Pith.

Robotics · arXiv:2610.17843 v1 · submitted 13 Oct 2026 · 52 references

Physics-Aware Object-Centric SE(3) Forecasting for Real-Time Proactive Dynamic Grasping

Yirui Zhou, Hongli Xu, Ying Zheng, Yi Wang, Lap-Pui Chau

The paper says

“Proactive dynamic grasping requires robots to predict the future state at which physical contact will occur, rather than reacting to the currently observed object pose. But what future must the robot predict, and at what level of abstraction?”

more of the abstract

We argue that contact-relevant rigid-body state in SE(3) is a compact task-sufficient abstraction for this problem: it directly determines where and when the robot should establish contact, without requiring visually complete future scenes. This insight leads to a lightweight object-centric forecast-then-select framework that predicts future rigid-body trajectories from short pose histories and uses a reachability-aware interception selector to choose feasible contact states under kinematic and timing constraints. Trained with randomized physics simulation over primitive shapes, the forecaster transfers to real objects once their metric geometry and SE(3) state histories are recovered. Experiments across simulation and real-world dynamic grasping tasks show that our framework improves the average success rate from \(24.5%\) to \(62.0%\) in simulation, and from \(15.3%\) to \(67.3%\) in real-world catching, conveyor, and turntable interception. The SE(3) forecaster runs in \(15\,\mathrmms\), orders of magnitude faster than dense visual future prediction, enabling real-time contact decision making. Our results suggest that effective proactive grasping does not require imagining the entire future scene. Instead, predicting the compact physical future that determines contact provides an effective and practical abstraction for real-time dynamic grasping.

from the abstract

Pith says

Machine review · deepseek-v4-flash · 14 Oct 2026 · what was read was not recorded · not checked by a person

It can be wrong. Challenge this review or reply to it.

Holds up, with conditions. A well-motivated paper arguing that compact SE(3) state prediction is the right abstraction for dynamic grasping, but the evaluation needs tighter baselines and robustness tests to fully back the claim. This line is wrong

3 major objections · 5 minor

How often objections from reviews like this one are right: not yet measured

Are you an author? What you can do.

Not checked: figures, code, what the 52 references say.

The paper says, at §4.1, Real-world experiments

The review did not quote this passage.

Pith says: Objection 1 of 3 · major

The inertia prior ψ=I_B is shown to reduce rotational final displacement error (Rot-FDE) from 11.24° to 4.75° on the toss task, indicating its crucial role for rotational forecasting. Yet the paper provides no quantitative analysis of how degradation in the estimated inertia prior (e.g., from VLM material classification errors, mesh reconstruction noise, or density estimation errors) affects forecasting accuracy or grasping success rate.

more of objection 1

The authors acknowledge in §5 that the inertia prior assumes a coarse single-material density but do not test robustness to realistic errors. Since real-world estimation of I_B is approximate, this omission weakens the evidence for the sim-to-real transfer claim. I recommend adding a simulation ablation that injects controlled noise into the inertia prior (e.g., scaling errors of ×0.5–×2, mass distribution misspecifications) and reports the impact on Rot-FDE, Pos-FDE, and success rate across motion families.

The paper says, at §4.1, Simulation setup

The review did not quote this passage.

Pith says: Objection 2 of 3 · major

The real-world comparison pits PAF-Grasp (which receives SE(3) pose histories, forecasts, and selected contact states) against π0.5 (which receives only RGB, language, and robot proprioception). The paper correctly states that this evaluates the effect of making contact states explicit, but the conflated factor of input modality prevents isolating the benefit of forecasting from the benefit of having object-centric SE(3) state inputs in general.

more of objection 2

A clearer ablation would be to evaluate a version of π0.5 augmented with current SE(3) poses (without forecasting), or to decompose the gain into the portions attributable to SE(3) input versus the forecast-then-select mechanism. This would clarify the core contribution of the forecasting component.

The paper says, at §4.1 and §F

The review did not quote this passage.

A classifier read this sentence of the paper as acknowledging the concern (probability 0.75):

“In our real-world setup, high-speed 6D tracking [46] of the asymmetric hammer was sensitive to motion blur and rapid pose changes, making the measured success rate strongly coupled to percep- tion stability rather than only to the interception policy.”

Pith says: Objection 3 of 3 · major

Simulation experiments use ground-truth object poses as forecaster inputs, which appropriately isolates forecasting performance. However, real-world results include perception errors from FoundationPose++ and mesh reconstruction. The paper does not quantify the sensitivity of overall system performance to perception noise. Adding a simulation ablation with realistic pose noise (based on measured perception accuracy) or reporting perception-noise statistics (e.g., translation/rotation RMSE of …

more of objection 3

… the tracker on the real setup) would improve the transparency of the sim-to-real transfer and help readers assess the robustness of the results.

5 minor objections

The paper says, at Abstract and §4.2

Pith says: minor objection 1 of 5

The abstract claims 'orders of magnitude faster than dense visual future prediction' but does not include the concrete runtime numbers for video models (30–180 s from Table 9) in the main text. Consider adding a brief runtime comparison in §4.2 to strengthen the point without relying on the appendix alone.

The paper says, at Table 1

Pith says: minor objection 2 of 5

The success rate for VLASH on Conveyor is listed as '32' without a percent sign. For consistency, all entries in the SR columns should include the '%' symbol or be explicitly stated as percentages.

show 3 more

The paper says, at Figure 7

Pith says: minor objection 3 of 5

The y-axis label on the left subfigure reads 'Position Error (m)' with an extra closing parenthesis in the PDF rendering. Please correct the label to 'Position Error (m)'. Also consider increasing font sizes in the legend for readability.

The paper says, at §3.2, Eq. (2)

Pith says: minor objection 4 of 5

The quaternion derivative '˙˜q = 1/2 ˜q ⊗ [0, ω^B]' is given correctly, but the earlier description of the state vector could benefit from explicitly stating that the quaternion is stored as the relative rotation from the last observed orientation. A brief clarifying sentence immediately before the equation would reduce potential confusion.

The paper says, at §3.3 and §5

Pith says: minor objection 5 of 5

The paper uses the phrase 'coarse single-material density' in the limitations but does not give an estimated error range for typical VLM-based density estimation. While not required, a rough uncertainty quantification (e.g., density error ±x% from the VLM) would help readers gauge the realism of the prior.

What would settle it

Run the toss forecasting experiment (Table 4 setting) with synthetic noise on I_B: scale the diagonal elements by factors 0.5, 0.75, 1.25, 1.5, and also add off-diagonal perturbations of up to 10% of the max moment. Record Pos-FDE, Rot-FDE, and the downstream success rate using the interception selector. If Rot-FDE stays below, say, 8° and success rate stays above 45% for all perturbations up to 25% error, the robustness is supported. If not, the sim-to-real transfer claim is materially weaker.

from the skeptic's reading

How the review went

4 readings by deepseek-v4-flash · 14 Oct 2026

  1. Reader holds up, with conditions confidence moderate
  2. Skeptic partly agrees 30 s · 20,284 tokens read, 3,497 written
  3. Editor wrote the verdict above 13 s · 21,180 tokens read, 1,133 written
  4. Referee major revision: 3 major objections, 5 minor 36 s · 20,609 tokens read, 4,225 written

Not run on this paper: the simulated authors' defence.

Pith did not check this review's objections against the paper.

The review in full

The review in brief, machine-written

The core idea here is clean and sensible: instead of generating pixel-level video futures for dynamic grasping, predict the object’s future SE(3) state directly. That’s the right question to ask, and the authors make a good case for why it matters. The framework (PAF-Grasp) is clearly described, the sim-to-real recipe using primitive shapes plus an inertia prior is clever, and the latency numbers (15 ms) are genuinely impressive. The success rate gains—from 24.5% to 62% in simulation, 15.3% to 67.3% in the real world—are large enough to catch attention.

Now the soft spots, and these are real. The comparison with video generation models is explicitly “not input-matched” ([b135]), which the authors acknowledge, but that makes the headline claim about “outperforming dense visual futures” weaker than it reads. The real-world baseline (π0.5) is trained on expert trajectories collected by the system itself, which conflates representation quality with data quality. There are no error bars reported anywhere, which matters when the absolute success rates are moderate. And the stress-test concern—that the inertia prior (IB) is crucial for rotational accuracy (Table 4: Rot-FDE drops from 11.24° to 4.75° with IB) yet the paper never tests how VLM material errors or mesh noise degrade forecasting—is a genuine gap. The authors list sensible limitations in the conclusion, but the central transfer claim would be stronger with a quantitative ablation on prior degradation.

Still, this is a serious piece of work. The reasoning is sound, the experiments are systematic (if not perfectly controlled), and the argument that contact-relevant SE(3) is a better abstraction than pixel futures is worth taking seriously. The paper is for robotics researchers working on dynamic manipulation, time-critical grasping, or prediction-based control. I’d want to see the comparisons tightened and error bars added before fully endorsing the strongest claims, but the core contribution deserves a proper hearing.

The strongest objection

The central claim that the forecaster transfers from simulated primitives to real objects via a recovered inertial prior is plausible but untested under realistic degradation. Section 3.3 states that the inertia prior ψ = I_B provides object-specific rotational structure, and Table 4 shows its large effect on toss Rot-FDE (11.24° → 4.75°). Yet the paper never injects noise into the prior in simulation to quantify robustness. The real-world pipeline (§F) uses a VLM to infer material category → density → inertia, and a single-view mesh recovry. Both sources can be significantly off e.g., wood vs. plastic density error >2×. Without an ablation, it is unknown whether the forecaster's learned acceleration field can compensate for such estmiation errors, or whether the real-world gains (67.3% vs. 15.3% depend on unusually accurate inertia estimates. The authors acknoledge 'coarse single-material density' and 'spatially varing mass distributions remain outside scope' but do not test how these limitions affect performance. This leaves a gap in the evidence for the sim-to-real transfer claim.

The reader's weakest_assumption flagged the sim-to-real transfer and noted that the paper does not quantify degradation in the inertial prior estimate. I agree that this is a key unaddressed issue. However, I consider it load-bearing because the success of the method hinges on the forecaster's ability to generalize across objects with different inertias, and the paper's own ablation shows I_B is critical for rotation. The proposed concrete test would either confirm robustness (strengthening the claim) or reveal a vulnerability (weakening it). The reader's verdict was CONDITIONAL already, and this concern adds a specific condition rather than overturning the existing judgment, so UNCHANGED is appropriate. The paper does provide other strong evidence: simulation comparisons with fair baselines (Table 1), NODE advantages over AR (Table 5, Fig. 7), and real-world results that, while comparing against a baseline trained on its own data, show large absolute margins. Thus the central claim remains plausible but requires the robustness check to become fully convincing.

The referee's report

If the claimed results hold, PAF-Grasp offers a practical and efficient approach for dynamic grasping by explicitly decoupling object-centric forecasting from robot-specific execution. The work provides systematic evidence that explicit physical state prediction outperforms dense visual world models in accuracy, physical consistency, and real-time efficiency for this contact-critical setting. The framework is lightweight, sim-to-real transferable with minimal object-specific setup, and compatible with real-time control loops. These findings support a broader shift from reactive policies and dense world models toward physically structured, action-relevant state abstractions for dynamic manipulation.

Pith's reading of the paper (a machine's summary, not the authors')

Robots that need to catch or grasp moving objects typically either react to the current observed pose or try to predict entire future videos. This paper argues that neither is efficient. Instead, the robot only needs to predict where and how the object will be at the moment of contact. That information is fully captured by the object's future position and orientation (its SE(3) pose) a compact six-number vector at each future time step. The authors build a lightweight neural network that learns rigid-body motion from short pose histories. It uses a continuous-time neural ODE formulation with Newton-Euler dynamics as a structural prior, and conditions on an approximate inertia estimate of the object. The network is trained entirely on simulated primitive shapes (spheres, cubes, cylinders), then transferred to real objects by estimating mesh and density from a single RGB-D image. At runtime, the forecaster predicts a trajectory of future SE(3 poses in about 15 milliseconds. A separate reachability-aware selector then chooses a specific future contact state that the robot can reach in time, avoiding stale targets. Experiments across four simulated and three real-world dynamic grasping tasks show success rate improvements of 37.5 percentage points in simulation (from 24.5% to 62.0%) and 52.0 percentage points in real-world tasks (from 15.3% to 67.3%) compared to reactive visuomotor policies. The forecaster also outperforms dense video generation models on contact-relevant prediction accuracy (0.276 mIoU vs at most 0.047 for video models) while being orders of magnitude faster.

Core claim. Contact-relevant rigid-body state in SE(3) is a compact task-sufficient abstraction for proactive dynamic grasping, and PAF-Grasp improves average success rate from 24.5% to 62.0% in simulation and from 15.3% to 67.3% in real-world tasks (Abstract, [b10]).

The premise it rests on. The forecaster trained exclusively on simulated primitive shapes (spheres, cuboids, cylinders) with randomized dynamics transfers to real objects without object-specific retraining, relying on a recovered inertial prior (I_B) and SE(3) pose histories. This assumption is stated in §3.3: 'This transfer is enabled by the dynamics abstraction in Eq. (2).' The paper does not quantify how degradation in the inertial prior estimate (e.g., from VLM material classification or mesh reconstruction errors) affects forecasting accuracy or success rate.

Check this review

Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.

How often reviews like this one are right

Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.

How often objections from reviews like this one are right: not yet measured

Every value above, in the review's public record (JSON)

Reviews by people

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review

How Pith checks papers · The full audit of this paper

The paper

Abstract

Proactive dynamic grasping requires robots to predict the future state at which physical contact will occur, rather than reacting to the currently observed object pose. But what future must the robot predict, and at what level of abstraction? We argue that contact-relevant rigid-body state in SE(3) is a compact task-sufficient abstraction for this problem: it directly determines where and when the robot should establish contact, without requiring visually complete future scenes. This insight leads to a lightweight object-centric forecast-then-select framework that predicts future rigid-body trajectories from short pose histories and uses a reachability-aware interception selector to choose feasible contact states under kinematic and timing constraints. Trained with randomized physics simulation over primitive shapes, the forecaster transfers to real objects once their metric geometry and SE(3) state histories are recovered. Experiments across simulation and real-world dynamic grasping tasks show that our framework improves the average success rate from \(24.5%\) to \(62.0%\) in simulation, and from \(15.3%\) to \(67.3%\) in real-world catching, conveyor, and turntable interception. The SE(3) forecaster runs in \(15\,\mathrmms\), orders of magnitude faster than dense visual future prediction, enabling real-time contact decision making. Our results suggest that effective proactive grasping does not require imagining the entire future scene. Instead, predicting the compact physical future that determines contact provides an effective and practical abstraction for real-time dynamic grasping.

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.

abstract compiled · 2026-10-14

The authors' main claim, as the classifier picked it out 0.92

“Instead, predicting the compact physical future that determines contact provides an effective and practical abstraction for real-time dynamic grasping.”

How the main claim is stated: asserted 0.93

What the abstract itself states, as probabilities: claims first 0.04 · states limitation 0.31 · human participants 0.02 · controlled comparison 0.72 · randomized assignment 0.13 · states quantitative result 0.99 · compares with prior results 0.80 · mentions code or data availability 0.03

The abstract, sentence by sentence

This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.

SentenceRoleThe authors' own workStates the main conclusionStrength
Proactive dynamic grasping requires robots to predict the future state at which physical contact will occur, rather than reacting to the currently observed object pose. background 1.00 0.07 0.06 asserted 1.00
But what future must the robot predict, and at what level of abstraction? gap 0.94 0.06 0.02 unclear 0.98
We argue that contact-relevant rigid-body state in SE(3) is a compact task-sufficient abstraction for this problem: it directly determines where and when the robot should establish contact, without requiring visually complete future scenes. contribution 0.75 0.30 0.39 asserted 0.93
This insight leads to a lightweight object-centric forecast-then-select framework that predicts future rigid-body trajectories from short pose histories and uses a reachability-aware interception selector to choose feasible contact states under kinematic and timing constraints. method 0.84 0.88 0.13 asserted 1.00
Trained with randomized physics simulation over primitive shapes, the forecaster transfers to real objects once their metric geometry and SE(3) state histories are recovered. method 0.99 0.89 0.05 asserted 0.97
Experiments across simulation and real-world dynamic grasping tasks show that our framework improves the average success rate from \(24.5\%\) to \(62.0\%\) in simulation, and from \(15.3\%\) to \(67.3\%\) in real-world catching, conveyor, and turntable interception. result 1.00 0.97 0.18 asserted 1.00
The SE(3) forecaster runs in \(15\,\mathrm{ms}\), orders of magnitude faster than dense visual future prediction, enabling real-time contact decision making. result 1.00 0.95 0.08 asserted 1.00
Our results suggest that effective proactive grasping does not require imagining the entire future scene. interpretation 1.00 0.78 0.49 hedged possibility 1.00
Instead, predicting the compact physical future that determines contact provides an effective and practical abstraction for real-time dynamic grasping. interpretation 1.00 0.67 0.73 asserted 0.93

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.98

What was examined, and what was not

Sources read

  • arXiv listing abstract v1

Not examined

  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers

Levels run

  • P abstract

Not run

  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

How the labels were made

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-10-14: 1 request, 7,077 input tokens.

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON

Figures (8)
Figure 1 of the paper
Figure 1. Left: Reactive policies act on the current object pose and fail under fast dynamics. Video
Figure 2 of the paper
Figure 2. Overview of PAF-Grasp. The Neural SE(3) forecaster is trained on Genesis-generated
Figure 3 of the paper
Figure 3. Real-world success rates across three dynamic
Figure 4 of the paper
Figure 4. Success rate on turntable handle grasping. GT Ours Kling Veo
Figure 6 of the paper
Figure 6. Fast-WAM (right) failure: the gripper reaches the cup but misses the rotating handle. Comparison with world action model. The pre￾vious comparison evaluates dense visual future prediction as an explicit future representation. We further ask whether such representations are suffi￾cient when used inside a real-time action model rather than queried as a separate video rollout. To this end, we compare with Fast-WAM [29] on turntable handle grasping, where success requires matching the future rotational phase of a rotating cup handle. Fast-WAM uses pretrained Wan2.2- 5B [25] as its video…
Figure 7 of the paper
Figure 7. Extrapolation beyond the supervised horizon. Left:
Figure 8 of the paper
Figure 8. Real-world experimental setup for dynamic grasping.

Showing 7 of the paper's 8 figures. Figures are the authors', reproduced from the paper.

References (52)
  1. [1] T. Z. Zhao, V . Kumar, S. Levine, and C. Finn. Learning fine-grained bimanual manipulation with low-cost hardware. InICML Workshop on New Frontiers in Learning, Control, and Dynamical Systems on Pith
  2. [2] C. Chi, Z. Xu, S. Feng, E. Cousineau, Y . Du, B. Burchfiel, R. Tedrake, and S. Song. Diffusion policy: Visuomotor policy learning via action diffusion.The International Journal of Robotics Research, 44(10-11):1684–1704, 2025 on Pith
  3. [3] Y . Ze, G. Zhang, K. Zhang, C. Hu, M. Wang, and H. Xu. 3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations. In2nd Workshop on Dexterous Manipulation: Design, Perception and Control (RSS) on Pith
  4. [4] T. Gervet, Z. Xian, N. Gkanatsios, and K. Fragkiadaki. Act3d: 3d feature field transformers for multi-task robotic manipulation. In7th Annual Conference on Robot Learning on Pith
  5. [5] H. Xie, B. Wen, J. Zheng, Z. Chen, F. Hong, H. Diao, and Z. Liu. Dynamicvla: A vision- language-action model for dynamic object manipulation.arXiv, 2601.22153, 2026 on PithPith's reviewarXiv
  6. [6] K. Black, N. Brown, D. Driess, A. Esmail, M. Equi, C. Finn, N. Fusai, L. Groom, K. Hausman, B. Ichter, et al.π 0: A vision-language-action flow model for general robot control.arXiv preprint arXiv:2410.24164, 2024 on PithPith's reviewarXiv
  7. [7] P. Intelligence, K. Black, N. Brown, J. Darpinian, K. Dhabalia, D. Driess, A. Esmail, M. Equi, C. Finn, N. Fusai, et al.π 0.5: a vision-language-action model with open-world generalization. arXiv preprint arXiv:2504.16054, 2025 on PithPith's reviewarXiv
  8. [8] P. Intelligence, A. Amin, R. Aniceto, A. Balakrishna, K. Black, K. Conley, G. Connors, J. Darpinian, K. Dhabalia, J. DiCarlo, et al.π ∗0.6: a vla that learns from experience.arXiv preprint arXiv:2511.14759, 2025 on PithPith's reviewarXiv
  9. [9] J. Tang, Y . Sun, Y . Zhao, S. Yang, Y . Lin, Z. Zhang, J. Hou, Y . Lu, Z. Liu, and S. Han. Vlash: Real-time vlas via future-state-aware asynchronous inference.arXiv preprint arXiv:2512.01031, 2025 on PithPith's reviewarXiv
  10. [10] J. Bjorck, F. Casta ˜neda, N. Cherniadev, X. Da, R. Ding, L. Fan, Y . Fang, D. Fox, F. Hu, S. Huang, et al. Gr00t n1: An open foundation model for generalist humanoid robots.arXiv preprint arXiv:2503.14734, 2025 on PithPith's reviewarXiv
  11. [11] Y . Meng, Z. Bing, X. Yao, K. Chen, K. Huang, Y . Gao, F. Sun, and A. Knoll. Preserving and combining knowledge in robotic lifelong reinforcement learning.Nature Machine Intelligence, 7(2):256–269, 2025 on Pith
  12. [12] X. Yao, Y . Zhou, Y . Meng, Y . Liu, Z. Zhang, L. Dong, Z. Bing, K. Huang, F. Sun, and A. Knoll. Inference-stage adaptation-projection strategy adapts diffusion policy to cross-manipulators scenarios. In2026 IEEE International Conference on Robotics and Automation (ICRA), page... on Pithdoi:10.1109/icra57385.2026.11697208
  13. [13] Z. Zhang, S. Yang, Q. Hu, L. J. Huang, J. Hou, Y . Sun, Y . Lu, and S. Han. Foreact: Steering your vla with efficient visual foresight planning.arXiv preprint arXiv:2602.12322, 2026. 10 on PitharXiv
  14. [14] M. J. Kim, Y . Gao, T.-Y . Lin, Y .-C. Lin, Y . Ge, G. Lam, P. Liang, S. Song, M.-Y . Liu, C. Finn, et al. Cosmos policy: Fine-tuning video models for visuomotor control and planning.arXiv preprint arXiv:2601.16163, 2026 on PithPith's reviewarXiv
  15. [15] L. Li, Q. Zhang, Y . Luo, S. Yang, R. Wang, F. Han, M. Yu, Z. Gao, N. Xue, X. Zhu, et al. Causal world modeling for robot control.arXiv preprint arXiv:2601.21998, 2026 on PithPith's reviewarXiv
  16. [16] S. Ye, Y . Ge, K. Zheng, S. Gao, S. Yu, G. Kurian, S. Indupuru, Y . L. Tan, C. Zhu, J. Xiang, et al. World action models are zero-shot policies.arXiv preprint arXiv:2602.15922, 2026 on PithPith's reviewarXiv
  17. [17] Z. Li, X. Wu, Z. Xu, and H. Zhao. Train once, deploy anywhere: Realize data-efficient dynamic object manipulation.arXiv preprint arXiv:2508.14042, 2025 on PitharXiv
  18. [18] Y . Zhang, R. Wang, and X. Chen. Dynamic behavior cloning with temporal feature prediction: Enhancing robotic arm manipulation in moving object tasks.IEEE Robotics and Automation Letters, 10(6):5209–5216, 2025. doi:10.1109/LRA.2025.3557746 on Pithdoi:10.1109/lra.2025.3557746
  19. [19] Y . Jia, J. Xu, D. Jayaraman, and S. Song. Dynamic grasping with a learned meta-controller. In2024 IEEE 20th International Conference on Automation Science and Engineering (CASE), pages 3608–3615, 2024. doi:10.1109/CASE59546.2024.10711690 on Pithdoi:10.1109/case59546.2024.10711690
  20. [20] I. Akinola, J. Xu, S. Song, and P. K. Allen. Dynamic grasping with reachability and motion awareness. In2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 9422–9429, 2021. doi:10.1109/IROS51168.2021.9636057 on Pithdoi:10.1109/iros51168.2021.9636057
show 32 more
  1. [21] Y . Zheng, L. Yao, Y . Su, Y . Zhang, Y . Wang, S. Zhao, Y . Zhang, and L.-P. Chau. A survey of embodied learning for object-centric robotic manipulation.Machine Intelligence Research, 22 (4):588–626, 2025 on Pith
  2. [22] H.-S. Fang, C. Wang, H. Fang, M. Gou, J. Liu, H. Yan, W. Liu, Y . Xie, and C. Lu. Anygrasp: Robust and efficient grasp perception in spatial and temporal domains.IEEE Transactions on Robotics, 39(5):3929–3945, 2023 on Pith
  3. [23] N. Chen, X.-M. Wu, G. Xu, J.-J. Jiang, Z. Chen, and W.-S. Zheng. Motiongrasp: Long-term grasp motion tracking for dynamic grasping.IEEE Robotics and Automation Letters, 10(1): 796–803, 2025. doi:10.1109/LRA.2024.3504792 on Pithdoi:10.1109/lra.2024.3504792
  4. [24] J. Liu, R. Zhang, H.-S. Fang, M. Gou, H. Fang, C. Wang, S. Xu, H. Yan, and C. Lu. Target- referenced reactive grasping for dynamic objects. In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 8824–8833, 2023. doi:10.1109/CVPR52729. 2023.00852 on Pithdoi:10.1109/cvpr52729
  5. [25] T. Wan, A. Wang, B. Ai, B. Wen, C. Mao, C.-W. Xie, D. Chen, F. Yu, H. Zhao, J. Yang, et al. Wan: Open and advanced large-scale video generative models.arXiv preprint arXiv:2503.20314, 2025 on PithPith's reviewarXiv
  6. [26] Z. Yang, J. Teng, W. Zheng, M. Ding, S. Huang, J. Xu, Y . Yang, W. Hong, X. Zhang, G. Feng, et al. Cogvideox: Text-to-video diffusion models with an expert transformer. InThe Thirteenth International Conference on Learning Representations on Pith
  7. [27] K. Team, J. Chen, Y . Ci, X. Du, Z. Feng, K. Gai, S. Guo, F. Han, J. He, K. He, et al. Kling-omni technical report.arXiv preprint arXiv:2512.16776, 2025 on PithPith's reviewarXiv
  8. [28] A. Hu, L. Russell, H. Yeo, Z. Murez, G. Fedoseev, A. Kendall, J. Shotton, and G. Corrado. Gaia-1: A generative world model for autonomous driving.arXiv preprint arXiv:2309.17080, 2023 on PithPith's reviewarXiv
  9. [29] T. Yuan, Z. Dong, Y . Liu, and H. Zhao. Fast-wam: Do world action models need test-time future imagination?arXiv preprint arXiv:2603.16666, 2026. 11 on PithPith's reviewarXiv
  10. [30] M. Team, C. Xiang, F. Bao, H. Liu, H. Tan, H. Bi, J. Li, J. Liu, J. Pang, K. Jing, et al. Mo- tubrain: An advanced world action model for robot control.arXiv preprint arXiv:2604.27792, 2026 on PithPith's reviewarXiv
  11. [31] M. Assran, Q. Duval, I. Misra, P. Bojanowski, P. Vincent, M. Rabbat, Y . LeCun, and N. Ballas. Self-supervised learning from images with a joint-embedding predictive architecture. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pag... on Pith
  12. [32] A. Bardes, Q. Garrido, J. Ponce, X. Chen, M. Rabbat, Y . LeCun, M. Assran, and N. Ballas. Revisiting feature prediction for learning visual representations from video.Transactions on Machine Learning Research on Pith
  13. [33] S. Jha, A. Zholus, S. Chandar, et al. Reconstruction or semantics? what makes a latent space useful for robotic world models.arXiv preprint arXiv:2605.06388, 2026 on PithPith's reviewarXiv
  14. [34] S. Cui and J. Ma. Physically native world models: A hamiltonian perspective on generative world modeling.arXiv preprint arXiv:2605.00412, 2026 on PithPith's reviewarXiv
  15. [35] G. Authors. Genesis: A generative and universal physics engine for robotics and beyond, December 2024. URLhttps://github.com/Genesis-Embodied-AI/Genesis on Pith
  16. [36] A. Singh, A. Fry, A. Perelman, A. Tart, A. Ganesh, A. El-Kishky, A. McLaughlin, A. Low, A. Ostrow, A. Ananthram, et al. Openai gpt-5 system card.arXiv preprint arXiv:2601.03267, 2025 on PithPith's reviewarXiv
  17. [37] N. Ravi, V . Gabeur, Y .-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. R ¨adle, C. Rolland, L. Gustafson, et al. Sam 2: Segment anything in images and videos. InThe Thirteenth In- ternational Conference on Learning Representations on Pith
  18. [38] J. Xiang, Z. Lv, S. Xu, Y . Deng, R. Wang, B. Zhang, D. Chen, X. Tong, and J. Yang. Structured 3d latents for scalable and versatile 3d generation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 21469–21480, 2025 on Pith
  19. [39] Z. Geng, N. Wang, S. Xu, C. Ye, B. Li, Z. Chen, S. Peng, and H. Zhao. One view, many worlds: Single-image to 3d object meets generative domain randomization for one-shot 6d pose estimation. In9th Annual Conference on Robot Learning on Pith
  20. [40] S. Bai, Y . Cai, R. Chen, K. Chen, X. Chen, Z. Cheng, L. Deng, W. Ding, C. Gao, C. Ge, et al. Qwen3-vl technical report.arXiv preprint arXiv:2511.21631, 2025 on PithPith's reviewarXiv
  21. [41] E. Croft, R. Fenton, and B. Benhabib. Optimal rendezvous-point selection for robotic inter- ception of moving objects.IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 28(2):192–204, 1998. doi:10.1109/3477.662759 on Pithdoi:10.1109/3477.662759
  22. [42] R. Lampariello, D. Nguyen-Tuong, C. Castellini, G. Hirzinger, and J. Peters. Trajectory planning for optimal robot catching in real-time. In2011 IEEE International Conference on Robotics and Automation, pages 3719–3726, 2011. doi:10.1109/ICRA.2011.5980114 on Pithdoi:10.1109/icra.2011.5980114
  23. [43] K. Pertsch, K. Stachowicz, B. Ichter, D. Driess, S. Nair, Q. Vuong, O. Mees, C. Finn, and S. Levine. Fast: Efficient action tokenization for vision-language-action models.arXiv preprint arXiv:2501.09747, 2025 on PithPith's reviewarXiv
  24. [44] J. K. Christopher, S. Baek, and F. Fioretto. Constrained synthesis with projected diffusion models.Advances in Neural Information Processing Systems, 37:89307–89333, 2024 on Pith
  25. [45] X. Yao, Y . Zhou, Y . Meng, Y . Liu, L. Dong, Z. Zhang, Z. Bing, K. Huang, F. Sun, and A. Knoll. Inference-stage adaptation-projection strategy adapts diffusion policy to cross-manipulators scenarios.arXiv preprint arXiv:2509.11621, 2025. 12 on PitharXiv
  26. [46] teal024. FoundationPose++: Simple tricks boost foundationpose performance in high-dynamic scenes.https://github.com/teal024/FoundationPose-plus-plus, 2025. Accessed: 2026-05-16 on Pith
  27. [47] B. Wen, W. Yang, J. Kautz, and S. Birchfield. Foundationpose: Unified 6d pose estimation and tracking of novel objects. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 17868–17879, 2024 on Pith
  28. [48] H. K. Cheng, S. W. Oh, B. Price, J.-Y . Lee, and A. Schwing. Putting the object back into video object segmentation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3151–3161, 2024 on Pith
  29. [49] K. Black, M. Galliker, and S. Levine. Real-time execution of action chunking flow policies. Advances in Neural Information Processing Systems, 38:33383–33407, 2026 on Pith
  30. [50] T. Wiedemer, Y . Li, P. Vicol, S. S. Gu, N. Matarese, K. Swersky, B. Kim, P. Jaini, and R. Geirhos. Video models are zero-shot learners and reasoners.arXiv preprint arXiv:2509.20328, 2025 on PithPith's reviewarXiv
  31. [51] J. Lee, J. Jung, J. Han, T. Narihira, K. Fukuda, J. Seo, S. Hong, Y . Mitsufuji, and S. Kim. 3d scene prompting for scene-consistent camera-controllable video generation. InInternational Conference on Learning Representations, volume 2026, pages 73753–73777, 2026 on Pith
  32. [52] J.-Y . Kang, H. Cho, T. Lee, M. Kang, B. Wen, Y . Kim, and K.-J. Yoon. Event6d: Event-based novel object 6d pose tracking. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026. 13 A Implementation Details SE(3) Forecaster training.Ta... on Pith

Challenges and discussion

Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.

No comments yet.

Sign in to comment