pith. machine review for the scientific record. sign in

arxiv: 2512.03743 · v6 · submitted 2025-12-03 · 💻 cs.RO · cs.LG

Recognition: 2 theorem links

· Lean Theorem

House of Dextra: Cross-embodied Co-design for Dexterous Hands

Kehlani Fay , Darin Anthony Djapri , Anya Zorin , James Clinton , Ali El Lahib , Hao Su , Michael T. Tolley , Sha Yi
show 1 more author
Xiaolong Wang
Authors on Pith no claims yet

Pith reviewed 2026-05-17 02:39 UTC · model grok-4.3

classification 💻 cs.RO cs.LG
keywords dexterous manipulationco-designrobotic handsmorphology optimizationcross-embodied controlrobot fabricationtask-specific designsim-to-real transfer
0
0 comments X

The pith

A co-design framework jointly optimizes robotic hand shapes and control policies to enable full design, training, fabrication, and deployment in under 24 hours.

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

The paper introduces a framework that simultaneously searches for effective hand morphologies and learns matching dexterous control policies for given tasks. It expands the design space to include choices for joints, fingers, and palms, then evaluates many candidate designs at scale by training control policies that are conditioned on the specific morphology. These designs can be quickly turned into physical hardware using accessible components, closing the loop from idea to working robot hand. A sympathetic reader would care because current dexterous manipulation research often treats hardware design and software control as separate, time-consuming steps that limit rapid progress on tasks such as in-hand rotation.

Core claim

We present a co-design framework that learns task-specific hand morphology and complementary dexterous control policies. The framework supports an expansive morphology search space including joint, finger, and palm generation, scalable evaluation across the wide design space via morphology-conditioned cross-embodied control, and real-world fabrication with accessible components. We evaluate the approach across multiple dexterous tasks, including in-hand rotation with simulation and real deployment. Our framework enables an end-to-end pipeline that can design, train, fabricate, and deploy a new robotic hand in under 24 hours.

What carries the argument

Morphology-conditioned cross-embodied control, which trains policies in simulation that are aware of varying hand shapes so that many morphologies can be evaluated efficiently before any physical build.

If this is right

  • Task-specific hands can be generated and deployed for multiple dexterous manipulation problems without manual redesign.
  • The separation between morphology search and policy learning is removed, allowing faster iteration on manipulation hardware.
  • Real hardware can be produced from simulation results using standard accessible parts rather than custom manufacturing.

Where Pith is reading between the lines

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

  • This approach could shorten the typical multi-month cycle of robotic hand development to daily cycles, letting researchers test more morphology ideas.
  • Similar co-design methods might apply to other robot components where shape and control interact strongly, such as legs or grippers for different environments.
  • If open-sourced as promised, the pipeline could serve as a testbed for studying how much hand design variety is actually needed for broad dexterity.

Load-bearing premise

Policies trained in simulation for a given hand morphology will transfer to the corresponding physical hand with little extra tuning, and the fabrication constraints will not exclude high-performing designs.

What would settle it

A fabricated hand performing substantially below its simulated performance on the target task even after basic real-world policy fine-tuning would show the transfer assumption does not hold.

Figures

Figures reproduced from arXiv: 2512.03743 by Ali El Lahib, Anya Zorin, Darin Anthony Djapri, Hao Su, James Clinton, Kehlani Fay, Michael T. Tolley, Sha Yi, Xiaolong Wang.

Figure 1
Figure 1. Figure 1: We present our cross-embodied co-design framework for dexterous hands that jointly [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overview of our method. Stage 1: We first randomly sample embodiments, including morphology and degrees of freedom. We then pre-train a morphology-conditioned policy across embodiments. Stage 2: The design is generated using modular grammars, with assembly rules se￾lected based on prior performance. The cross-embodied policy is then deployed to evaluate different designs in simulation. Stage 3: Best design… view at source ↗
Figure 3
Figure 3. Figure 3: Left to Right. 1) Variable finger length of 1-10 using flat modular stacks. 2) Example [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Top Left: Three finger co-design hand, best design found by algorithm. Top Right: Five finger symmetric co-design hand. Bottom Left: Four finger co-design hand with thin fingertips. Bottom Right: Five finger anthropomorphic baseline. fall of 4.63 seconds for the in-hand rotation task. By contrast, our best designs without fine tuning achieves a 1.85 rad/sec in simulation with no time to fall over the evalu… view at source ↗
Figure 5
Figure 5. Figure 5: Top Left: Grasping task with hold time over a minute and quick grasps (< 0.1 sec).Bottom: Best found design of flipping task where an object rotates along the z-axis using a fixed wrist. Bottom Right: Normalized reward of best design for Flipping, Grasping, and In-Hand Rotation. formed asymmetric designs, including those with thumb layouts. As the hardest task due to a fixed wrist and variable object heigh… view at source ↗
Figure 6
Figure 6. Figure 6: Correlation of physical parameters to improving rotation from Bayesian Sampling on a LEAP hand. Parameters of highest impact are morphology and control parameters. Our analysis ( [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Exploded view of modular robot hand, anthropomorphic baseline. [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Two sampled co-design hands including the 5 finger with standard fingertips and four [PITH_FULL_IMAGE:figures/full_fig_p019_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Modular anthropomorphic robot hand baseline. [PITH_FULL_IMAGE:figures/full_fig_p019_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: The object dataset used for evaluating designs across tasks in simulation. Objects were [PITH_FULL_IMAGE:figures/full_fig_p020_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Five parameters are optimized including fingertip type from selection shapes inspired [PITH_FULL_IMAGE:figures/full_fig_p020_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Random examples of generated robot hands. [PITH_FULL_IMAGE:figures/full_fig_p021_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Design generated using RoboGrammar. A. RoboGrammar with direction-aware reward, [PITH_FULL_IMAGE:figures/full_fig_p022_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Top four highest parameters of impact with 100 samples completed for each parameter. [PITH_FULL_IMAGE:figures/full_fig_p022_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: All sampled parameters using Bayesian sampling on the LEAP robot hand and found [PITH_FULL_IMAGE:figures/full_fig_p023_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Ranking of each found parameter on reward impact and sensitivity of each parameter to [PITH_FULL_IMAGE:figures/full_fig_p023_16.png] view at source ↗
read the original abstract

Dexterous manipulation is limited by both control and design, without consensus as to what makes manipulators best for performing dexterous tasks. This raises a fundamental challenge: how should we design and control robot manipulators that are optimized for dexterity? We present a co-design framework that learns task-specific hand morphology and complementary dexterous control policies. The framework supports 1) an expansive morphology search space including joint, finger, and palm generation, 2) scalable evaluation across the wide design space via morphology-conditioned cross-embodied control, and 3) real-world fabrication with accessible components. We evaluate the approach across multiple dexterous tasks, including in-hand rotation with simulation and real deployment. Our framework enables an end-to-end pipeline that can design, train, fabricate, and deploy a new robotic hand in under 24 hours. The full framework will be open-sourced and available on our website: https://an-axolotl.github.io/HouseofDextra/ .

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper presents House of Dextra, a co-design framework for dexterous robotic hands that jointly optimizes task-specific morphologies (including joints, fingers, and palm) and complementary control policies via morphology-conditioned cross-embodied control. It supports an expansive design space, scalable simulation-based evaluation, and fabrication with accessible components. The central claim is that the framework enables a complete end-to-end pipeline to design, train, fabricate, and deploy a new hand in under 24 hours, demonstrated through evaluations on dexterous tasks such as in-hand rotation in both simulation and real hardware.

Significance. If the quantitative results and timing claims hold, this would represent a meaningful advance in robotic manipulation by providing a practical, rapid co-design pipeline that addresses both morphology and control without requiring extensive manual tuning. The commitment to open-sourcing the full framework and demonstrating real deployment are notable strengths that could support reproducibility and adoption in the field.

major comments (1)
  1. [Abstract / Evaluation] The abstract and description claim an end-to-end pipeline completing in under 24 hours with real deployment on tasks like in-hand rotation, yet no wall-clock timing breakdown, quantitative performance metrics, ablation studies, or explicit analysis of sim-to-real transfer gaps are provided. This directly impacts the load-bearing central claim, as unmodeled effects in contact-rich tasks (e.g., friction variance or actuator dynamics) could necessitate additional real-world adaptation steps that exceed the stated time bound.
minor comments (2)
  1. [Method] The description of the morphology search space and cross-embodied policy training would benefit from a diagram or pseudocode to clarify how morphology parameters condition the policy network.
  2. [Fabrication] Ensure all claims about 'accessible components' for fabrication include specific examples or a bill of materials to aid reproducibility.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive feedback and for acknowledging the potential significance of the House of Dextra co-design framework. We agree that the central claim of an end-to-end pipeline completing in under 24 hours requires more detailed substantiation, and we will revise the manuscript to provide the requested timing breakdown, metrics, ablations, and sim-to-real analysis.

read point-by-point responses

  1. Referee: [Abstract / Evaluation] The abstract and description claim an end-to-end pipeline completing in under 24 hours with real deployment on tasks like in-hand rotation, yet no wall-clock timing breakdown, quantitative performance metrics, ablation studies, or explicit analysis of sim-to-real transfer gaps are provided. This directly impacts the load-bearing central claim, as unmodeled effects in contact-rich tasks (e.g., friction variance or actuator dynamics) could necessitate additional real-world adaptation steps that exceed the stated time bound.

    Authors: We acknowledge that the current manuscript does not provide a granular wall-clock timing breakdown or the other requested analyses. In the revised version we will add a dedicated timing section with a breakdown of each pipeline stage (morphology optimization, cross-embodied policy training, fabrication, and real-world deployment) drawn from our experimental logs, confirming the total remains under 24 hours. We will also report quantitative performance metrics (e.g., success rates and rotation accuracy) for the in-hand rotation task in both simulation and hardware. Ablation studies comparing the full morphology-conditioned controller against fixed-morphology and non-cross-embodied baselines will be included. Finally, we will add an explicit sim-to-real analysis section discussing contact-rich effects such as friction variance and actuator dynamics, and report that our deployed policies transferred directly without additional adaptation steps that would exceed the time bound. These revisions will be placed in the evaluation and supplementary material sections. revision: yes

Circularity Check

0 steps flagged

Empirical co-design pipeline with no derivation chain or circular reductions

full rationale

The paper presents a framework for co-designing dexterous hands via morphology search, morphology-conditioned cross-embodied policies trained in simulation, and accessible fabrication. The end-to-end 24-hour claim is an empirical assertion demonstrated through task evaluations (e.g., in-hand rotation) with real deployment, rather than a mathematical derivation. No equations, first-principles results, fitted parameters renamed as predictions, or load-bearing self-citations appear in the provided text that would reduce the central claims to their own inputs by construction. The approach is self-contained as an experimental pipeline evaluated against external hardware benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework rests on the assumption that simulation-based evaluation with morphology-conditioned policies is predictive of real-world dexterity and that accessible components suffice for high-performing designs.

free parameters (1)
  • Morphology search parameters
    Parameters defining joint types, finger counts, and palm geometry are searched and optimized within the framework.
axioms (1)
  • domain assumption Morphology-conditioned policies can be trained scalably across a wide design space
    This underpins the claim of scalable evaluation without per-design retraining from scratch.

pith-pipeline@v0.9.0 · 5502 in / 1205 out tokens · 30463 ms · 2026-05-17T02:39:04.989051+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Function-based Parametric Co-Design Optimization of Dexterous Hands

    cs.RO 2026-04 unverdicted novelty 6.0

    A unified parametric framework optimizes dexterous hand designs by combining structure, kinematics, and fine surface geometry for grasp stability in simulation and real-world tests.

Reference graph

Works this paper leans on

72 extracted references · 72 canonical work pages · cited by 1 Pith paper

  1. [1]

    The reality gap in robotics: Challenges, solutions, and best practices

    Elie Aljalbout, Jiaxu Xing, Angel Romero, Iretiayo Akinola, Caelan Garrett, Eric Heiden, Abhishek Gupta, Tucker Hermans, Yash Narang, Dieter Fox, Davide Scaramuzza, and Fabio Ramos. The reality gap in robotics: Challenges, solutions, and best practices. arXiv preprint, 2025

    work page 2025

  2. [2]

    AH V4 Main

    Allegro Hand . AH V4 Main . URL https://www.allegrohand.com/v4-main

    work page

  3. [3]

    Learning dexterous in-hand manipulation

    OpenAI: Marcin Andrychowicz, Bowen Baker, Maciek Chociej, Rafal Józefowicz, Bob McGrew, Jakub Pachocki, Arthur Petron, Matthias Plappert, Glenn Powell, Alex Ray, Jonas Schneider, Szymon Sidor, Josh Tobin, Peter Welinder, Lilian Weng, and Wojciech Zaremba. Learning dexterous in-hand manipulation. The International Journal of Robotics Research, 39 0 (1): 0 ...

    work page doi:10.1177/0278364919887447 2020

  4. [4]

    Geometry Matching for Multi - Embodiment Grasping

    Maria Attarian, Muhammad Adil Asif, Jingzhou Liu, Ruthrash Hari, Animesh Garg, Igor Gilitschenski, and Jonathan Tompson. Geometry Matching for Multi - Embodiment Grasping . In Proceedings of The 7th Conference on Robot Learning , pp.\ 1242--1256. PMLR, December 2023. ISSN: 2640-3498

    work page 2023

  5. [5]

    Tolley, Yan Duan, Pieter Abbeel, Xiaolong Wang, and Sha Yi

    Xueqian Bai, Nicklas Hansen, Adabhav Singh, Michael T. Tolley, Yan Duan, Pieter Abbeel, Xiaolong Wang, and Sha Yi. Learning to design soft hands using reward models, 2025. URL https://arxiv.org/abs/2510.17086

    work page arXiv 2025

  6. [6]

    More than a feeling: Learning to grasp and regrasp using vision and touch

    Roberto Calandra, Andrew Owens, Dinesh Jayaraman, Justin Lin, Wenzhen Yuan, Jitendra Malik, Edward H Adelson, and Sergey Levine. More than a feeling: Learning to grasp and regrasp using vision and touch. IEEE Robotics and Automation Letters, 3 0 (4): 0 3300--3307, 2018

    work page 2018

  7. [7]

    Robot co-design: Beyond the monotone case

    Luca Carlone and Carlo Pinciroli. Robot co-design: Beyond the monotone case. In 2019 International Conference on Robotics and Automation (ICRA), pp.\ 3024--3030, 2019. doi:10.1109/ICRA.2019.8793926

    work page doi:10.1109/icra.2019.8793926 2019

  8. [8]

    Co-designing hardware and control for robot hands

    Tianjian Chen, Zhanpeng He, and Matei Ciocarlie. Co-designing hardware and control for robot hands. Science Robotics, 6 0 (54): 0 eabg2133, 2021

    work page 2021

  9. [9]

    Unshackling evolution: evolving soft robots with multiple materials and a powerful generative encoding

    Nick Cheney, Robert MacCurdy, Jeff Clune, and Hod Lipson. Unshackling evolution: evolving soft robots with multiple materials and a powerful generative encoding. In Proceedings of the 15th annual conference on Genetic and evolutionary computation , GECCO '13, pp.\ 167--174, New York, NY, USA, July 2013. Association for Computing Machinery. ISBN 978-1-4503...

    work page doi:10.1145/2463372.2463404 2013

  10. [10]

    Xianyi Cheng, Eric Huang, Yifan Hou, and Matthew T. Mason. Contact mode guided sampling-based planning for quasistatic dexterous manipulation in 2d. In 2021 IEEE International Conference on Robotics and Automation (ICRA), pp.\ 6520--6526, 2021. doi:10.1109/ICRA48506.2021.9560766

    work page doi:10.1109/icra48506.2021.9560766 2021

  11. [11]

    Christoph, Maximilian Eberlein, Filippos Katsimalis, Arturo Roberti, Aristotelis Sympetheros, Michel R

    Clemens C. Christoph, Maximilian Eberlein, Filippos Katsimalis, Arturo Roberti, Aristotelis Sympetheros, Michel R. Vogt, Davide Liconti, Chenyu Yang, Barnabas Gavin Cangan, Ronan J. Hinchet, and Robert K. Katzschmann. ORCA : An Open - Source , Reliable , Cost - Effective , Anthropomorphic Robotic Hand for Uninterrupted Dexterous Task Learning , September ...

    work page arXiv 2025

  12. [12]

    A versatile co-design approach for dynamic legged robots

    Traiko Dinev, Carlos Mastalli, Vladimir Ivan, Steve Tonneau, and Sethu Vijayakumar. A versatile co-design approach for dynamic legged robots. In 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems, pp.\ 10343--10349, United States, December 2022. IEEE. doi:10.1109/IROS47612.2022.9981378. Funding Information: This research was supporte...

    work page doi:10.1109/iros47612.2022.9981378 2022

  13. [13]

    Symmetry-aware robot design with structured subgroups

    Heng Dong, Junyu Zhang, Tonghan Wang, and Chongjie Zhang. Symmetry-aware robot design with structured subgroups. In International Conference on Machine Learning, pp.\ 8334--8355. PMLR, 2023 a

    work page 2023

  14. [14]

    Leveraging hyperbolic embeddings for coarse-to-fine robot design

    Heng Dong, Junyu Zhang, and Chongjie Zhang. Leveraging hyperbolic embeddings for coarse-to-fine robot design. arXiv preprint arXiv:2311.00462, 2023 b

    work page arXiv 2023

  15. [15]

    Cagecoopt: enhancing manipulation robustness through caging-guided morphology and policy co-optimization

    Yifei Dong, Shaohang Han, Xianyi Cheng, Werner Friedl, Rafael I Cabral Muchacho, M \'a ximo A Roa, Jana Tumova, and Florian T Pokorny. Cagecoopt: enhancing manipulation robustness through caging-guided morphology and policy co-optimization. arXiv preprint arXiv:2409.11113, 2024

    work page arXiv 2024

  16. [16]

    Scaling Cross - Embodied Learning : One Policy for Manipulation , Navigation , Locomotion and Aviation

    Ria Doshi, Homer Rich Walke, Oier Mees, Sudeep Dasari, and Sergey Levine. Scaling Cross - Embodied Learning : One Policy for Manipulation , Navigation , Locomotion and Aviation . September 2024

    work page 2024

  17. [17]

    Underwater Soft Robot Modeling and Control With Differentiable Simulation

    Tao Du, Josie Hughes, Sebastien Wah, Wojciech Matusik, and Daniela Rus. Underwater Soft Robot Modeling and Control With Differentiable Simulation . IEEE Robotics and Automation Letters, 6 0 (3): 0 4994--5001, July 2021. ISSN 2377-3766. doi:10.1109/LRA.2021.3070305

    work page doi:10.1109/lra.2021.3070305 2021

  18. [18]

    A system for morphology-task generalization via unified representation and behavior distillation

    Hiroki Furuta, Yusuke Iwasawa, Yutaka Matsuo, and Shixiang Shane Gu. A system for morphology-task generalization via unified representation and behavior distillation. arXiv preprint arXiv:2211.14296, 2022

    work page arXiv 2022

  19. [19]

    Embodied intelligence via learning and evolution

    Agrim Gupta, Silvio Savarese, Surya Ganguli, and Li Fei-Fei. Embodied intelligence via learning and evolution. Nature Communications, 12 0 (1): 0 5721, October 2021. ISSN 2041-1723. doi:10.1038/s41467-021-25874-z. Publisher: Nature Publishing Group

    work page doi:10.1038/s41467-021-25874-z 2021

  20. [20]

    Fit2form: 3d generative model for robot gripper form design

    Huy Ha, Shubham Agrawal, and Shuran Song. Fit2form: 3d generative model for robot gripper form design. In Conference on Robot Learning, pp.\ 176--187. PMLR, 2021

    work page 2021

  21. [21]

    Task-based limb optimization for legged robots

    Sehoon Ha, Stelian Coros, Alexander Alspach, Joohyung Kim, and Katsu Yamane. Task-based limb optimization for legged robots. In 2016 IEEE / RSJ International Conference on Intelligent Robots and Systems ( IROS ) , pp.\ 2062--2068, October 2016. doi:10.1109/IROS.2016.7759324. ISSN: 2153-0866

    work page doi:10.1109/iros.2016.7759324 2016

  22. [22]

    What went wrong? closing the sim-to-real gap via differentiable causal discovery

    Peide Huang, Xilun Zhang, Ziang Cao, Shiqi Liu, Mengdi Xu, Wenhao Ding, Jonathan Francis, Bingqing Chen, and Ding Zhao. What went wrong? closing the sim-to-real gap via differentiable causal discovery. In Jie Tan, Marc Toussaint, and Kourosh Darvish (eds.), Proceedings of The 7th Conference on Robot Learning, volume 229 of Proceedings of Machine Learning ...

    work page 2023

  23. [23]

    One Policy to Control Them All : Shared Modular Policies for Agent - Agnostic Control

    Wenlong Huang, Igor Mordatch, and Deepak Pathak. One Policy to Control Them All : Shared Modular Policies for Agent - Agnostic Control . In Proceedings of the 37th International Conference on Machine Learning , pp.\ 4455--4464. PMLR, November 2020. ISSN: 2640-3498

    work page 2020

  24. [24]

    Task-based design and policy co-optimization for tendon-driven underactuated kinematic chains

    Sharfin Islam, Zhanpeng He, and Matei Ciocarlie. Task-based design and policy co-optimization for tendon-driven underactuated kinematic chains. In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp.\ 12016--12023. IEEE, 2024

    work page 2024

  25. [25]

    Computational design of passive grippers

    Milin Kodnongbua, Ian Good Yu Lou, Jeffrey Lipton, and Adriana Schulz. Computational design of passive grippers. arXiv preprint arXiv:2306.03174, 2023

    work page arXiv 2023

  26. [26]

    Generating freeform endoskeletal robots

    Muhan Li, Lingji Kong, and Sam Kriegman. Generating freeform endoskeletal robots. arXiv preprint arXiv:2412.01036, 2024

    work page arXiv 2024

  27. [27]

    Hod Lipson and Jordan B. Pollack. Automatic design and manufacture of robotic lifeforms. Nature, 406 0 (6799): 0 974--978, August 2000. ISSN 1476-4687. doi:10.1038/35023115. Publisher: Nature Publishing Group

    work page doi:10.1038/35023115 2000

  28. [28]

    Learning to design and use tools for robotic manipulation

    Ziang Liu, Stephen Tian, Michelle Guo, C Karen Liu, and Jiajun Wu. Learning to design and use tools for robotic manipulation. arXiv preprint arXiv:2311.00754, 2023

    work page arXiv 2023

  29. [29]

    Bodygen: Advancing towards efficient embodiment co-design

    Haofei Lu, Zhe Wu, Junliang Xing, Jianshu Li, Ruoyu Li, Zhe Li, and Yuanchun Shi. Bodygen: Advancing towards efficient embodiment co-design. arXiv preprint arXiv:2503.00533, 2025

    work page arXiv 2025

  30. [30]

    Data-efficient Co - Adaptation of Morphology and Behaviour with Deep Reinforcement Learning

    Kevin Sebastian Luck, Heni Ben Amor, and Roberto Calandra. Data-efficient Co - Adaptation of Morphology and Behaviour with Deep Reinforcement Learning . In Proceedings of the Conference on Robot Learning , pp.\ 854--869. PMLR, May 2020. ISSN: 2640-3498

    work page 2020

  31. [31]

    Designing Anthropomorphic Soft Hands Through Interaction

    Pragna Mannam, Kenneth Shaw, Dominik Bauer, Jean Oh, Deepak Pathak, and Nancy Pollard. Designing Anthropomorphic Soft Hands Through Interaction . In 2023 IEEE - RAS 22nd International Conference on Humanoid Robots ( Humanoids ) , pp.\ 1--8, December 2023. doi:10.1109/Humanoids57100.2023.10375195. ISSN: 2164-0580

    work page doi:10.1109/humanoids57100.2023.10375195 2023

  32. [32]

    Efficient automatic design of robots

    David Matthews, Andrew Spielberg, Daniela Rus, Sam Kriegman, and Josh Bongard. Efficient automatic design of robots. Proceedings of the National Academy of Sciences, 120 0 (41): 0 e2305180120, 2023

    work page 2023

  33. [33]

    Neural Posterior Domain Randomization

    Fabio Muratore, Theo Gruner, Florian Wiese, Boris Belousov, Michael Gienger, and Jan Peters. Neural Posterior Domain Randomization . In Proceedings of the 5th Conference on Robot Learning , pp.\ 1532--1542. PMLR, January 2022. ISSN: 2640-3498

    work page 2022

  34. [34]

    Fearing, Pieter Abbeel, Sergey Levine, and Chelsea Finn

    Anusha Nagabandi, Ignasi Clavera, Simin Liu, Ronald S. Fearing, Pieter Abbeel, Sergey Levine, and Chelsea Finn. Learning to Adapt in Dynamic , Real - World Environments through Meta - Reinforcement Learning . September 2018

    work page 2018

  35. [35]

    Open x-embodiment: Robotic learning datasets and rt-x models: Open x-embodiment collaboration 0

    Abby O’Neill, Abdul Rehman, Abhiram Maddukuri, Abhishek Gupta, Abhishek Padalkar, Abraham Lee, Acorn Pooley, Agrim Gupta, Ajay Mandlekar, Ajinkya Jain, et al. Open x-embodiment: Robotic learning datasets and rt-x models: Open x-embodiment collaboration 0. In 2024 IEEE International Conference on Robotics and Automation (ICRA), pp.\ 6892--6903. IEEE, 2024

    work page 2024

  36. [36]

    Deepak Pathak, Chris Lu, Trevor Darrell, Phillip Isola, and Alexei A. Efros. Learning to control self-assembling morphologies: a study of generalization via modularity. In Proceedings of the 33rd International Conference on Neural Information Processing Systems , pp.\ 2295--2305, Red Hook, NY, USA, December 2019. Curran Associates Inc

    work page 2019

  37. [37]

    Sim-to- Real Transfer of Robotic Control with Dynamics Randomization

    Xue Bin Peng, Marcin Andrychowicz, Wojciech Zaremba, and Pieter Abbeel. Sim-to- Real Transfer of Robotic Control with Dynamics Randomization . In 2018 IEEE International Conference on Robotics and Automation ( ICRA ) , pp.\ 3803--3810, May 2018. doi:10.1109/ICRA.2018.8460528. ISSN: 2577-087X

    work page doi:10.1109/icra.2018.8460528 2018

  38. [38]

    Piazza, G

    C. Piazza, G. Grioli, M. G. Catalano, and A. Bicchi. A Century of Robotic Hands . Annual Review of Control, Robotics, and Autonomous Systems, 2 0 (Volume 2, 2019): 0 1--32, May 2019. ISSN 2573-5144. doi:10.1146/annurev-control-060117-105003. Publisher: Annual Reviews

    work page doi:10.1146/annurev-control-060117-105003 2019

  39. [39]

    In- Hand Object Rotation via Rapid Motor Adaptation

    Haozhi Qi, Ashish Kumar, Roberto Calandra, Yi Ma, and Jitendra Malik. In- Hand Object Rotation via Rapid Motor Adaptation . In Proceedings of The 6th Conference on Robot Learning , pp.\ 1722--1732. PMLR, March 2023. ISSN: 2640-3498

    work page 2023

  40. [40]

    Combining sampling- and gradient-based planning for contact-rich manipulation

    Filippo Rozzi, Loris Roveda, and Kevin Haninger. Combining sampling- and gradient-based planning for contact-rich manipulation. In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp.\ 9901--9907, 2024. doi:10.1109/IROS58592.2024.10802369

    work page doi:10.1109/iros58592.2024.10802369 2024

  41. [41]

    Charles Schaff, David Yunis, Ayan Chakrabarti, and Matthew R. Walter. Jointly Learning to Construct and Control Agents using Deep Reinforcement Learning . In 2019 International Conference on Robotics and Automation ( ICRA ) , pp.\ 9798--9805, Montreal, QC, Canada, May 2019. IEEE Press. doi:10.1109/ICRA.2019.8793537

    work page doi:10.1109/icra.2019.8793537 2019

  42. [42]

    Charles Schaff, Audrey Sedal, and Matthew R. Walter. Soft robots learn to crawl: Jointly optimizing design and control with sim-to-real transfer, 2022

    work page 2022

  43. [43]

    Charles Schaff, Audrey Sedal, Shiyao Ni, and Matthew R. Walter. Sim-to-real transfer of co-optimized soft robot crawlers. Autonomous Robots, 47 0 (8): 0 1195--1211, December 2023. ISSN 1573-7527. doi:10.1007/s10514-023-10130-8

    work page doi:10.1007/s10514-023-10130-8 2023

  44. [44]

    LEAP Hand : Low - Cost , Efficient , and Anthropomorphic Hand for Robot Learning

    Kenneth Shaw, Ananye Agarwal, and Deepak Pathak. LEAP Hand : Low - Cost , Efficient , and Anthropomorphic Hand for Robot Learning . volume 19, July 2023. ISBN 978-0-9923747-9-2

    work page 2023

  45. [45]

    Accelerated co-design of robots through morphological pretraining, February 2025

    Luke Strgar and Sam Kriegman. Accelerated co-design of robots through morphological pretraining, February 2025. arXiv:2502.10862 [cs]

    work page arXiv 2025

  46. [46]

    Domain randomization for transferring deep neural networks from simulation to the real world

    Josh Tobin, Rachel Fong, Alex Ray, Jonas Schneider, Wojciech Zaremba, and Pieter Abbeel. Domain randomization for transferring deep neural networks from simulation to the real world. In 2017 IEEE / RSJ International Conference on Intelligent Robots and Systems ( IROS ) , pp.\ 23--30, September 2017. doi:10.1109/IROS.2017.8202133. ISSN: 2153-0866

    work page doi:10.1109/iros.2017.8202133 2017

  47. [47]

    Unitree Dex5 -1 Smart Adaptability , Instant Responsiveness

    Unitree Robotics . Unitree Dex5 -1 Smart Adaptability , Instant Responsiveness . URL https://www.unitree.com/Dex5-1

    work page

  48. [48]

    Lessons from Learning to Spin “ Pens ”

    Jun Wang, Ying Yuan, Haichuan Che, Haozhi Qi, Yi Ma, Jitendra Malik, and Xiaolong Wang. Lessons from Learning to Spin “ Pens ”. In Proceedings of The 8th Conference on Robot Learning , pp.\ 3124--3138. PMLR, January 2025. ISSN: 2640-3498

    work page 2025

  49. [49]

    NerveNet : Learning Structured Policy with Graph Neural Networks

    Tingwu Wang, Renjie Liao, Jimmy Ba, and Sanja Fidler. NerveNet : Learning Structured Policy with Graph Neural Networks . February 2018

    work page 2018

  50. [50]

    Tenenbaum, Daniela Rus, and Chuang Gan

    Tsun-Hsuan Wang, Pingchuan Ma, Andrew Everett Spielberg, Zhou Xian, Hao Zhang, Joshua B. Tenenbaum, Daniela Rus, and Chuang Gan. SoftZoo : A Soft Robot Co -design Benchmark For Locomotion In Diverse Environments . In ICLR , 2023 a

    work page 2023

  51. [51]

    Curriculum-based co-design of morphology and control of voxel-based soft robots

    Yuxing Wang, Shuang Wu, Haobo Fu, Qiang Fu, Tiantian Zhang, Yongzhe Chang, and Xueqian Wang. Curriculum-based co-design of morphology and control of voxel-based soft robots. In The Eleventh International Conference on Learning Representations, 2023 b

    work page 2023

  52. [52]

    Preco: Enhancing generalization in co-design of modular soft robots via brain-body pre-training

    Yuxing Wang, Shuang Wu, Tiantian Zhang, Yongzhe Chang, Haobo Fu, Qiang Fu, and Xueqian Wang. Preco: Enhancing generalization in co-design of modular soft robots via brain-body pre-training. In Conference on Robot Learning, pp.\ 478--498. PMLR, 2023 c

    work page 2023

  53. [53]

    Aggressive driving with model predictive path integral control

    Grady Williams, Paul Drews, Brian Goldfain, James M Rehg, and Evangelos A Theodorou. Aggressive driving with model predictive path integral control. In 2016 IEEE international conference on robotics and automation (ICRA), pp.\ 1433--1440. IEEE, 2016

    work page 2016

  54. [54]

    Universal morphology control via contextual modulation

    Zheng Xiong, Jacob Beck, and Shimon Whiteson. Universal morphology control via contextual modulation. In International Conference on Machine Learning, pp.\ 38286--38300. PMLR, 2023

    work page 2023

  55. [55]

    An end-to-end differentiable framework for contact-aware robot design

    Jie Xu, Tao Chen, Lara Zlokapa, Michael Foshey, Wojciech Matusik, Shinjiro Sueda, and Pulkit Agrawal. An end-to-end differentiable framework for contact-aware robot design. Robotics: Science and Systems, 2021 a

    work page 2021

  56. [56]

    Accelerated Policy Learning with Parallel Differentiable Simulation

    Jie Xu, Viktor Makoviychuk, Yashraj Narang, Fabio Ramos, Wojciech Matusik, Animesh Garg, and Miles Macklin. Accelerated Policy Learning with Parallel Differentiable Simulation . October 2021 b

    work page 2021

  57. [57]

    Multi-objective graph heuristic search for terrestrial robot design

    Jie Xu, Andrew Speilberg, Allan Zhao, Daniela Rus, and Wojciech Matusik. Multi-objective graph heuristic search for terrestrial robot design. In 2021 International conference on robotics and automation (ICRA). IEEE, 2021 c

    work page 2021

  58. [58]

    Dynamics-guided diffusion model for robot manipulator design

    Xiaomeng Xu, Huy Ha, and Shuran Song. Dynamics-guided diffusion model for robot manipulator design. CoRR, 2024

    work page 2024

  59. [59]

    Tolley, and Xiaolong Wang

    Sha Yi, Xueqian Bai, Adabhav Singh, Jianglong Ye, Michael T. Tolley, and Xiaolong Wang. Co-design of soft gripper with neural physics. In Joseph Lim, Shuran Song, and Hae-Won Park (eds.), Proceedings of The 9th Conference on Robot Learning, volume 305 of Proceedings of Machine Learning Research, pp.\ 4313--4327. PMLR, 27--30 Sep 2025 a . URL https://proce...

    work page 2025

  60. [60]

    Co-design of soft gripper with neural physics

    Sha Yi, Xueqian Bai, Adabhav Singh, Jianglong Ye, Michael T Tolley, and Xiaolong Wang. Co-design of soft gripper with neural physics. Conference on Robot Learning (CoRL), 2025 b

    work page 2025

  61. [61]

    Geometric retargeting: A principled, ultrafast neural hand retargeting algorithm

    Zhao-Heng Yin, Changhao Wang, Luis Pineda, Krishna Bodduluri, Tingfan Wu, Pieter Abbeel, and Mustafa Mukadam. Geometric retargeting: A principled, ultrafast neural hand retargeting algorithm. arXiv preprint arXiv:2503.07541, 2025

    work page arXiv 2025

  62. [62]

    Yako, Alex Gruebele, and J

    Shenli Yuan, Lin Shao, Connor L. Yako, Alex Gruebele, and J. Kenneth Salisbury. Design and Control of Roller Grasper V2 for In - Hand Manipulation . In 2020 IEEE / RSJ International Conference on Intelligent Robots and Systems ( IROS ) , pp.\ 9151--9158, October 2020. doi:10.1109/IROS45743.2020.9340953. ISSN: 2153-0866

    work page doi:10.1109/iros45743.2020.9340953 2020

  63. [63]

    Transform2act: Learning a transform-and-control policy for efficient agent design

    Ye Yuan, Yuda Song, Zhengyi Luo, Wen Sun, and Kris Kitani. Transform2act: Learning a transform-and-control policy for efficient agent design. arXiv preprint arXiv:2110.03659, 2021

    work page arXiv 2021

  64. [64]

    RoboPianist : Dexterous Piano Playing with Deep Reinforcement Learning

    Kevin Zakka, Philipp Wu, Laura Smith, Nimrod Gileadi, Taylor Howell, Xue Bin Peng, Sumeet Singh, Yuval Tassa, Pete Florence, Andy Zeng, and Pieter Abbeel. RoboPianist : Dexterous Piano Playing with Deep Reinforcement Learning . August 2023

    work page 2023

  65. [65]

    TossingBot : Learning to Throw Arbitrary Objects with Residual Physics

    Andy Zeng, Shuran Song, Johnny Lee, Alberto Rodriquez, and Thomas A.Funkouser. TossingBot : Learning to Throw Arbitrary Objects with Residual Physics . volume 15, June 2019. ISBN 978-0-9923747-5-4

    work page 2019

  66. [66]

    RoboGrammar : graph grammar for terrain-optimized robot design

    Allan Zhao, Jie Xu, Mina Konaković-Luković, Josephine Hughes, Andrew Spielberg, Daniela Rus, and Wojciech Matusik. RoboGrammar : graph grammar for terrain-optimized robot design. ACM Trans. Graph., 39 0 (6): 0 188:1--188:16, November 2020 a . ISSN 0730-0301. doi:10.1145/3414685.3417831

    work page doi:10.1145/3414685.3417831 2020

  67. [67]

    Sim-to-real transfer in deep reinforcement learning for robotics: a survey

    Wenshuai Zhao, Jorge Peña Queralta, and Tomi Westerlund. Sim-to-real transfer in deep reinforcement learning for robotics: a survey. In 2020 IEEE Symposium Series on Computational Intelligence (SSCI), pp.\ 737--744, 2020 b . doi:10.1109/SSCI47803.2020.9308468

    work page doi:10.1109/ssci47803.2020.9308468 2020

  68. [68]

    Bhattasali, and Lerrel Pinto

    Anya Zorin, Irmak Guzey, Nikhil X. Bhattasali, and Lerrel Pinto. RUKA : Rethinking the Design of Humanoid Hands with Learning . volume 21, June 2025. ISBN 9798990284814

    work page 2025

  69. [69]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

    work page

  70. [70]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should not add it explicitly Type <Return> for now, but then later remove the command n...

    work page

  71. [71]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@first@sw \@firstoftwo \@ifundefined NAT@b*@#2 \@firstoftwo @num @NAT@ctr \@secondoft...

    work page

  72. [72]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibsetup #1 @NAT@ctr @ @openbib .11em \@plus.33em \@minus.07em 4000 4000 `\.\@m @bibit...

    work page