Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

AnyBimanual: Transferring Unimanual Policy for General Bimanual Manipulation

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This paper claims a pretrained one-arm policy can drive two-arm tasks by re-combining its skills and masking the scene per arm, beating PerAct2 by 17.33 points across 12 bimanual tasks.

desk verdict A sensible transfer method with good ablations, but the 'general' claim outruns the architecture's factorization. read the letter →

arxiv 2412.06779 v2 pith:QTFMWLVX submitted 2024-12-09 cs.RO cs.AI

Guanxing Lu , Tengbo Yu , Haoyuan Deng , Season Si Chen , Yansong Tang , Ziwei Wang This is my paper · ORCID
classification cs.ROcs.AI
keywords bimanualmanipulationunimanualpolicytransferskillprimitivesschedulingvisualalignervoxelsoftmaskinglanguage-conditionedfew-demonstration
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 claims that the general, language-conditioned manipulation knowledge inside a pretrained one-arm (unimanual) policy can be transferred to two-arm (bimanual) tasks using only a small number of bimanual demonstrations, avoiding the expensive data collection that bimanual training normally requires. AnyBimanual installs two lightweight learnable modules in front of the pretrained policy: a skill manager that re-expresses the instruction for each arm as a sparse weighted sum of reusable skill primitives plus a small per-arm compensation, and a visual aligner that soft-masks the shared 3D voxel scene so each arm observes something close to what the unimanual policy saw during pretraining. The authors report that this plug-and-play transfer beats the PerAct2 baseline by 17.33 points in average success rate across 12 RLBench2 bimanual tasks, reaches 84.62% average success on 9 real-world two-arm tasks, and lifts two different base policies (PerAct and RVT) with the same add-ons. If the claim holds, a general bimanual agent can be built from an existing one-arm foundation policy plus tens of demonstrations per task, rather than from large bimanual teleoperation datasets.

What carries the argument

The load-bearing objects are the two learned front-ends inserted between the shared observation and two copies of the pretrained unimanual policy. The skill manager $f_\theta$, a small transformer, predicts at each step per-arm softmax weights $\hat{w}^{\mathrm{arm}}_t$ over $K=18$ skill primitives $z_k$ — embeddings initialized from the text-encoder templates of the unimanual policy's own pretraining tasks — together with compensation vectors $\epsilon^{\mathrm{arm}}_t$, forming each arm's prompt via Eq. (1); its sparse objective (Eq. (2)) is what forces the skill primitives to stay disjoint and reusable. The visual aligner $q_\theta$ predicts soft spatial masks $\hat{v}^{\mathrm{left}}_t$ and $\hat{v}^{\mathrm{right}}_t$ over the voxel embedding, and each arm's input becomes $v^{\mathrm{arm}}_t = (\hat{v}^{\mathrm{arm}}_t \odot v_t) \oplus v_t$ (Eq. (3)), with a Jensen-Shannon divergence term (Eq. (4)) that drives the two masks to be mutually exclusive. The two modules are supervised jointly with per-arm behavior cloning by the total objective (Eq. (6)); their job is to recreate for each arm the input distribution — a language prompt and a scene view — that the unimanual policy encountered in its own pretraining.

What would settle it

Take a task whose success requires within-step physical coupling between the arms — for example, two grippers carrying a single rigid beam to a target without letting it slip, or twisting a stiff lid off a fixed jar where one arm's torque must track the other arm's grip — and compare AnyBimanual against a single unimanual policy with the second arm following a fixed script. If the success rates match, then the claimed decomposition into two conditionally independent unimanual policies has failed for that task.

Watch

Extended reading notes

Core claim

The central discovery the paper argues for is that bimanual manipulation does not need its own scaled-up dataset: an arbitrary pretrained language-conditioned unimanual policy can be converted into a general bimanual policy by learning, from few demonstrations, how to prompt each arm separately from the same instruction and the same scene. The skill manager $f_\theta$ takes instruction, voxel observation, and proprioception, and outputs per-arm combination weights and compensation terms, reconstructing each arm's language embedding as $\hat{l}^{\mathrm{arm}}_t = \sum_{k=1}^K \hat{w}^{\mathrm{arm}}_{k,t} z_k + \epsilon^{\mathrm{arm}}_t$ from a bank of $K=18$ skill primitives; an $\ell^1$-on-weights, $\ell^{2,1}$-on-compensation penalty pushes each arm to use few primitives, keeping the bank interpretable and task-oriented. The visual aligner $q_\theta$ predicts two soft masks over the voxel grid and maximizes their mutual divergence, so the left and right arms attend to disjoint regions and each recovered view resembles the unimanual pretraining distribution. Both modules are trained end-to-end together with a per-arm behavior-cloning loss, and the experiments show the same recipe raises the success of PerAct-style and RVT-style base policies, with the largest gains in long-horizon, multi-variation, and synchronized tasks.

Load-bearing premise

Every bimanual task must be expressible as two independent one-arm sub-policies that coordinate only through a shared instruction and two masked views of the same scene; if a task demands that the arms react to each other's physical state within a single step, the method has no channel for that coupling.

Editorial extensions

If this is right

  • A general bimanual agent can be built from a pretrained one-arm policy plus roughly 20–100 demonstrations per task, and the supplementary reports that 5 demonstrations per real task still gives 53.33% average success.
  • The recipe is not tied to one base policy: the same skill-manager and visual-aligner add-ons raise PerAct-style and RVT-style leader-follower baselines by 72.76% and 39.41% relative on average.
  • The largest gains sit exactly where bimanual methods are traditionally weakest — long-horizon tasks such as put in fridge and take out tray, multi-variation tasks such as press buttons, and synchronized tasks such as straighten rope and lift tray — with only a small regression on the simple short-horizon lift ball task.
  • Decomposability is monitorable: the supplementary reports that the entropy of the predicted skill-combination weights tracks how well a task decomposes and correlates strongly with task success, so the skill manager itself can flag tasks that resist this kind of transfer.

Reading between the lines

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

  • A scaling prediction the paper leaves implicit: because the skill bank is initialized from the base policy's own pretraining templates, a unimanual base trained on a broader task set should widen the range of bimanual tasks the scheduler can express.
  • A testable extension: the mutually exclusive mask prior is a soft partition, so tasks where both arms must attend to the same workspace region (jointly handling one object or one tool) sit outside the demonstrated regime, and swapping the divergence term for a graded-overlap penalty would probe that boundary.
  • The authors' own limitation note implies a next step: because appearance differences between one-arm and two-arm robots are not handled, pairing the aligner with inpainting that erases the other arm rather than merely down-weighting it should strengthen cross-embodiment transfer.
  • The paper also leaves implicit that, since per-arm prompts couple only through shared inputs, adding an explicit cross-arm channel (feeding one arm's predicted action into the other arm's prompt) is a natural architecture change for tightly coupled tasks.
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

4 major / 6 minor

Summary. The paper proposes AnyBimanual, a framework for transferring a pretrained unimanual policy to bimanual manipulation with few bimanual demonstrations. The method has two main components: a skill manager that reconstructs per-arm language embeddings as sparse linear combinations of K skill primitives plus a task compensation term, and a visual aligner that predicts soft spatial masks to decompose voxel observations for each arm. Two copies of the pretrained unimanual policy are then fine-tuned with a behavior-cloning loss plus regularization. The paper reports experiments on 12 RLBench2 tasks and 9 real-world tasks, with an average success improvement of 17.33% over PerAct2 in simulation, and an 84.62% average success rate on the real-robot suite.

Significance. If the claimed results hold, this is a significant contribution: it offers a model-agnostic way to reuse strong unimanual policies for bimanual tasks with limited data, and it provides evidence that skill scheduling and visual alignment can transfer across embodiments. The paper includes a useful ablation (Table 2), a control that pre-trains PerAct2 on unimanual data (PerAct2+Pretraining), real-robot evaluation, an explicit limitations section in the supplement, and the authors state an intent to release code. The main qualification is that the scope of "general bimanual manipulation" is not established, because the architecture rests on a factorization assumption that is neither tested nor bounded.

major comments (4)
  1. [Section 3.2/3.5 and Eq. (1)] The policy factorizes into two conditionally independent unimanual action decoders; coordination can enter only through the shared inputs produced by the skill manager and visual aligner. No term in Ltotal couples the left and right action distributions. The paper does not characterize which bimanual tasks are expressible under this factorization. Table 2 shows the Sync category at only 25.00% success even with both components, and the supplementary "Bimanual Decomposability" metric (Fig. 17) measures the entropy of predicted combination weights, an internal property of the trained manager, not an independent test of task decomposability. The title/abstract claim of "general bimanual manipulation" therefore exceeds the demonstrated scope. Please either provide evidence on more tightly coupled tasks (e.g., handover where the receiving gripper closes as a function of the giver's commanded pose) or explicitly restrict the claim.
  2. [Eq. (4)] The expression -1/2 DKL(v̂_left || v̂_right) - 1/2 DKL(v̂_right || v̂_left) is the negative symmetrized KL divergence, not the Jensen-Shannon divergence, which is defined with respect to the midpoint distribution. Since Lvoxel is part of the training objective and the visual aligner is shown to contribute to performance (Table 2, Row 3 vs Row 2), the manuscript mis-specifies the loss being optimized. Please correct the equation and clarify whether the implemented loss matches the intended JS divergence.
  3. [Section 4.2/Table 1 and Section 4.5/Table 3] Results are reported from what appears to be a single training run per condition, with no error bars or standard deviations across multiple seeds. Given that per-task success rates are often near zero in Table 1 (e.g., handover item at 0-15%) and the real-robot results are based on only 65 total episodes (5 per task/variation), the claimed 17.33% improvement over PerAct2 is not shown to be statistically reliable. Reporting multiple seeds with confidence intervals, or at least per-task episode counts and variances, is necessary to support the headline comparison.
  4. [Section 3.3 and Section 7.3] The "discovered" skill primitives are initialized from the language templates of the very unimanual policy being transferred (e.g., 'open the drawer'), and the compensation term ε can in principle carry the full per-task language signal. The statement that the L1 sparsity loss "requires" orthogonal and disjoint skill subspaces is not justified: sparsity on the combination weights does not by itself enforce orthogonality of the z_k. The supplement reports that learned skill representations cluster near the initialization templates, so please clarify what is actually learned beyond the initialization and whether the sparsity prior has an identifiable effect independent of the initialization.
minor comments (6)
  1. [Abstract] The improvement percentage is given as 12.67% in the first abstract paragraph and as 17.33% in both the second abstract paragraph and the main text; please reconcile this inconsistency.
  2. [Section 4.1] The text says the evaluation uses "100 episodes per task" but does not specify the number of training seeds; please add this information to align with the request for error bars.
  3. [Section 3.5] The phrase "multi-model multi-task neural network" should likely read "multimodal multi-task neural network".
  4. [Section 4.4] The sentence "we use 18 task embeddings from PerAct [50] as the initial skill set" should clarify whether these are the same as the K=18 skill primitives in Table 6 and whether the z_k are frozen or updated during training; Section 7.3 says they are initialized from CLIP embeddings, but the learning objective in Eq. (2) does not explicitly describe their update rule.
  5. [Supplementary Figure 17] The x-axis label of the Bimanual Decomposability plot is not defined in the text, and the fitted logarithmic curve (y = 388.274 + 271.920 ln x) is presented without explaining which quantity is on each axis or how the four tasks were selected.
  6. [Introduction, reference [5]] Reference [5] (Baumgartner et al., "Lemma learning in the model evolution calculus") appears unrelated to the claim about shareable atomic movements and cooperation patterns; please verify the citation.

Circularity Check

1 steps flagged · score 2.0 of 10

No significant circularity; the only by-construction element is that the 'discovered' skills are initialized from the unimanual policy's own language templates, and that element is not load-bearing for the empirical bimanual results.

  1. renaming known result [Section 3.3 (Scheduling Unimanual Skill Primitives) and Section 7.3 / Fig. 12 (Skill Clustering)]
    "each potential skill is an implicit embedding zk ∈ RD, which can be initialized with the corresponding language template tokens of the pre-trained unimanual policy to mitigate the domain gap. ... The skill primitives use CLIP embeddings of unimanual language templates (e.g., 'open the drawer') from Table 1 of the supplementary file, as we expect them to represent foundational motions that generalize across object categories."

    The skill vocabulary is taken, by construction, from the pretrained unimanual policy's own language-template embeddings (K=18, matching the 18 pretraining tasks). Figure 12 then presents as 'discovered' the finding that learned skill representations cluster near those same unimanual task embeddings. This correspondence is guaranteed by the initialization choice rather than by independent discovery: the 'discovered skills' are the policy's input vocabulary renamed. However, this step is interpretive only; the bimanual success rates are measured on held-out episodes against external baselines and do not depend on the clustering interpretation, so the central empirical claim remains independent.

full rationale

The central claim that AnyBimanual transfers a pretrained unimanual policy to bimanual tasks with few demonstrations is supported by held-out success rates on 12 RLBench2 tasks and 9 real-world tasks, compared with external baselines such as PerAct2, PerAct-LF, and RVT-LF. The skill manager and visual aligner are trained with a behavior-cloning loss on the bimanual demonstrations; their outputs are not defined by the evaluation metric, so no fitted parameter is relabeled as a prediction. No load-bearing self-citation chain appears: the only author-overlapping reference (ManiGaussian) is cited as a previous RLBench-based manipulation method, not as the justification for AnyBimanual's design or its claimed superiority. The one mildly circular element is the skill-clustering interpretation: since the skill primitives are initialized from the pretrained unimanual policy's own language templates, the Fig. 12 observation that the learned skills are nearest to those same task embeddings is expected by construction. This is a renaming of the input vocabulary rather than an emergent discovery, but it does not force the bimanual success rates, which are empirically benchmarked. Overall, the derivation chain is self-contained with only a minor, non-load-bearing self-referential design choice.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities. Its skill primitives are a learned codebook parameterized by the unimanual policy's own language templates, and the soft masks are outputs of a learned module. The main assumptions are the bimanual decomposability of tasks, the expressiveness of the convex-hull skill representation, and the validity of the mutual exclusivity prior for visual alignment. The L1-sparsity-to-orthogonality assertion is unproven.

free parameters (3)
  • Number of skill primitives K = 18
    Set to the number of unimanual pretraining tasks; no sensitivity study reported (Section 7.2, Table 6).
  • Skill primitive embeddings z_k = 18 x 512 vectors
    Initialized with CLIP embeddings of unimanual language templates and refined on bimanual demonstrations (Section 7.3). These are the codebook the method composes; their initialization ties the transfer to the specific unimanual policy vocabulary.
  • Regularization weights lambda_skill, lambda_voxel = 0.0001, 0.001
    Chosen to keep auxiliary losses in the same magnitude as L_BC (Section 7.1); no tuning analysis provided.
assumptions (5)
  • domain assumption Bimanual tasks decompose into independent unimanual sub-tasks
    Skill manager Eq. 1 and Figure 3 assume a bimanual task can be expressed as per-arm linear combinations of unimanual skill embeddings; the failure mode Handover Easy at low success (Table 1: 31-44%) hints at the boundary.
  • domain assumption Softmax-weighted combination of 18 primitives plus bounded compensation can express the per-arm instruction embedding
    Eq. 1 restricts each arm's language embedding to the convex hull of the skill codebook plus epsilon; novel task instructions outside this hull are approximated only via the penalized residual.
  • domain assumption Mutually exclusive soft masks align bimanual voxel observations with the unimanual pretraining distribution
    Eq. 4 maximizes divergence between left and right masks without access to pretraining data; if the two arms must attend to the same region (e.g., handover), the prior may conflict with the task.
  • ad hoc to paper L1 sparsity on weights yields orthogonal and disjoint skill subspaces
    Section 3.3 asserts orthogonality follows from sparsity; no orthogonality constraint or proof is given, and L1 sparsity alone does not enforce orthogonality of the codebook vectors.
  • domain assumption The unimanual policy is pretrained on tasks semantically related to the bimanual tasks
    The 18 RLBench tasks in Table 4 include sweep to dustpan, push buttons, and put-in tasks, which overlap with RLBench2 bimanual tasks like sweep to dustpan, press buttons, put in fridge; transfer gains may partly reflect this overlap.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AnyBimanual: Transferring Unimanual Policy for General Bimanual Manipulation." pith.science (2026). https://pith.science/paper/QTFMWLVX

@misc{pith2026241206779,
  author       = {Pith},
  title        = {Pith review of: AnyBimanual: Transferring Unimanual Policy for General Bimanual Manipulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QTFMWLVX}},
  note         = {Machine review of arXiv:2412.06779}
}
read the original abstract

Performing general language-conditioned bimanual manipulation tasks is of great importance for many applications ranging from household service to industrial assembly. However, collecting bimanual manipulation data is expensive due to the high-dimensional action space, which poses challenges for conventional methods to handle general bimanual manipulation tasks. In contrast, unimanual policy has recently demonstrated impressive generalizability across a wide range of tasks because of scaled model parameters and training data, which can provide sharable manipulation knowledge for bimanual systems. To this end, we propose a plug-and-play method named AnyBimanual, which transfers pre-trained unimanual policy to general bimanual manipulation policy with few bimanual demonstrations. Specifically, we first introduce a skill manager to dynamically schedule the skill representations discovered from pre-trained unimanual policy for bimanual manipulation tasks, which linearly combines skill primitives with task-oriented compensation to represent the bimanual manipulation instruction. To mitigate the observation discrepancy between unimanual and bimanual systems, we present a visual aligner to generate soft masks for visual embedding of the workspace, which aims to align visual input of unimanual policy model for each arm with those during pretraining stage. AnyBimanual shows superiority on 12 simulated tasks from RLBench2 with a sizable 12.67% improvement in success rate over previous methods. Experiments on 9 real-world tasks further verify its practicality with an average success rate of 84.62%.

Figures

Figures reproduced from arXiv: 2412.06779 by the authors.

Figure 1
Figure 1. AnyBimanual enables plug-and-play transferring from pretrained unimanual policies to bimanual manipulation pol￾icy, which preserves the generalizability with the proposed skill scheduling framework. in household service [72], robotic surgery [33], and com￾ponent assembly in factories [9]. Compared to unimanual systems, bimanual systems enlarge the workspace and are able to handle more complex manipulation tasks by s… view at source ↗
Figure 2
Figure 2. The overall pipeline of AnyBimanual, which primarily consists of a skill manager and a perception manager. The skill manager adaptively coordinates primitive skills for each robot arm, while the perception manager mitigates the distributional shift from unimanual to bimanual by decomposing the 3D voxel observation for each arm. where intelligent agents acquire new abilities that are trans￾ferable across different ta… view at source ↗
Figure 3
Figure 3. Shareable skills across unimanual and bimanual settings. We observe that bimanual tasks are often originated from the combination of unimanual sub-tasks, which thus can be solved by effectively coordinating unimanual skills synchronously or asynchronously. combining the primitives from the skill set, the language embedding for the unimanual policy model can be repre￾sented as a linear combination of these primitives… view at source ↗
Figures from the paper (13 more)
Figure 5
Figure 5. Figure 5: Real-World Tasks. The real-world experiments are per￾formed in a tabletop setup with objects randomized in location ev￾ery episode. AnyBimanual can simultaneously conduct 9 complex real-world bimanual manipulation tasks with one model. Different colors mean different s…
Figure 6
Figure 6. Figure 6: Unseen Data Generalization. We include multiple distractors in real-world experiments, and find AnyBimanual generalizes to these settings successfully by unlocking the commonsense held by unimanual base models. creased input processing complexity, it leads to significa…
Figure 7
Figure 7. Figure 7: Visual Observation in RLBench2. We adopt 6 RGB-D cameras to cover the whole workspace [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Real-Robot Setup with one RealSense L515 RGB-D Camera and Two UR5e Manipulators. Tasksuite Descriptions. We provide a detailed descrip￾tion of the real-world tasks we used to evaluate our AnyBimanual in [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: Real-Robot Task Taxonomy. Following [42], we classify the reported 9 real-world tasks into 5 categories according to their collaboration patterns. Fold Clothes Play Ping Pong [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Real-world Keyframes. We manually select keyframes from the collected trajectories to simplify training. i-VisionGroup Perturbed right arm action Baseline Ours Perturbed left arm action Perturbed right arm action Perturbed left arm action [PITH_FULL_IMAGE:figures/ful…
Figure 11
Figure 11. Figure 11: SE(3) Augmentation. We slightly modify the SE(3) augmentation strategy to ensure that the perturbed ground-truth actions of both arms are reasonable. After resetting the manipulators to the home positions, we query the policy with the language instruction, initial obs…
Figure 12
Figure 12. Figure 12: Skill Clustering. We cluster the skill representations to further interpret the discovered skills of our AnyBimanual. to produce logits over 18 classes, and softmaxed to obtain probabilities. These probabilities are then used to recon￾struct the skill representation v…
Figure 13
Figure 13. Figure 13: Comparions of the rollouts of PerAct2 and our AnyBimanual. AnyBimanual demonstrates complex collaboration patterns by incorporating the skill manager and the visual aligner to schedule the manipulators correctly. Pick&Place Sync Rotate Toothbrush Pick&Place Async Coar…
Figure 14
Figure 14. Figure 14: Error Modes in Real-world Rollouts. We study the common failure cases of our AnyBimanual in real-world settings, where the incorrect end-effector actions are marked with red circles. We also provide the correct actions with green circles for reference. ables compact s…
Figure 15
Figure 15. Figure 15: Visualization of Skill Manager. 8.4. Error Modes To further study the limitations and risks of the pro￾posed method, we visualize the common failure cases in [PITH_FULL_IMAGE:figures/full_fig_p018_15.png]
Figure 17
Figure 17. Figure 17: Bimanual Decomposability [PITH_FULL_IMAGE:figures/full_fig_p018_17.png]
Figure 16
Figure 16. Figure 16: Visualization of Visual Aligner. 8.5. Bimanual Decomposability We complement a heuristic measurement to assess the feasi￾bility of decomposing a given bimanual task with a learned skill manager at the semantic level. Specifically, higher en￾tropy in the predicted comb…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Diffusion-Based Imaginative Coordination for Bimanual Manipulation

    cs.RO 2025-07 conditional novelty 6.0 of 10

    A diffusion-based policy that jointly predicts future video latents and actions improves bimanual manipulation success, with video prediction used only during training.

Reference graph

Works this paper leans on

76 extracted references · 49 canonical work pages · cited by 1 Pith paper

  1. [1]

    Autort: Embodied foundation models for large scale orchestration of robotic agents

    Michael Ahn, Debidatta Dwibedi, Chelsea Finn, Montse Gonzalez Arenas, Keerthana Gopalakrishnan, Karol Hausman, Brian Ichter, Alex Irpan, Nikhil Joshi, Ryan Julian, et al. Autort: Embodied foundation models for large scale orchestration of robotic agents. arXiv preprint arXiv:2401.12963, 2024. 2

    arXiv 2024

  2. [2]

    Exploiting symmetries in reinforcement learning of bimanual robotic tasks

    Fabio Amadio, Adri `a Colom´e, and Carme Torras. Exploiting symmetries in reinforcement learning of bimanual robotic tasks. IEEE Robotics and Automation Letters (RAL) , 4(2): 1838–1845, 2019. 3

    work page 2019

  3. [3]

    Speedfolding: Learning effi- cient bimanual folding of garments

    Yahav Avigal, Lars Berscheid, Tamim Asfour, Torsten Kr¨oger, and Ken Goldberg. Speedfolding: Learning effi- cient bimanual folding of garments. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 1–8. IEEE, 2022

    work page 2022

  4. [4]

    Compliant movement primitives in a bimanual setting

    Aleksandar Batinica, Bojan Nemec, Ale ˇs Ude, Mirko Rakovi´c, and Andrej Gams. Compliant movement primitives in a bimanual setting. In IEEE-RAS International Confer- ence on Humanoid Robotics (Humanoids) , pages 365–371,

    work page

  5. [5]

    Lemma learning in the model evolution calculus

    Peter Baumgartner, Alexander Fuchs, and Cesare Tinelli. Lemma learning in the model evolution calculus. In Interna- tional Conference on Logic for Programming Artificial Intel- ligence and Reasoning (LPAR), pages 572–586, 2006. 2

    work page 2006

  6. [6]

    Bi- manual robotic cloth manipulation for laundry folding

    Christian Bersch, Benjamin Pitzer, and S ¨oren Kammel. Bi- manual robotic cloth manipulation for laundry folding. In Proceedings of Robotics: Science and Systems (RSS) , pages 1413–1419, 2011. 2

    work page 2011

  7. [7]

    Rt-1: Robotics transformer for real-world control at scale

    Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, Joseph Dabis, Chelsea Finn, Keerthana Gopalakr- ishnan, Karol Hausman, Alex Herzog, Jasmine Hsu, et al. Rt-1: Robotics transformer for real-world control at scale. In Proceedings of Robotics: Science and Systems (RSS) , 2023. 2, 4

    work page 2023

  8. [8]

    Rt-2: Vision-language-action models trans- fer web knowledge to robotic control

    Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, et al. Rt-2: Vision-language-action models trans- fer web knowledge to robotic control. In Conference on Robot Learning (CoRL), pages 2165–2183, 2023. 2, 4

    work page 2023

Show all 76 references
  1. [9]

    A dual-arm collaborative robot system for the smart factories of the future

    Jens F Buhl, Rune Grønhøj, Jan K Jørgensen, Guilherme Mateus, Daniela Pinto, Jacob K Sørensen, Simon Bøgh, and Dimitrios Chrysostomou. A dual-arm collaborative robot system for the smart factories of the future. Procedia manu- facturing, 38:333–340, 2019. 1, 2

    2019

  2. [10]

    Pali-3 vision language models: Smaller, faster, stronger

    Xi Chen, Xiao Wang, Lucas Beyer, Alexander Kolesnikov, Jialin Wu, Paul V oigtlaender, Basil Mustafa, Sebastian Goodman, Ibrahim Alabdulmohsin, Piotr Padlewski, et al. Pali-3 vision language models: Smaller, faster, stronger. arXiv preprint arXiv:2310.09199, 2023. 2

    arXiv

  3. [11]

    Bi-dexhands: Towards human-level bimanual dex- terous manipulation

    Yuanpei Chen, Yiran Geng, Fangwei Zhong, Jiaming Ji, Jiechuang Jiang, Zongqing Lu, Hao Dong, and Yaodong Yang. Bi-dexhands: Towards human-level bimanual dex- terous manipulation. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence (T-PAMI) , 46(5):2804–2818,

  4. [12]

    Open-television: Teleoperation with immersive active visual feedback

    Xuxin Cheng, Jialong Li, Shiqi Yang, Ge Yang, and Xiao- long Wang. Open-television: Teleoperation with immersive active visual feedback. arXiv preprint arXiv:2407.01512 ,

    arXiv

  5. [13]

    Universal manipulation interface: In-the-wild robot teaching without in-the-wild robots

    Cheng Chi, Zhenjia Xu, Chuer Pan, Eric Cousineau, Ben- jamin Burchfiel, Siyuan Feng, Russ Tedrake, and Shu- ran Song. Universal manipulation interface: In-the-wild robot teaching without in-the-wild robots. arXiv preprint arXiv:2402.10329, 2024. 2

    2024 arXiv

  6. [14]

    Intrinsic motivation for encouraging synergistic behavior

    Rohan Chitnis, Shubham Tulsiani, Saurabh Gupta, and Ab- hinav Gupta. Intrinsic motivation for encouraging synergistic behavior. In Proceedings of the International Conference on Learning Representations (ICLR), 2020. 3

    2020

  7. [15]

    Efficient bimanual manipulation using learned task schemas

    Rohan Chitnis, Shubham Tulsiani, Saurabh Gupta, and Ab- hinav Gupta. Efficient bimanual manipulation using learned task schemas. In IEEE International Conference on Robotics and Automation (ICRA), pages 1149–1155, 2020. 2, 3

    2020

  8. [16]

    Active vision might be all you need: Exploring active vision in bimanual robotic manipulation

    Ian Chuang, Andrew Lee, Dechen Gao, and Iman Soltani. Active vision might be all you need: Exploring active vision in bimanual robotic manipulation. arXiv preprint arXiv:2409.17435, 2024. 2

    2024 arXiv

  9. [17]

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

    Embodiment Collaboration and Abby O’Neill et al. Open x-embodiment: Robotic learning datasets and rt-x models. In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) , pages 6892–6903, 2024. 2, 4

    2024

  10. [18]

    Bunny-visionpro: Real-time bimanual dexterous teleopera- tion for imitation learning.arXiv preprint arXiv:2407.03162,

    Runyu Ding, Yuzhe Qin, Jiyue Zhu, Chengzhe Jia, Shiqi Yang, Ruihan Yang, Xiaojuan Qi, and Xiaolong Wang. Bunny-visionpro: Real-time bimanual dexterous teleopera- tion for imitation learning.arXiv preprint arXiv:2407.03162,

    arXiv

  11. [19]

    Interactive imitation learn- ing of bimanual movement primitives

    Giovanni Franzese, Leandro de Souza Rosa, Tim Verburg, Luka Peternel, and Jens Kober. Interactive imitation learn- ing of bimanual movement primitives. IEEE/ASME Trans- actions on Mechatronics (T-Mech), 2023. 3

    2023

  12. [20]

    Drive like a human: Rethink- ing autonomous driving with large language models

    Daocheng Fu, Xin Li, Licheng Wen, Min Dou, Pinlong Cai, Botian Shi, and Yu Qiao. Drive like a human: Rethink- ing autonomous driving with large language models. In IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 910–919, 2024. 3

    2024

  13. [21]

    Safe self-supervised learning in real of visuo-tactile feedback policies for industrial inser- tion

    Letian Fu, Huang Huang, Lars Berscheid, Hui Li, Ken Gold- berg, and Sachin Chitta. Safe self-supervised learning in real of visuo-tactile feedback policies for industrial inser- tion. In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) , pages...

  14. [22]

    Mobile aloha: Learning bimanual mobile manipulation with low-cost whole-body teleoperation

    Zipeng Fu, Tony Z Zhao, and Chelsea Finn. Mobile aloha: Learning bimanual mobile manipulation with low-cost whole-body teleoperation. arXiv preprint arXiv:2401.02117,

    arXiv

  15. [23]

    Learning dense visual correspondences in simulation to smooth and fold real fabrics

    Aditya Ganapathi, Priya Sundaresan, Brijen Thananjeyan, Ashwin Balakrishna, Daniel Seita, Jennifer Grannen, Minho Hwang, Ryan Hoque, Joseph E Gonzalez, Nawid Jamali, et al. Learning dense visual correspondences in simulation to smooth and fold real fabrics. In Proceedings of t...

    2021

  16. [24]

    Bi-kvil: Keypoints-based visual 9 imitation learning of bimanual manipulation tasks

    Jianfeng Gao, Xiaoshu Jin, Franziska Krebs, No ´emie Jaquier, and Tamim Asfour. Bi-kvil: Keypoints-based visual 9 imitation learning of bimanual manipulation tasks. In Pro- ceedings of the IEEE International Conference on Robotics and Automation (ICRA), pages 16850–16857, 2024. 2

    2024

  17. [26]

    Dag-plan: Generating directed acyclic dependency graphs for dual-arm cooperative plan- ning

    Zeyu Gao, Yao Mu, Jinye Qu, Mengkang Hu, Lingyue Guo, Ping Luo, and Yanfeng Lu. Dag-plan: Generating directed acyclic dependency graphs for dual-arm cooperative plan- ning. arXiv preprint arXiv:2406.09953, 2024. 2

    2024 arXiv

  18. [27]

    Bi-vla: Vision-language-action model-based sys- tem for bimanual robotic dexterous manipulations

    Koffivi Fid `ele Gbagbe, Miguel Altamirano Cabrera, Ali Al- abbas, Oussama Alyunes, Artem Lykov, and Dzmitry Tset- serukou. Bi-vla: Vision-language-action model-based sys- tem for bimanual robotic dexterous manipulations. arXiv preprint arXiv:2405.06039, 2024. 2

    2024 arXiv

  19. [28]

    Rvt: Robotic view transformer for 3d ob- ject manipulation

    Ankit Goyal, Jie Xu, Yijie Guo, Valts Blukis, Yu-Wei Chao, and Dieter Fox. Rvt: Robotic view transformer for 3d ob- ject manipulation. In Conference on Robot Learning (CoRL), pages 694–710, 2023. 6

    2023

  20. [29]

    Un- tangling dense knots by learning task-relevant keypoints

    Jennifer Grannen, Priya Sundaresan, Brijen Thananjeyan, Jeffrey Ichnowski, Ashwin Balakrishna, Vainavi Viswanath, Michael Laskey, Joseph Gonzalez, and Ken Goldberg. Un- tangling dense knots by learning task-relevant keypoints. In Conference on Robot Learning (CoRL) , pages 782–800,

  21. [30]

    Stabilize to act: Learning to coordinate for bimanual manip- ulation

    Jennifer Grannen, Yilin Wu, Brandon Vu, and Dorsa Sadigh. Stabilize to act: Learning to coordinate for bimanual manip- ulation. In Conference on Robot Learning (CoRL) , pages 563–576, 2023. 1, 2

    2023

  22. [31]

    Peract2: Benchmarking and learning for robotic bimanual manipulation tasks

    Markus Grotz, Mohit Shridhar, Yu-Wei Chao, Tamim As- four, and Dieter Fox. Peract2: Benchmarking and learning for robotic bimanual manipulation tasks. In Conference on Robot Learning (CoRL), 2024. 1, 2, 3, 6

    2024

  23. [32]

    Flingbot: The unreasonable ef- fectiveness of dynamic manipulation for cloth unfolding

    Huy Ha and Shuran Song. Flingbot: The unreasonable ef- fectiveness of dynamic manipulation for cloth unfolding. In Conference on Robot Learning (CoRL), pages 24–33, 2021. 3

    2021

  24. [33]

    Towards human-robot collaborative surgery: Trajectory and strategy learning in bimanual peg transfer

    Zhaoyang Jacopo Hu, Ziwei Wang, Yanpei Huang, Aran Sena, Ferdinando Rodriguez y Baena, and Etienne Burdet. Towards human-robot collaborative surgery: Trajectory and strategy learning in bimanual peg transfer. IEEE Robotics and Automation Letters (RAL), 8(8):4553–4560, 2023. 1, 2

    2023

  25. [34]

    V oxposer: Composable 3d value maps for robotic manipulation with language models

    Wenlong Huang, Chen Wang, Ruohan Zhang, Yunzhu Li, Jiajun Wu, and Li Fei-Fei. V oxposer: Composable 3d value maps for robotic manipulation with language models. In Conference on Robot Learning (CoRL) , pages 540–562,

  26. [35]

    Perceiver: General perception with iterative attention

    Andrew Jaegle, Felix Gimeno, Andy Brock, Oriol Vinyals, Andrew Zisserman, and Joao Carreira. Perceiver: General perception with iterative attention. In Proceedings of the In- ternational Conference on Machine Learning (ICML), pages 4651–4664, 2021. 6

    2021

  27. [36]

    Rlbench: The robot learning benchmark & learning environment

    Stephen James, Zicong Ma, David Rovick Arrojo, and An- drew J Davison. Rlbench: The robot learning benchmark & learning environment. IEEE Robotics and Automation Let- ters (RAL), 5(2):3019–3026, 2020. 6, 1

    2020

  28. [37]

    Copal: corrective planning of robot actions with large lan- guage models

    Frank Joublin, Antonello Ceravola, Pavel Smirnov, Felix Ocker, Joerg Deigmoeller, Anna Belardinelli, Chao Wang, Stephan Hasler, Daniel Tanneberg, and Michael Gienger. Copal: corrective planning of robot actions with large lan- guage models. In Proceedings of the IEEE Internati...

    2024

  29. [38]

    Bi-manual manipulation and attachment via sim-to-real reinforcement learning

    Satoshi Kataoka, Seyed Kamyar Seyed Ghasemipour, Daniel Freeman, and Igor Mordatch. Bi-manual manipulation and attachment via sim-to-real reinforcement learning. arXiv preprint arXiv:2203.08277, 2022. 2

    2022 arXiv

  30. [39]

    3d diffuser actor: Policy diffusion with 3d scene rep- resentations

    Tsung-Wei Ke, Nikolaos Gkanatsios, and Katerina Fragki- adaki. 3d diffuser actor: Policy diffusion with 3d scene rep- resentations. arXiv preprint arXiv:2402.10885, 2024. 2, 4, 5, 6

    2024 arXiv

  31. [40]

    Droid: A large-scale in-the-wild robot manipulation dataset

    Alexander Khazatsky, Karl Pertsch, Suraj Nair, Ash- win Balakrishna, Sudeep Dasari, Siddharth Karamcheti, Soroush Nasiriany, Mohan Kumar Srirama, Lawrence Yun- liang Chen, Kirsty Ellis, et al. Droid: A large-scale in-the-wild robot manipulation dataset. arXiv preprint arXiv:24...

    2024 arXiv

  32. [41]

    Openvla: An open-source vision-language-action model

    Moo Jin Kim, Karl Pertsch, Siddharth Karamcheti, Ted Xiao, Ashwin Balakrishna, Suraj Nair, Rafael Rafailov, Ethan Foster, Grace Lam, Pannag Sanketi, et al. Openvla: An open-source vision-language-action model. arXiv preprint arXiv:2406.09246, 2024. 2, 4, 5, 8

    2024 arXiv

  33. [42]

    A bimanual manip- ulation taxonomy

    Franziska Krebs and Tamim Asfour. A bimanual manip- ulation taxonomy. IEEE Robotics and Automation Letters (RAL), 7(4):11031–11038, 2022. 8, 1, 3

    2022

  34. [43]

    Skilldiffuser: Interpretable hi- erarchical planning via skill abstractions in diffusion-based task execution

    Zhixuan Liang, Yao Mu, Hengbo Ma, Masayoshi Tomizuka, Mingyu Ding, and Ping Luo. Skilldiffuser: Interpretable hi- erarchical planning via skill abstractions in diffusion-based task execution. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVP...

    2024

  35. [44]

    Llm+ p: Empower- ing large language models with optimal planning proficiency

    Bo Liu, Yuqian Jiang, Xiaohan Zhang, Qiang Liu, Shiqi Zhang, Joydeep Biswas, and Peter Stone. Llm+ p: Empower- ing large language models with optimal planning proficiency. arXiv preprint arXiv:2304.11477, 2023. 2

    2023 arXiv

  36. [45]

    V oxact-b: V oxel-based acting and stabi- lizing policy for bimanual manipulation

    I Liu, Chun Arthur, Sicheng He, Daniel Seita, and Gau- rav Sukhatme. V oxact-b: V oxel-based acting and stabi- lizing policy for bimanual manipulation. arXiv preprint arXiv:2407.04152, 2024. 1, 2

    2024 arXiv

  37. [46]

    Rdt-1b: a diffusion foundation model for bimanual manipu- lation

    Songming Liu, Lingxuan Wu, Bangguo Li, Hengkai Tan, Huayu Chen, Zhengyi Wang, Ke Xu, Hang Su, and Jun Zhu. Rdt-1b: a diffusion foundation model for bimanual manipu- lation. arXiv preprint arXiv:2410.07864, 2024. 2

    2024 arXiv

  38. [47]

    Manigaussian: Dynamic gaus- sian splatting for multi-task robotic manipulation

    Guanxing Lu, Shiyi Zhang, Ziwei Wang, Changliu Liu, Ji- wen Lu, and Yansong Tang. Manigaussian: Dynamic gaus- sian splatting for multi-task robotic manipulation. In Euro- pean Conference on Computer Vision (ECCV) , pages 349– 366, 2025. 6

    2025

  39. [48]

    Maestrob: A robotics frame- work for integrated orchestration of low-level control and high-level reasoning

    Asim Munawar, Giovanni De Magistris, Tu-Hoa Pham, Daiki Kimura, Michiaki Tatsubori, Takao Moriyama, Ryuki Tachibana, and Grady Booch. Maestrob: A robotics frame- work for integrated orchestration of low-level control and high-level reasoning. In Proceedings of the IEEE Inter- ...

    2018

  40. [49]

    Learn- ing transferable visual models from natural language super- vision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learn- ing transferable visual models from natural language super- vision. In Proceedings of the International Conference ...

    2021

  41. [50]

    Perceiver- actor: A multi-task transformer for robotic manipulation

    Mohit Shridhar, Lucas Manuelli, and Dieter Fox. Perceiver- actor: A multi-task transformer for robotic manipulation. In Conference on Robot Learning (CoRL) , pages 785–799,

  42. [51]

    Octo: An open-source generalist robot policy

    Octo Model Team, Dibya Ghosh, Homer Walke, Karl Pertsch, Kevin Black, Oier Mees, Sudeep Dasari, Joey Hejna, Tobias Kreiman, Charles Xu, et al. Octo: An open-source generalist robot policy. arXiv preprint arXiv:2405.12213, 2024. 2, 4, 5

    2024 arXiv

  43. [52]

    Regression shrinkage and selection via the lasso

    Robert Tibshirani. Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society Series B: Statistical Methodology, 58(1):267–288, 1996. 5

    1996

  44. [53]

    Llama 2: Open foundation and fine-tuned chat models.arXiv preprint arXiv:2307.09288, 2023

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. Llama 2: Open foundation and fine-tuned chat models.arXiv preprint arXiv:2307.09288, 2023. 2

    2023 arXiv

  45. [54]

    Embodied ai with two arms: Zero-shot learning, safety and modularity

    Jake Varley, Sumeet Singh, Deepali Jain, Krzysztof Choro- manski, Andy Zeng, Somnath Basu Roy Chowdhury, Avinava Dubey, and Vikas Sindhwani. Embodied ai with two arms: Zero-shot learning, safety and modularity. arXiv preprint arXiv:2404.03570, 2024. 2

    2024 arXiv

  46. [55]

    Bridgedata v2: A dataset for robot learning at scale

    Homer Rich Walke, Kevin Black, Tony Z Zhao, Quan Vuong, Chongyi Zheng, Philippe Hansen-Estruch, An- dre Wang He, Vivek Myers, Moo Jin Kim, Max Du, et al. Bridgedata v2: A dataset for robot learning at scale. In Con- ference on Robot Learning (CoRL), pages 1723–1736, 2023. 2

    2023

  47. [56]

    Dexcap: Scalable and portable mocap data collection system for dexterous manipulation

    Chen Wang, Haochen Shi, Weizhuo Wang, Ruohan Zhang, Li Fei-Fei, and C Karen Liu. Dexcap: Scalable and portable mocap data collection system for dexterous manipulation. arXiv preprint arXiv:2403.07788, 2024. 2

    2024 arXiv

  48. [57]

    V oyager: An open-ended embodied agent with large language models

    Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, and Anima Anandku- mar. V oyager: An open-ended embodied agent with large language models. Transactions on Machine Learning Re- search (TMLR), 2024, 2024. 3

    2024

  49. [58]

    Scaling proprioceptive-visual learning with heterogeneous pre-trained transformers

    Lirui Wang, Xinlei Chen, Jialiang Zhao, and Kaiming He. Scaling proprioceptive-visual learning with heterogeneous pre-trained transformers. arXiv preprint arXiv:2409.20537,

    arXiv

  50. [59]

    Imagen editor and editbench: Advancing and evaluating text-guided image inpainting

    Su Wang, Chitwan Saharia, Ceslee Montgomery, Jordi Pont- Tuset, Shai Noy, Stefano Pellegrini, Yasumasa Onoe, Sarah Laszlo, David J Fleet, Radu Soricut, et al. Imagen editor and editbench: Advancing and evaluating text-guided image inpainting. In Proceedings of the IEEE Confere...

    2023

  51. [60]

    Influencing towards stable multi-agent inter- actions

    Woodrow Zhouyuan Wang, Andy Shih, Annie Xie, and Dorsa Sadigh. Influencing towards stable multi-agent inter- actions. In Conference on Robot Learning (CoRL) , pages 1132–1143, 2022. 2

    2022

  52. [61]

    Sparse representation for computer vision and pattern recognition

    John Wright, Yi Ma, Julien Mairal, Guillermo Sapiro, Thomas S Huang, and Shuicheng Yan. Sparse representation for computer vision and pattern recognition. Proceedings of the IEEE, 98(6):1031–1044, 2010. 5

    2010

  53. [62]

    Gello: A general, low-cost, and intuitive tele- operation framework for robot manipulators

    Philipp Wu, Yide Shentu, Zhongke Yi, Xingyu Lin, and Pieter Abbeel. Gello: A general, low-cost, and intuitive tele- operation framework for robot manipulators. arXiv preprint arXiv:2309.13037, 2023. 2

    2023 arXiv

  54. [63]

    Chaineddiffuser: Unify- ing trajectory diffusion and keypose prediction for robotic manipulation

    Zhou Xian, Nikolaos Gkanatsios, Theophile Gervet, Tsung- Wei Ke, and Katerina Fragkiadaki. Chaineddiffuser: Unify- ing trajectory diffusion and keypose prediction for robotic manipulation. In Conference on Robot Learning (CoRL) , pages 2323–2339, 2023. 6

    2023

  55. [64]

    Clara De Paolis Kaluza, Linfeng Zhao, Lawson L

    Fan Xie, Alexander Chowdhury, M. Clara De Paolis Kaluza, Linfeng Zhao, Lawson L. S. Wong, and Rose Yu. Deep im- itation learning for bimanual robotic manipulation. In Ad- vances in Neural Information Processing Systems (NeurIPS),

  56. [65]

    Ace: A cross-platform visual-exoskeletons sys- tem for low-cost dexterous teleoperation

    Shiqi Yang, Minghuan Liu, Yuzhe Qin, Runyu Ding, Jia- long Li, Xuxin Cheng, Ruihan Yang, Sha Yi, and Xiao- long Wang. Ace: A cross-platform visual-exoskeletons sys- tem for low-cost dexterous teleoperation. arXiv preprint arXiv:2408.11805, 2024. 2

    2024 arXiv

  57. [66]

    Learning nonlinear dynam- ical system for movement primitives

    Xiaochuan Yin and Qijun Chen. Learning nonlinear dynam- ical system for movement primitives. In IEEE International Conference on Systems, Man, and Cybernetics (SMC), pages 3761–3766, 2014. 3

    2014

  58. [67]

    Reddi, Jonathan Hseu, Sanjiv Kumar, Srinadh Bhojanapalli, Xiaodan Song, James Dem- mel, Kurt Keutzer, and Cho-Jui Hsieh

    Yang You, Jing Li, Sashank J. Reddi, Jonathan Hseu, Sanjiv Kumar, Srinadh Bhojanapalli, Xiaodan Song, James Dem- mel, Kurt Keutzer, and Cho-Jui Hsieh. Large batch opti- mization for deep learning: Training BERT in 76 minutes. In Proceedings of the International Conference on L...

    2020

  59. [68]

    Bikc: Keypose-conditioned consistency policy for biman- ual robotic manipulation

    Dongjie Yu, Hang Xu, Yizhou Chen, Yi Ren, and Jia Pan. Bikc: Keypose-conditioned consistency policy for biman- ual robotic manipulation. arXiv preprint arXiv:2406.10093,

    arXiv

  60. [69]

    Gnfactor: Multi-task real robot learning with generalizable neural feature fields

    Yanjie Ze, Ge Yan, Yueh-Hua Wu, Annabella Macaluso, Yuying Ge, Jianglong Ye, Nicklas Hansen, Li Erran Li, and Xiaolong Wang. Gnfactor: Multi-task real robot learning with generalizable neural feature fields. In Conference on Robot Learning (CoRL), pages 284–301, 2023. 6

    2023

  61. [70]

    Jesse Zhang, Jiahui Zhang, Karl Pertsch, Ziyi Liu, Xiang Ren, Minsuk Chang, Shao-Hua Sun, and Joseph J. Lim. Bootstrap your own skills: Learning to solve new tasks with large language model guidance. In Conference on Robot Learning (CoRL), pages 302–325. PMLR, 2023. 2

    2023

  62. [71]

    Dair: Disentangled attention intrinsic regu- larization for safe and efficient bimanual manipulation.arXiv preprint arXiv:2106.05907, 2021

    Minghao Zhang, Pingcheng Jian, Yi Wu, Huazhe Xu, and Xiaolong Wang. Dair: Disentangled attention intrinsic regu- larization for safe and efficient bimanual manipulation.arXiv preprint arXiv:2106.05907, 2021. 2

    2021 arXiv

  63. [72]

    Empowering embodied manipulation: A bimanual-mobile robot manipulation dataset for household tasks

    Tianle Zhang, Dongjiang Li, Yihang Li, Zecui Zeng, Lin Zhao, Lei Sun, Yue Chen, Xuelong Wei, Yibing Zhan, Lu- song Li, et al. Empowering embodied manipulation: A bimanual-mobile robot manipulation dataset for household tasks. arXiv preprint arXiv:2405.18860, 2024. 1, 2 11

    2024 arXiv

  64. [73]

    Zhao, Vikash Kumar, Sergey Levine, and Chelsea Finn

    Tony Z. Zhao, Vikash Kumar, Sergey Levine, and Chelsea Finn. Learning fine-grained bimanual manipulation with low-cost hardware. In Proceedings of Robotics: Science and Systems (RSS), 2023. 2 12 AnyBimanual: Transferring Unimanual Policy for General Bimanual Manipulation Suppl...

    2023

  65. [74]

    open the drawer

    Additional Experimental Details 6.1. Simulation We utilize RLBench2 [31] as our main simulated task suite for mulit-task learning. Table 5 is an overview of the 12 selected tasks we use in the experiments. Table 4 is an overview of the 18 selected tasks from RLBench [36] used ...

  66. [75]

    Hyperparameters The hyperparameters used in AnyBimanual are shown in Table 6

    Additional Implementation Details 7.1. Hyperparameters The hyperparameters used in AnyBimanual are shown in Table 6. To ensure that the auxiliary objectives Lskill and Lvoxel remain in the same magnitude as LBC, we set λskill = 0.0001 and λvoxel = 0.001. Other hyperparameters ...

  67. [76]

    Push the box to the red area

    Additional Experimental Results 8.1. Skill Clustering The goal of the skill manager is to obtain the language embedding via linear skill representations, so that the pre-trained unimanual policy model ( i.e., PerAct) can be prompted to generate feasible manipulation actions. I...

  68. [77]

    Discussions on Limitations Cross-embodiment Skill Transferring. AnyBimanual needs careful visual alignment between unimanual and bi- manual systems to ensure robust policy transfer, which means it is not designed for cross-embodiment skill trans- ferring. Even though the 7-DoF...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.