to a scalar gain.
    Section II-B derives Eq. (2) from kinematic equivalence and then drops the linear/angular decomposition, assuming the linear component dominates and no online pseudo-inversion is needed.
  • domain assumption GELLO's Dynamixel motors can faithfully execute current-based torque references with the modified power supply.
    Section III and footnote 4 describe using current control with a stronger power supply and scaled-down reflected force because of motor disconnect issues, implying hardware limits.
  • domain assumption ACT's default hyperparameters transfer to force-conditioned inputs without retuning.
    Section III says ACT is used 'with its default hyperparameters' apart from beta=50 and chunk size 50; no tuning analysis for force-conditioned inputs is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improving Low-Cost Teleoperation: Augmenting GELLO with Force." pith.science (2026). https://pith.science/paper/QB5KGMHQ

@misc{pith2026250713602,
  author       = {Pith},
  title        = {Pith review of: Improving Low-Cost Teleoperation: Augmenting GELLO with Force},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QB5KGMHQ}},
  note         = {Machine review of arXiv:2507.13602}
}
read the original abstract

In this work we extend the low-cost GELLO teleoperation system, initially designed for joint position control, with additional force information. Our first extension is to implement force feedback, allowing users to feel resistance when interacting with the environment. Our second extension is to add force information into the data collection process and training of imitation learning models. We validate our additions by implementing these on a GELLO system with a Franka Panda arm as the follower robot, performing a user study, and comparing the performance of policies trained with and without force information on a range of simulated and real dexterous manipulation tasks. Qualitatively, users with robotics experience preferred our controller, and the addition of force inputs improved task success on the majority of tasks.

Figures

Figures reproduced from arXiv: 2507.13602 by the authors.

Figure 1
Figure 1. The low-cost GELLO teleoperation system (right), [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Current feedback for joints two (blue) and four [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The ACT architecture. During deployment, the Transformer encoder (middle) takes in images, joint information, and [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Set of manipulation tasks for IL with different action types: grasping, placing, pulling, and wiping. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 32 canonical work pages

  1. [1]

    Champion-level Drone Racing Using Deep Reinforcement Learning,

    E. Kaufmann, L. Bauersfeld, A. Loquercio, M. M ¨uller, V . Koltun, and D. Scaramuzza, “Champion-level Drone Racing Using Deep Reinforcement Learning,” Nature, vol. 620, no. 7976, pp. 982–987, 2023

    work page 2023

  2. [2]

    Robot Parkour Learning,

    Z. Zhuang, Z. Fu, J. Wang, C. G. Atkeson, S. Schwertfeger, C. Finn, and H. Zhao, “Robot Parkour Learning,” in Conference on Robot Learning. PMLR, 2023, pp. 73–92

    work page 2023

  3. [3]

    Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation,

    Z. Fu, T. Z. Zhao, and C. Finn, “Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation,” arXiv preprint arXiv:2401.02117 , 2024

    arXiv 2024

  4. [4]

    Deep Imitation Learning for Humanoid Loco-manipulation Through Human Teleoperation,

    M. Seo, S. Han, K. Sim, S. H. Bang, C. Gonzalez, L. Sentis, and Y . Zhu, “Deep Imitation Learning for Humanoid Loco-manipulation Through Human Teleoperation,” in IEEE-RAS Humanoids . IEEE, 2023, pp. 1–8

    work page 2023

  5. [5]

    ALVINN: An Autonomous Land Vehicle in a Neural Network,

    D. A. Pomerleau, “ALVINN: An Autonomous Land Vehicle in a Neural Network,” in NeurIPS, 1988

    work page 1988

  6. [6]

    RT-1: Robotics Transformer for Real-World Control at Scale,

    A. Brohan, N. Brown, J. Carbajal, Y . Chebotar, J. Dabis, C. Finn, K. Gopalakrishnan, K. Hausman, A. Herzog, J. Hsu, et al. , “RT-1: Robotics Transformer for Real-World Control at Scale,” arXiv preprint arXiv:2212.06817, 2022

    arXiv 2022

  7. [7]

    RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control,

    A. Brohan, N. Brown, J. Carbajal, Y . Chebotar, X. Chen, K. Choro- manski, T. Ding, D. Driess, A. Dubey, C. Finn, et al. , “RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control,” arXiv preprint arXiv:2307.15818 , 2023

    arXiv 2023

  8. [8]

    Robotics for Surgery,

    R. D. Howe and Y . Matsuoka, “Robotics for Surgery,” Annual review of biomedical engineering , vol. 1, no. 1, pp. 211–240, 1999

    work page 1999

Show all 39 references
  1. [9]

    Combined Perception, Control, and Learning for Teleoperation: Key Technologies, Applications, and Chal- lenges,

    J. Luo, W. He, and C. Yang, “Combined Perception, Control, and Learning for Teleoperation: Key Technologies, Applications, and Chal- lenges,” Cognitive Computation and Systems , vol. 2, no. 2, pp. 33–43, 2020

    2020

  2. [10]

    Avatar Work: Telework for Disabled People Unable to Go Outside by Using Avatar Robots,

    K. Takeuchi, Y . Yamazaki, and K. Yoshifuji, “Avatar Work: Telework for Disabled People Unable to Go Outside by Using Avatar Robots,” in Companion of the 2020 ACM/IEEE international conference on human-robot interaction, 2020, pp. 53–60

    2020

  3. [11]

    A survey on deep reinforcement learning algorithms for robotic manipulation,

    D. Han, B. Mulyana, V . Stankovic, and S. Cheng, “A survey on deep reinforcement learning algorithms for robotic manipulation,” Sensors, vol. 23, no. 7, 2023. [Online]. Available: https://www.mdpi.com/ 1424-8220/23/7/3762

    2023

  4. [12]

    Deep Imitation Learning for Complex Manipulation Tasks from Virtual Reality Teleoperation,

    T. Zhang, Z. McCarthy, O. Jow, D. Lee, X. Chen, K. Goldberg, and P. Abbeel, “Deep Imitation Learning for Complex Manipulation Tasks from Virtual Reality Teleoperation,” in 2018 IEEE international conference on robotics and automation (ICRA) . IEEE, 2018, pp. 5628–5635

    2018

  5. [13]

    GELLO: A General, Low-Cost, and Intuitive Teleoperation Framework for Robot Manipulators,

    P. Wu, Y . Shentu, Z. Yi, X. Lin, and P. Abbeel, “GELLO: A General, Low-Cost, and Intuitive Teleoperation Framework for Robot Manipulators,” in Towards Generalist Robots Workshop, CoRL , 2023

    2023

  6. [14]

    Learning Fine-grained Bimanual Manipulation with Low-Cost Hardware,

    T. Z. Zhao, V . Kumar, S. Levine, and C. Finn, “Learning Fine-grained Bimanual Manipulation with Low-Cost Hardware,” in RSS, 2023

    2023

  7. [15]

    Stability and Transparency in Bilateral Teleopera- tion,

    D. A. Lawrence, “Stability and Transparency in Bilateral Teleopera- tion,” IEEE transactions on robotics and automation , vol. 9, no. 5, pp. 624–637, 1993

    1993

  8. [16]

    Still Not Solved: A Call for Renewed Focus on User-Centered Teleoperation Interfaces,

    D. J. Rea and S. H. Seo, “Still Not Solved: A Call for Renewed Focus on User-Centered Teleoperation Interfaces,” Frontiers in Robotics and AI, vol. 9, p. 704225, 2022

    2022

  9. [17]

    The impact of haptic feedback quality on the performance of teleoperated assembly tasks,

    J. G. Wildenbeest, D. A. Abbink, C. J. Heemskerk, F. C. Van Der Helm, and H. Boessenkool, “The impact of haptic feedback quality on the performance of teleoperated assembly tasks,” IEEE Transactions on Haptics , vol. 6, no. 2, pp. 242–252, 2012

    2012

  10. [18]

    Study of multimodal interfaces and the improvements on teleoperation,

    E. Triantafyllidis, C. Mcgreavy, J. Gu, and Z. Li, “Study of multimodal interfaces and the improvements on teleoperation,” IEEE Access , vol. 8, pp. 78 213–78 227, 2020

    2020

  11. [19]

    A meta-analysis of the effects of haptic interfaces on task performance with teleoperation systems,

    V . Nitsch and B. F ¨arber, “A meta-analysis of the effects of haptic interfaces on task performance with teleoperation systems,” IEEE transactions on haptics , vol. 6, no. 4, pp. 387–398, 2012

    2012

  12. [20]

    Experimental quantitative comparison of different control architectures for master-slave teleop- eration,

    I. Aliaga, A. Rubio, and E. Sanchez, “Experimental quantitative comparison of different control architectures for master-slave teleop- eration,” IEEE transactions on control systems technology , vol. 12, no. 1, pp. 2–11, 2004

    2004

  13. [21]

    Design fundamentals of master-slave systems with a force-position bilateral control scheme,

    L. Pe ˜n´ın, M. Ferre, J. Fernandez-Pello, R. Aracil, and A. Barrientos, “Design fundamentals of master-slave systems with a force-position bilateral control scheme,” IF AC Proceedings V olumes, vol. 30, no. 20, pp. 605–612, 1997

    1997

  14. [22]

    Sabanovic and K

    A. Sabanovic and K. Ohnishi, Motion Control Systems . John Wiley & Sons, 2011

    2011

  15. [23]

    Bilateral Control of Master-Slave Manipulators for Ideal Kinesthetic Coupling-Formulation and Exper- iment,

    Y . Yokokohji and T. Yoshikawa, “Bilateral Control of Master-Slave Manipulators for Ideal Kinesthetic Coupling-Formulation and Exper- iment,” IEEE T-RO, vol. 10, no. 5, pp. 605–620, 1994

    1994

  16. [24]

    Imitation learning based on bilateral control for human–robot cooperation,

    A. Sasagawa, K. Fujimoto, S. Sakaino, and T. Tsuji, “Imitation learning based on bilateral control for human–robot cooperation,” IEEE Robotics and Automation Letters , vol. 5, no. 4, pp. 6169–6176, 2020

    2020

  17. [25]

    Imitation learn- ing for variable speed contact motion for operation up to control bandwidth,

    S. Sakaino, K. Fujimoto, Y . Saigusa, and T. Tsuji, “Imitation learn- ing for variable speed contact motion for operation up to control bandwidth,” IEEE Open Journal of the Industrial Electronics Society , vol. 3, pp. 116–127, 2022

    2022

  18. [26]

    Bi-ACT: Bilateral Control-Based Imitation Learning via Action Chunking with Transformer,

    T. Buamanee, M. Kobayashi, Y . Uranishi, and H. Takemura, “Bi-ACT: Bilateral Control-Based Imitation Learning via Action Chunking with Transformer,” in Manipulation Skills Workshop, ICRA , 2024

    2024

  19. [27]

    isora: humanoid robot arm for intelligent haptic interaction with the environment,

    D. Tsetserukou, N. Kawakami, and S. Tachi, “isora: humanoid robot arm for intelligent haptic interaction with the environment,” Advanced Robotics, vol. 23, no. 10, pp. 1327–1358, 2009

    2009

  20. [28]

    Diffusion Policy: Visuomotor Policy Learning via Action Diffusion,

    C. Chi, S. Feng, Y . Du, Z. Xu, E. Cousineau, B. Burchfiel, and S. Song, “Diffusion Policy: Visuomotor Policy Learning via Action Diffusion,” in RSS, 2023

    2023

  21. [29]

    Attention is All You Need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is All You Need,” in NeurIPS, 2017

    2017

  22. [30]

    Learning Structured Output Repre- sentation Using Deep Conditional Generative Models,

    K. Sohn, H. Lee, and X. Yan, “Learning Structured Output Repre- sentation Using Deep Conditional Generative Models,” in NeurIPS, 2015

    2015

  23. [31]

    BERT: Pre-Training of Deep Bidirectional Transformers for Language Understanding,

    J. Devlin, “BERT: Pre-Training of Deep Bidirectional Transformers for Language Understanding,” arXiv preprint arXiv:1810.04805 , 2018

    2018 arXiv

  24. [32]

    Learning Vari- able Compliance Control From a Few Demonstrations for Bimanual Robot with Haptic Feedback Teleoperation System,

    T. Kamijo, C. C. Beltran-Hernandez, and M. Hamaya, “Learning Vari- able Compliance Control From a Few Demonstrations for Bimanual Robot with Haptic Feedback Teleoperation System,” in IROS, 2024

    2024

  25. [33]

    Forward Dynamics Compliance Control (FDCC): A New Approach to Cartesian Com- pliance for Robotic Manipulators,

    S. Scherzinger, A. Roennau, and R. Dillmann, “Forward Dynamics Compliance Control (FDCC): A New Approach to Cartesian Com- pliance for Robotic Manipulators,” in 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2017, pp. 4568–4575

    2017

  26. [34]

    RoboHive: A Unified Framework for Robot Learning,

    V . Kumar, R. Shah, G. Zhou, V . Moens, V . Caggiano, A. Gupta, and A. Rajeswaran, “RoboHive: A Unified Framework for Robot Learning,” in NeurIPS, 2024

    2024

  27. [35]

    MuJoCo: A Physics Engine for Model-Based Control,

    E. Todorov, T. Erez, and Y . Tassa, “MuJoCo: A Physics Engine for Model-Based Control,” in IROS, 2012

    2012

  28. [36]

    robosuite: A Modular Simulation Framework and Benchmark for Robot Learning,

    Y . Zhu, J. Wong, A. Mandlekar, R. Mart´ın-Mart´ın, A. Joshi, S. Nasiri- any, and Y . Zhu, “robosuite: A Modular Simulation Framework and Benchmark for Robot Learning,” arXiv preprint arXiv:2009.12293 , 2020

    2009 arXiv

  29. [37]

    Development of NASA-TLX (Task Load Index): Results of Empirical and Theoretical Research,

    S. G. Hart and L. E. Staveland, “Development of NASA-TLX (Task Load Index): Results of Empirical and Theoretical Research,” in Advances in psychology . Elsevier, 1988, vol. 52, pp. 139–183

    1988

  30. [38]

    Bridge data: Boosting gener- alization of robotic skills with cross-domain datasets,

    F. Ebert, Y . Yang, K. Schmeckpeper, B. Bucher, G. Georgakis, K. Daniilidis, C. Finn, and S. Levine, “Bridge data: Boosting gener- alization of robotic skills with cross-domain datasets,” arXiv preprint arXiv:2109.13396, 2021

    2021 arXiv

  31. [39]

    Open x-embodiment: Robotic learning datasets and rt-x models,

    A. Padalkar, A. Pooley, A. Jain, A. Bewley, A. Herzog, A. Ir- pan, A. Khazatsky, A. Rai, A. Singh, A. Brohan, et al. , “Open x-embodiment: Robotic learning datasets and rt-x models,” arXiv preprint arXiv:2310.08864, 2023

    2023 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.