Robot Drummer: Learning Rhythmic Skills for Humanoid Drumming

Asad Ali Shahid; Francesco Braghin; Loris Roveda

REVIEW 3 major objections 4 minor 20 references

Robot Drummer: Learning Rhythmic Skills for Humanoid Drumming

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper argues that a full drum track can be learned as a Rhythmic Contact Chain and that a single policy, trained on fixed-length segments with dense contact rewards, plays over thirty songs with F1 mostly above 0.9.

desk verdict First RL-trained humanoid drummer with a clean formulation and honest ablations; the F1 metric needs to be specified before the headline numbers can be trusted. read the letter →

arxiv 2507.11498 v3 pith:2KDWBLYO submitted 2025-07-15 cs.RO

Asad Ali Shahid , Francesco Braghin , Loris Roveda This is my paper

classification cs.RO

keywords humanoidcontrolroboticmusicianshipreinforcementlearningrhythmiccontactchaindrummingcontact-richmanipulationtemporaldecompositionF1score

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

This paper argues that reinforcement learning can take humanoid robots into expressive, process-driven tasks, and uses drumming as the testbed. Its central claim is that a full drum track can be encoded as a Rhythmic Contact Chain---a sequence of timed contact events specifying which drums to strike and when---and that a single policy trained on fixed-length segments can execute that chain over songs several minutes long. Across more than thirty popular rock, metal, and jazz tracks, the trained specialist policies reach mean F1 scores mostly above 0.9, including on demanding drum fills and snare rolls. The learned policies also show strategies the reward never encoded explicitly, such as crossing arms to reach distant drums and reassigning sticks depending on which hit comes next. A sympathetic reader would care because drumming demands split-second timing and sustained multi-limb coordination over a long horizon, the very properties that make process-driven skills hard for current robot learning methods.

What carries the argument

The central object is the Rhythmic Contact Chain (RCC): a sequence of contact steps $C = \{C_1, \ldots, C_N\}$, each step $C_i = (D_i, t_i, S_i)$ specifying the subset of drums to strike at a given time and the sticks allowed to strike them. It converts a MIDI drum track into spatiotemporal targets without prescribing any reference trajectory. Two mechanisms make the chain learnable: the policy observes the next $L$ one-hot contact goals as a discrete phase marker, telling it which drum is coming next, and a dense contact reward scores correct, wrong, and missed strikes plus a proximity bonus at every timestep. Temporal decomposition then divides the chain into fixed-length segments that are trained in parallel by a single policy, turning a several-minute sparse-reward task into sample-efficient reinforcement learning. The RCC carries the argument because it turns aesthetic timing into a dense contact-prediction and contact-execution problem.

What would settle it

Record the actual contact timestamps of a trained policy on a full song, define a hit as correctly matched to a target only when it falls within a stated tolerance window (for example, plus or minus 20 ms), and recompute F1; if the scores drop well below the reported values, the paper's central performance claim depends on the undefined hit-matching rule rather than on genuine rhythmic precision.

Watch

Extended reading notes

Core claim

The authors' claim is that a musical performance is, for a robot, a chain of contact-fulfillment events, and that this chain is learnable. They formalize the Rhythmic Contact Chain as $C = \{C_1, \ldots, C_N\}$, with each contact step $C_i = (D_i, t_i, S_i)$ marking which drum(s) must be hit at time $t_i$ and which stick(s) may be used; the chain is converted into one-hot vectors that serve as lookahead goals in the policy's observation. The reward scores every contact event at each control step, rewarding correct strikes, penalizing wrong and missed strikes, and adding a proximity bonus, while temporal decomposition splits each song into fixed-length segments trained in parallel. With this setup, a single specialist policy reproduces hundreds of precisely timed hits per song, scoring F1 mostly above 0.9 across over thirty tracks. The policies also develop emergent human-like drumming strategies, including cross-over strikes and adaptive stick assignment, that were not hand-authored. The paper presents this as evidence that reinforcement learning can bring humanoid robots into creative musical performance.

Load-bearing premise

The reported F1 scores count a strike as correct only if it lands on the exact 50 Hz timestep of the intended target, and the paper never defines how a physical hit is matched to a timestep, so the high scores implicitly assert sub-20 ms precision without demonstrating it.

Editorial extensions

If this is right

Specialist policies achieve near-perfect rhythmic fidelity on many full songs, with F1 scores above 0.9 on most of the more than thirty tested tracks.
Songs with highly irregular hit timing (high nPVI) or many different drums are the hardest, and polyphony places an upper bound on F1 because the robot can hit at most two drums at once.
A single generalist policy trained across multiple songs performs markedly worse than song-specific specialists, showing negative transfer in multi-song rhythmic control.
Removing any dense-contact reward component or the one-hot contact-goal observations degrades F1 substantially, and contact goals cannot be replaced by a continuous time-phase variable.
Temporal decomposition does not change final F1 but cuts training time from about 8-9 hours to 2-3 hours on the same hardware.

Reading between the lines

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

Beyond the paper: if the F1 scores are robust to a defined hit-matching tolerance, the Rhythmic Contact Chain formulation should transfer to other percussion instruments and to any contact-timing task where the goal is a temporal sequence of impacts rather than a final pose.
Beyond the paper: the strong nPVI-to-F1 correlation suggests a testable curriculum---training segments ordered from low to high rhythmic irregularity---that might reduce the specialist-to-generalist performance drop the paper observes.
Beyond the paper: because the paper never defines hit matching, a direct check is to recompute F1 with a one-timestep tolerance and with a 20 ms tolerance; the comparison would show whether the reported high scores reflect true rhythmic precision or a permissive scoring rule.

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.

Desk Editor's Note

read the letter

Colleague,

Here's my read on the Robot Drummer paper. The headline: it's the first RL-trained full humanoid drummer, and the core formulation—drumming as a sequence of timed contact events (RCC)—is clean and works. The paper is a solid empirical study, not a toy demonstration.

What I like: the ablations are genuinely informative. Removing wrong/missed strike penalties drops F1 substantially, and removing the one-hot contact goals collapses performance. The temporal decomposition is shown to cut training time from ~8 hours to ~2-3 hours with no loss in F1, which is a real practical contribution. The complexity analysis (nPVI, polyphony, drum count) matches intuition and gives the field a useful way to characterize drumming difficulty. The listener study is a nice addition; the low naturalness score (2.46) is honestly reported.

The main soft spot is the evaluation metric. The paper says F1 compares ground-truth RCC to 'the robot's actual hits at each timestep,' with no tolerance or matching rule. At 50 Hz, that implies a 20 ms window, which is extremely strict. The contact detection threshold in Isaac Gym is also unreported. So the headline F1 values are not portable—another group cannot reproduce them without the exact evaluation script. This is the load-bearing weakness: the paper's claim of split-second timing rests on a metric that is not fully specified. That said, I don't think there's a logical error. The reward uses the same per-timestep matching, so the policy is trained toward the same objective. The numbers are internally consistent. It's a reporting gap, not a fatal flaw.

Other smaller issues: segment length P, lookahead L, PD gains, and action scaling are unlisted, which makes replication harder. The 'emergent human-like strategies' claim is qualitative and the listener study says otherwise on naturalness. But those are minor.

Bottom line: this is a worthwhile contribution to robot musicianship and contact-rich RL. It deserves a serious referee, but the metric definition and code release need to be fixed before it can be fully trusted. I'd accept for review.

Referee Report

3 major / 4 minor

Summary. The paper proposes a simulation framework, Robot Drummer, in which a Unitree G1 humanoid learns to play MIDI-derived drum tracks by treating drumming as a sequence of timed contact events, the Rhythmic Contact Chain (RCC). The RCC is segmented into fixed-length intervals, and a single PPO policy is trained across all segments in parallel with dense contact rewards (correct, wrong, missed, proximity) plus a lookahead of future contact goals. The authors report specialist policies achieving mean F1 scores mostly above 0.9 across 32 rock and metal songs, present ablations supporting each reward term and the contact-goal observation, show that a single generalist policy over multiple songs degrades performance, and include a small listener study. The overarching claim is that RL can produce temporally precise, expressive humanoid drumming in simulation.

Significance. If the reported F1 scores are trustworthy, the paper advances process-driven whole-body control: it demonstrates that a single policy can sustain long-horizon rhythmic contact sequences, with emergent behaviors such as cross-arm strikes and adaptive stick assignment. The strengths include a clear contact-chain formulation, a sensible temporal decomposition, hypothesis-driven song-complexity analysis using nPVI and polyphony, and ablations reported as mean ± std over five seeds. The main weakness is that the central metric — F1 against the RCC — is not defined precisely enough to support the claimed rhythmic precision; the paper needs to specify the hit-to-target matching rule, the contact-detection threshold, and a few key hyperparameters before the empirical contribution can be fully assessed.

major comments (3)

[§IV-B and Figure 7] The F1 evaluation metric is not sufficiently defined. The text says scores are computed by comparing the ground-truth RCC to "the robot's actual hits at each timestep," but no matching rule is stated. With the 50 fps discretization mentioned in the Figure 7 caption, a strict per-timestep match would imply a 20 ms timing tolerance; any hit one timestep early or late would count as wrong or missed. If instead a tolerance window is used, its size and alignment procedure must be reported. Without this information, the reported F1 scores are not interpretable, and the headline claim of near-perfect rhythmic precision cannot be verified. Please specify the matching rule, report the onset-asynchrony distribution for representative tracks, and ideally release the evaluation script.
[§III (opening paragraph) and Table I] The contact-event detection underlying both the reward and the metric is underspecified. The paper states that each drum "registers an impact when struck with a sufficient force," but the force threshold is not reported, and the text does not explain how the set E_t of drums actually hit at time t is formed. A single physical strike could plausibly register over multiple simulation timesteps, which would inflate correct-strike counts and distort precision and recall. Please report the contact threshold, any debouncing or duplicate-removal procedure, and the raw confusion-matrix counts for at least one song.
[§III-C and §III-D] Several quantities that determine the training setup are missing. Equation (3) defines the segment length P as N/M, but the actual value of P (or M) is never given; the lookahead horizon L in Section III-D.1 is not reported; and in Equation (4), the action scaling β and the PD gains Kp and Kd are not specified. These parameters directly affect both sample efficiency and the timing precision the policy can achieve, so their absence makes the experiments unreproducible and weakens the "efficient learning" claim. Please add a full hyperparameter table.

minor comments (4)

[§III-A] The decision to keep only the most frequent articulation of each drum is justified by saying evaluation "showed no significant drop in musical quality," but no quantitative evidence is provided; consider reporting the affected tracks or a small ablation.
[Table III] The checkmark notation in the observation-ablation rows is difficult to parse: the row label "Proprioception & Spatial Phase Variable Contact Goals" with three checkmarks does not make clear which component is present. Use explicit column headers or a clearer legend.
[§V-D] The listener study is described as "small-scale," but the report would benefit from stating the number of ratings per video, whether the 15 participants all rated overlapping sets of videos, and how the Likert scores were aggregated across participants.
[§VI] The Discussion states that the robot "kept perfect timing," but the highest mean F1 is 0.985 and the paper does not report any direct timing-error statistic. Consider replacing "perfect timing" with a qualification such as "near-perfect per-timestep hit alignment" or reporting the mean absolute onset error in milliseconds.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the RCC, reward, and F1 metric all derive from the same external MIDI targets, but no fitted parameter or self-citation is used to produce the headline results.

full rationale

The paper's derivation chain is self-contained with respect to circularity. The Rhythmic Contact Chain is constructed from external MIDI transcriptions (Section III-A and III-B), and the dense contact reward in Section III-D.3 and Table I is a hand-designed function of those same external contact targets. The F1 metric in Section IV-B compares policy rollouts to the same external target sequence over 20 independent evaluation rollouts per seed. No model parameter is fitted to the reported F1 metric, no per-song calibration is applied to generate the scores, and no load-bearing claim is justified by a self-citation. The reward and metric sharing the same ground-truth source is standard practice in reinforcement-learning evaluation and does not make the prediction equivalent to its input by construction. The undefined hit-to-target matching tolerance, while a legitimate reproducibility and verification concern, is not a circular reduction. Therefore no circular steps are identified.

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

The central results depend on several unstated or hand-chosen quantities: reward weights, segment length P, lookahead L, and simulator fidelity. These are not derived from data or external benchmarks.

free parameters (4)

Reward weights (wc, ww, wm, wp) = 1.0, -0.5, -2.0, -1.0
Hand-chosen from intuition (Section III-D, Table I); directly shape the policy and the reported F1.
Segment length P (contact steps per segment) = not reported
Chosen by the authors; decomposition quality and training parallelism depend on it, but the paper never states P.
Lookahead horizon L = not reported
Controls how many future contact goals the policy sees; ablations do not vary L, so its value is undisclosed.
Action scaling and PD gains (beta, Kp, Kd) = not reported
Controller parameters determine movement speed and contact forces; not specified, making exact reproduction impossible.

assumptions (4)

domain assumption Isaac Gym contact simulation faithfully represents drum strike timing and detection.
All metrics are measured inside the simulator; no real robot validation is presented.
domain assumption The humanoid's lower body can be ignored or fixed; only 15 upper-body DOF are controlled.
No balance or leg control is described, so the humanoid claim rests on an unstated base condition.
domain assumption A 50 Hz discretization is sufficient to represent musical timing.
Hit targets are encoded per timestep with no tolerance; the F1 claim depends on this time quantization.
domain assumption Mapping each MIDI pitch to one drum and keeping only the most frequent articulation preserves the song's rhythmic content.
Section III-A says less frequent articulations are discarded and that this caused no significant drop, but no quantitative evidence is given.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Robot Drummer: Learning Rhythmic Skills for Humanoid Drumming." pith.science (2026). https://pith.science/paper/2KDWBLYO

@misc{pith2026250711498,
  author       = {Pith},
  title        = {Pith review of: Robot Drummer: Learning Rhythmic Skills for Humanoid Drumming},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2KDWBLYO}},
  note         = {Machine review of arXiv:2507.11498}
}

read the original abstract

Humanoid robots have seen remarkable advances in dexterity, balance, and locomotion, yet their role in expressive domains such as music performance remains largely unexplored. Musical tasks, like drumming, present unique challenges such as split-second timing, rapid contacts, and multi-limb coordination over performances lasting minutes. In this paper, we introduce Robot Drummer, a simulation framework for humanoid drumming across a diverse repertoire of songs. We formulate humanoid drumming as the realization of timed contact events encoded as a Rhythmic Contact Chain. To handle the long-horizon nature of musical performance, we decompose each track into fixed-length segments and train a single policy across all segments in parallel using reinforcement learning. Through extensive experiments on over thirty popular tracks, our results demonstrate that Robot Drummer consistently achieves high F1 scores and enables efficient learning of long-horizon musical performances. The learned behaviors exhibit emergent human-like drumming strategies, such as cross-arm strikes, and adaptive stick assignments, demonstrating the potential of reinforcement learning to bring humanoid robots into the domain of creative musical performance. Project page: robotdrummer.github.io

Figures

Figures reproduced from arXiv: 2507.11498 by the authors.

**Figure 1.** The humanoid robot demonstrates expressive drumming skills across three songs: In the top row, the robot plays jazz [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗

**Figure 2.** Drum kit configuration used for training and evaluating [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Overview of the Robot Drummer: Starting from a raw MIDI drum track (left), each note-onset is first mapped to a [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (6 more)

**Figure 2.** Figure 2: Each contact event in RCC thus corresponds to spa [PITH_FULL_IMAGE:figures/full_fig_p004_2.png]

**Figure 4.** Figure 4: Sequential snapshots of a complex drum-fill segment from ”Linkin Park – In the End”, showing the humanoid robot [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: (a) Relative influence of song metrics on drumming performance across all specialist policies. Bars show the absolute [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

**Figure 6.** Figure 6: Sequence of frames, illustrating a robot performing a fast drum roll (within a one-second interval) on the snare drum [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

**Figure 7.** Figure 7: Performance comparison of Nirvana - Lithium (top) and [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

**Figure 8.** Figure 8: F1 scores for specialist versus generalist policies on [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

20 extracted references · 9 canonical work pages

[1]

Humanplus: Humanoid shadowing and imitation from humans,

Z. Fu, Q. Zhao, Q. Wu, G. Wetzstein, and C. Finn, “Humanplus: Humanoid shadowing and imitation from humans,” in Conference on Robot Learning (CoRL) , 2024

2024
[2]

Learning getting-up policies for real-world humanoid robots,

X. He, R. Dong, Z. Chen, and S. Gupta, “Learning getting-up policies for real-world humanoid robots,” arXiv preprint arXiv:2502.12152 , 2025

arXiv 2025
[3]

Learning humanoid standing-up control across diverse postures,

T. Huang, J. Ren, H. Wang, Z. Wang, Q. Ben, M. Wen, X. Chen, J. Li, and J. Pang, “Learning humanoid standing-up control across diverse postures,” arXiv preprint arXiv:2502.08378 , 2025

arXiv 2025
[4]

Learn- ing human-to-humanoid real-time whole-body teleoperation,

T. He, Z. Luo, W. Xiao, C. Zhang, K. Kitani, C. Liu, and G. Shi, “Learn- ing human-to-humanoid real-time whole-body teleoperation,” 2024

2024
[5]

Asap: Aligning simulation and real-world physics for learning agile humanoid whole-body skills,

T. He, J. Gao, W. Xiao, Y . Zhang, Z. Wang, J. Wang, Z. Luo, G. He, N. Sobanbabu, C. Pan, Z. Yi, G. Qu, K. Kitani, J. Hodgins, L. J. Fan, Y . Zhu, C. Liu, and G. Shi, “Asap: Aligning simulation and real-world physics for learning agile humanoid whole-body skills,” arXiv preprint arXiv:2502.01143, 2025

arXiv 2025
[6]

Robopianist: Dexterous piano playing with deep reinforcement learning,

K. Zakka, P. Wu, L. Smith, N. Gileadi, T. Howell, X. B. Peng, S. Singh, Y . Tassa, P. Florence, A. Zeng, and P. Abbeel, “Robopianist: Dexterous piano playing with deep reinforcement learning,” inConference on Robot Learning (CoRL), 2023

work page 2023
[7]

Em- bodied intelligence for drumming; a reinforcement learning approach to drumming robots,

S. M. Karbasi, A. R. Jensenius, R. I. Godøy, and J. Torresen, “Em- bodied intelligence for drumming; a reinforcement learning approach to drumming robots,” Frontiers in Robotics and AI , vol. 11, p. 1450097, 2024

work page 2024
[8]

Ex- pressive whole-body control for humanoid robots,

X. Cheng, Y . Ji, J. Chen, R. Yang, G. Yang, and X. Wang, “Ex- pressive whole-body control for humanoid robots,” arXiv preprint arXiv:2402.16796, 2024

arXiv 2024

Show all 20 references

[9]

Visual imitation enables contextual humanoid control,

A. Allshire, H. Choi, J. Zhang, D. McAllister, A. Zhang, C. M. Kim, T. Darrell, P. Abbeel, J. Malik, and A. Kanazawa, “Visual imitation enables contextual humanoid control,” arXiv preprint arXiv:2505.03729, 2025

2025 arXiv
[10]

Hub: Learning extreme humanoid balance,

T. Zhang, B. Zheng, R. Nai, Y . Hu, Y .-J. Wang, G. Chen, F. Lin, J. Li, C. Hong, K. Sreenath et al., “Hub: Learning extreme humanoid balance,” arXiv preprint arXiv:2505.07294 , 2025

2025 arXiv
[11]

Mobile-television: Predictive motion priors for humanoid whole-body control,

C. Lu, X. Cheng, J. Li, S. Yang, M. Ji, C. Yuan, G. Yang, S. Yi, and X. Wang, “Mobile-television: Predictive motion priors for humanoid whole-body control,” arXiv preprint arXiv:2412.07773 , 2024

2024 arXiv
[12]

Wococo: Learning whole-body humanoid control with sequential contacts,

C. Zhang, W. Xiao, T. He, and G. Shi, “Wococo: Learning whole-body humanoid control with sequential contacts,” in 8th Annual Conference on Robot Learning
[13]

The design of a perceptual and impro- visational robotic marimba player,

G. Weinberg and S. Driscoll, “The design of a perceptual and impro- visational robotic marimba player,” in RO-MAN 2007-The 16th IEEE International Symposium on Robot and Human Interactive Communica- tion. IEEE, 2007, pp. 769–774

2007
[14]

The interactive robotic percussionist: new developments in form, mechanics, perception and interaction design,

——, “The interactive robotic percussionist: new developments in form, mechanics, perception and interaction design,” in Proceedings of the ACM/IEEE international conference on Human-robot interaction , 2007, pp. 97–104

2007
[15]

Drumming arm: an upper-limb prosthetic system to restore grip con- trol for a transradial amputee drummer,

N. Yang, R. Sha, R. Sankaranarayanan, Q. Sun, and G. Weinberg, “Drumming arm: an upper-limb prosthetic system to restore grip con- trol for a transradial amputee drummer,” in 2021 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2021, pp. 10 317–10 323

2021
[16]

Multisensory learning framework for robot drumming,

“Multisensory learning framework for robot drumming,” arXiv preprint arXiv:1907.09775, 2019

1907 arXiv
[17]

Prox- imal policy optimization algorithms,

J. Schulman, F. Wolski, P. Dhariwal, A. Radford, and O. Klimov, “Prox- imal policy optimization algorithms,” arXiv preprint arXiv:1707.06347, 2017

2017 arXiv
[18]

Deepmimic: Example-guided deep reinforcement learning of physics-based character skills,

X. B. Peng, P. Abbeel, S. Levine, and M. Van de Panne, “Deepmimic: Example-guided deep reinforcement learning of physics-based character skills,” ACM Transactions On Graphics (TOG), vol. 37, no. 4, pp. 1–14, 2018

2018
[19]

Isaac gym: High performance gpu based physics simulation for robot learning,

V . Makoviychuk, L. Wawrzyniak, Y . Guo, M. Lu, K. Storey, M. Macklin, D. Hoeller, N. Rudin, A. Allshire, A. Handa et al. , “Isaac gym: High performance gpu based physics simulation for robot learning,” in Thirty- fifth Conference on Neural Information Processing Systems Datas...
[20]

The pairwise variability index as a measure of rhythm complexity,

G. T. Toussaint, “The pairwise variability index as a measure of rhythm complexity,” Analytical approaches to world music , vol. 2, no. 2, pp. 1–42, 2013

2013

Pith tools

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

[1] [1]

Humanplus: Humanoid shadowing and imitation from humans,

Z. Fu, Q. Zhao, Q. Wu, G. Wetzstein, and C. Finn, “Humanplus: Humanoid shadowing and imitation from humans,” in Conference on Robot Learning (CoRL) , 2024

2024

[2] [2]

Learning getting-up policies for real-world humanoid robots,

X. He, R. Dong, Z. Chen, and S. Gupta, “Learning getting-up policies for real-world humanoid robots,” arXiv preprint arXiv:2502.12152 , 2025

arXiv 2025

[3] [3]

Learning humanoid standing-up control across diverse postures,

T. Huang, J. Ren, H. Wang, Z. Wang, Q. Ben, M. Wen, X. Chen, J. Li, and J. Pang, “Learning humanoid standing-up control across diverse postures,” arXiv preprint arXiv:2502.08378 , 2025

arXiv 2025

[4] [4]

Learn- ing human-to-humanoid real-time whole-body teleoperation,

T. He, Z. Luo, W. Xiao, C. Zhang, K. Kitani, C. Liu, and G. Shi, “Learn- ing human-to-humanoid real-time whole-body teleoperation,” 2024

2024

[5] [5]

Asap: Aligning simulation and real-world physics for learning agile humanoid whole-body skills,

T. He, J. Gao, W. Xiao, Y . Zhang, Z. Wang, J. Wang, Z. Luo, G. He, N. Sobanbabu, C. Pan, Z. Yi, G. Qu, K. Kitani, J. Hodgins, L. J. Fan, Y . Zhu, C. Liu, and G. Shi, “Asap: Aligning simulation and real-world physics for learning agile humanoid whole-body skills,” arXiv preprint arXiv:2502.01143, 2025

arXiv 2025

[6] [6]

Robopianist: Dexterous piano playing with deep reinforcement learning,

K. Zakka, P. Wu, L. Smith, N. Gileadi, T. Howell, X. B. Peng, S. Singh, Y . Tassa, P. Florence, A. Zeng, and P. Abbeel, “Robopianist: Dexterous piano playing with deep reinforcement learning,” inConference on Robot Learning (CoRL), 2023

work page 2023

[7] [7]

Em- bodied intelligence for drumming; a reinforcement learning approach to drumming robots,

S. M. Karbasi, A. R. Jensenius, R. I. Godøy, and J. Torresen, “Em- bodied intelligence for drumming; a reinforcement learning approach to drumming robots,” Frontiers in Robotics and AI , vol. 11, p. 1450097, 2024

work page 2024

[8] [8]

Ex- pressive whole-body control for humanoid robots,

X. Cheng, Y . Ji, J. Chen, R. Yang, G. Yang, and X. Wang, “Ex- pressive whole-body control for humanoid robots,” arXiv preprint arXiv:2402.16796, 2024

arXiv 2024