Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Pixel Motion as Universal Representation for Robot Control

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read Language-driven robot control can be built on predicted pixel motion rather than generated images.

desk verdict LangToMo is a solid, incremental-plus step in the flow-as-action representation line; the benchmarks and ablations are convincing, but 'universal' and 'unsupervised' overclaim what the hand-crafted mappings actually support. read the letter →

arxiv 2505.07817 v2 pith:WIDBK3J3 submitted 2025-05-12 cs.RO cs.CV

classification cs.ROcs.CV
keywords pixelmotionopticalflowvision-language-actionrobotmanipulationdiffusionmodelszero-shotcontroluniversalrepresentationvideo-captionpretraining
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

LangToMo claims that dense pixel motion—the apparent movement of every pixel between frames—is a universal intermediate representation for robot control: embodiment-agnostic, interpretable, and learnable from plain video-caption data. A text-conditioned diffusion model (System 2) predicts future pixel motion from a single current frame and a language instruction, without seeing the future. A lightweight, embodiment-specific module (System 1) then converts that predicted motion into executable robot actions, using either a hand-crafted mapping or one learned from a small number of demonstrations. The paper reports that this dual-system design outperforms prior vision-language-action baselines on the MetaWorld benchmark (57.7% average success) and works zero-shot on real-world tabletop tasks, arguing that motion is a more transferable currency than RGB images across embodiments.

What carries the argument

The load-bearing object is the dense pixel-motion field $y_{i,i+k} \in \mathbb{R}^{h \times w \times 2}$, the apparent motion of every pixel between the current frame $x_i$ and a future frame $x_{i+k}$, normalized to $(0,1)$ per channel. It is generated by a 2D conditional U-Net diffusion model whose input is a 7-channel tensor (current RGB image, previous motion field, and noised target motion) with language injected via a Universal Sentence Encoder embedding, and whose output is the clean 2-channel motion field. The hierarchy carries the argument: System 2 runs at sparse intervals ($k$ frames) to set the motion direction, and System 1 runs at dense intervals ($j<k$ steps) to convert that motion into actions, either through hand-crafted mappings that exploit interpretability or through a learned vision transformer.

What would settle it

Run LangToMo's hand-crafted variant on a task that requires 3D reasoning or contact, such as stacking objects of different heights, using only the single-plane calibration; if the predicted 2D flow cannot be converted into successful actions, the claim that dense 2D pixel motion suffices as a universal action representation fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that predicting pixel motion, rather than full RGB images, is the right way to connect language to action. Because pixel motion is defined in image space, it is independent of the robot's physical form; because it is dense, it captures both the manipulator and the object; and because it can be computed automatically with optical-flow algorithms such as RAFT, training data can be any video-caption collection. System 2 is a 2D conditional U-Net diffusion model that generates the next motion field from the current image, the previous motion field, and a sentence embedding, making the language-to-motion step a generative one-to-many mapping. System 1 then deterministically translates that motion into actions; the hand-crafted variant requires known geometry (segmentation and depth in simulation, or a calibrated plane in the real world), while the learned variant is a lightweight transformer trained on a few demonstrations. The authors claim this decomposition yields strong performance with less data, benefits from mixed human and robot demonstrations, and enables zero-shot control after pretraining on a large multi-embodiment video-caption dataset.

Load-bearing premise

The zero-shot and unsupervised claims rest on System 1 mappings that need privileged geometric information—ground-truth segmentation and depth in simulation, or a calibrated single-plane plus fixed-camera assumption in the real world—so if that geometry is unavailable, predicted 2D pixel motion cannot be converted into correct actions.

Editorial extensions

If this is right

  • Training data for the motion generator can be any video-caption collection, without action labels or pixel annotations, since RAFT computes the supervision signal automatically.
  • The same System 2 can be fine-tuned on both robot and human demonstrations, because pixel-motion distributions are more similar across embodiments than RGB appearance (symmetrized KL divergence 0.0199 vs 0.7881).
  • A single pretrained System 2 transfers zero-shot to a new tabletop environment with a hand-crafted System 1 mapping (33.8% average over four real-world tasks, versus 22.5% for the LLaRA baseline).
  • Decoupling the two systems at different frequencies reduces inference cost and allows independent training of each module, so the learned mapping needs only a modest number of demonstrations (e.g., 20 per task on MetaWorld).
  • On MetaWorld, the learned mapping variant reaches 57.7% average success across 11 tasks, surpassing the closest dense-motion baseline AVDC (43.1%) and approaches that predict point tracks instead of dense flow.

Reading between the lines

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

  • If dense flow from a single image and language proves as transferable as reported, neighboring problems—video captioning, human-robot imitation, and simulation-to-real transfer—could be reformulated as motion-forecasting tasks, reusing the same System 2 without retraining.
  • The paper leaves depth out; a testable extension is to condition on or predict 3D scene flow or object-centric depth, which would likely be needed for tasks like stacking or grasping in cluttered scenes.
  • The fixed-camera assumption limits the claim of universality; an explicit stress test is to add ego-motion videos to pretraining and measure whether zero-shot control degrades for navigation-like tasks, where the paper's iThor result (31.9% overall) suggests headroom.
  • One could also treat System 1 itself as a learned, embodiment-agnostic module that maps flow to actions across multiple robots, which would remove the per-embodiment calibration cost the paper names as an open challenge.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper presents LangToMo, a dual-system vision-language-action framework in which a high-level diffusion model (System 2) forecasts dense optical flow ('pixel motion') from a single RGB frame, a language instruction, and previous flow, using RAFT pseudo-labels from video-caption data with no action labels. A low-level System 1 then converts the predicted flow into robot actions, either through a learned ViT trained on 20 demonstrations per task (LTM-S) or through hand-crafted mappings (LTM-H) that use segmentation/depth in simulation and a calibrated single-plane model in the real world. The authors evaluate on 11 MetaWorld tasks, four real-world xArm tasks, CALVIN, and iThor, reporting LTM-S at 57.7% MetaWorld average, LTM-H at 52.1%, and real-world zero-shot and finetuned success rates of 33.8% and 71.3% respectively, with ablations over conditioning inputs, pretraining, diffusion vs regression, and system frequency.

Significance. If the results hold, the paper makes a useful empirical contribution: an actionless, video-caption-driven training objective for a motion-based intermediate representation that is interpretable and can be reused across embodiments, including human demonstrations. The evaluation is unusually broad (sim, real, ego-motion, long-horizon), and the ablation of System 2 inputs (image, language, previous flow, pretraining) is careful. The paper is also honest about limitations such as the cost of System 1 per embodiment and lack of depth cues. The main weaknesses are that the headline claims of universality and 'unsupervised' control are broader than the experimental protocol actually establishes, and some central comparisons lack variance information.

major comments (4)
  1. [Abstract, §1, §4.1, Appendix F] The claims that pixel motion enables 'even unsupervised control' and is a 'universal representation for robot control' are stronger than the experiments support. In simulation, LTM-H follows Ko et al. (2023) and requires ground-truth segmentations of the robot controller and a depth map; in the real world it relies on a single-plane assumption and per-environment camera calibration (Appendix F). These are privileged geometric signals, not properties of the predicted flow itself, so the paper has not demonstrated that pixel motion alone can be converted into actions across embodiments and scenes. Please qualify the abstract and introduction (e.g., 'universal intermediate representation with embodiment-specific action mappings') and, ideally, add an ablation using estimated depth/segmentation or no calibration to show how much of the pipeline works without these priors.
  2. [§4.2, Table 4] The claim that dense pixel motion is more useful than point tracks is not supported by the main comparison with ATM: LTM-H obtains 52.1% vs ATM's 52.0%, a 0.1-point difference, while LTM-S's 57.7% is achieved with 20 expert demonstrations per task. The paper should report per-task standard errors or confidence intervals across the 75 trials per task and either demonstrate the dense-vs-sparse advantage statistically or soften the claim to a qualitative one.
  3. [§4.1, Table 3] The AVDC RD+HD entry of 0.0 is an extreme result that is used to argue that pixel motion is more embodiment-agnostic than RGB. A zero success rate on 80 trials is difficult to interpret without diagnostics; please provide rollout videos or action statistics (e.g., whether the model collapsed to static predictions) to confirm the failure mode. If this baseline result is not reproducible, the 'benefits from human demonstrations' conclusion loses its main comparison point.
  4. [Limitations vs Appendix B.3] The Limitations paragraph states that 'we limit our training to fixed camera videos (no ego motion),' but Appendix B.3 reports iThor results after training LTM-H and AVDC 'on the same data under common training settings' in an environment with robot ego-motion. This is an apparent contradiction. Please clarify whether the iThor training data includes ego-motion and, if so, revise the limitation; if not, specify exactly what the iThor models were trained on.
minor comments (5)
  1. [§3.2, Eq. (1) and Training] The notation for the previous-flow conditioning is inconsistent: Eq. (1) uses y_{i-k,i}, while the Training paragraph writes y_{i-1,i}. Please use a single interval notation throughout.
  2. [Appendix B.2, Table 7] The column header 'ith Task Success Rate' should read 'i-th Task Success Rate'.
  3. [Appendix J] The KL-divergence analysis compares aggregate pixel-value histograms of 40 human and 40 robot demonstrations, which supports the intuition but does not directly measure whether the learned System 2 converges more easily on the two data types; the claim 'pixel motion is a more embodiment-agnostic metric' would be strengthened by reporting training loss or sample efficiency on HD vs RD data.
  4. [Abstract] 'Checkout' should be 'Check out' before the project link.
  5. [Table 6] The appendix reports control frequencies and episode counts but not the total number of caption annotations or the exact OpenX split used; please add these details for reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: LangToMo's success rates are measured on external benchmarks, System 2 is trained on RAFT-derived flow targets, and System 1 mappings are either learned from expert actions or explicitly stated hand-crafted geometric transforms.

full rationale

The paper's derivation chain is self-contained and externally benchmarked. System 2 is trained to predict pixel motion (optical flow) between frames using RAFT pseudo-labels derived from video-caption pairs; the same flow is then mapped to actions by System 1. The learned System 1 (LTM-S) is trained on ground-truth action trajectories, while the hand-crafted System 1 (LTM-H) is a deterministic geometric transform imported from prior work (Ko et al., 2023; Li et al., 2024). Neither stage defines its output in terms of the quantity it is supposed to predict: the success rates in Tables 2-4 are measured against benchmark tasks, not against the training targets or the mapping functions themselves. The real-world hand-crafted mapping does require a single-plane assumption and camera calibration, and the simulated mapping requires segmentation and depth maps, but this is a stated limitation of the system's scope rather than a circular reduction: the pixel-motion prediction must still be correct for the mapping to produce successful actions. Self-citations to Li et al. (2024) appear for environment construction and evaluation protocol, but they do not supply the load-bearing argument; they provide a benchmark recipe. The Limitations section explicitly concedes the cost of per-embodiment System 1 mappings and the absence of depth and ego-motion handling, which further confirms that no hidden fitted input is being renamed as a prediction. No equation or fitted parameter in the paper is equivalent, by construction, to the claimed result; therefore no circular step is present.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on a chain of domain assumptions: RAFT pseudo-labels as supervision, 2D flow sufficiency, fixed-camera data, and geometry-dependent hand-crafted mappings. No new physical or mathematical entities are introduced, and no quantities are fitted to the headline success rates.

free parameters (3)
  • Motion forecast horizon k = 10 (implied by '10 control steps' per System-2 invocation)
    Sets the sparse interval for System 2 and the dense interval for System 1; chosen by hand and not swept, yet central to the dual-frequency claim.
  • DDIM inference steps = 25
    Used for all System 2 inference; chosen by hand with no sensitivity analysis reported.
  • Previous-flow corruption probability = 0.5
    Training-time corruption of the previous-flow condition (Appendix E); chosen ad hoc to bridge the train and inference distributions.
assumptions (4)
  • domain assumption RAFT optical flow is a reliable target for action-relevant pixel motion.
    Section 3.2: 'We utilize the RAFT algorithm to calculate our target pixel motion'; all training supervision derives from this pseudo-label, including its acknowledged background-motion noise.
  • domain assumption Dense 2D pixel motion is sufficient to encode robot actions relevant to the benchmark tasks.
    Section 3.2: 'Our experiments indicate the sufficiency of such 2D spaces to encode motions relevant to robot actions'; 3D depth cues are deliberately excluded.
  • domain assumption Fixed-camera training videos without ego motion are sufficient for the motion generator.
    Limitations: 'we limit our training to fixed camera videos (no ego motion)'; the iThor results show only modest performance on ego-motion tasks.
  • domain assumption Hand-crafted mappings can reliably convert 2D flow into executable actions given privileged or calibrated geometry.
    Appendix F: simulation uses ground-truth segmentation and depth maps; real world uses a single-plane assumption and initial camera calibration.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pixel Motion as Universal Representation for Robot Control." pith.science (2026). https://pith.science/paper/WIDBK3J3

@misc{pith2026250507817,
  author       = {Pith},
  title        = {Pith review of: Pixel Motion as Universal Representation for Robot Control},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WIDBK3J3}},
  note         = {Machine review of arXiv:2505.07817}
}
read the original abstract

We present LangToMo, a vision-language-action framework structured as a dual-system architecture that uses pixel motion forecasts as intermediate representations. Our high-level System 2, an image diffusion model, generates text-conditioned pixel motion sequences from a single frame to guide robot control. Pixel motion-a universal, interpretable, and motion-centric representation-can be extracted from videos in a weakly-supervised manner, enabling diffusion model training on any video-caption data. Treating generated pixel motion as learned universal representations, our low level System 1 module translates these into robot actions via motion-to-action mapping functions, which can be either hand-crafted or learned with minimal supervision. System 2 operates as a high-level policy applied at sparse temporal intervals, while System 1 acts as a low-level policy at dense temporal intervals. This hierarchical decoupling enables flexible, scalable, and generalizable robot control under both unsupervised and supervised settings, bridging the gap between language, motion, and action. Checkout https://kahnchana.github.io/LangToMo

Figures

Figures reproduced from arXiv: 2505.07817 by the authors.

Figure 1
Figure 1. Dual-System VLA Framework, LangToMo, with pixel motion representations. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Dense Motions: Most prior work that use pixel trajectories focus on a subset of pixels often lim￾ited to objects of interest. The example from Xu et al. (2024) (left) focuses on the cup movement, but ignores important action information relevant to manipulator movement. In contrast, proposed LangToMo gener￾ates dense pixel motions that account for both object and manipulator movements (right). annotations, pretraine… view at source ↗
Figure 3
Figure 3. Overview of LangToMo: (Left) We learn to forecast pixel motion as universal motion features from video-caption pairs using scalable, self-supervised training of a diffusion model. (Right) Our System 2 forecasts motion at sparse intervals (k), while System 1 maps it to dense action vectors at j intervals (j < k). which are then decoded into robot actions. This dual-system architecture comprises: System 2, a condition… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: LangToMo Architecture: (Left) Diffusion model generates pixel motion conditioned on RGB image, prior motion, and caption. Visualized predictions are overlaid as arrows. (Right) ViT-T network maps predicted motion to robot actions in supervised setting, conditioned on i…
Figure 5
Figure 5. Figure 5: Human (HD) & Robot (RD) Demonstrations: We visualize frames from two sample demonstrations on our real world environment. Pixel motion overlaid on intermediate frames. These human (top) and robot (bottom) demonstrations can both be used to fine-tune our System-2, highl…
Figure 6
Figure 6. Figure 6: Semantic Awareness Visualization: We visualize outputs from our System-2 module (ours; left two figures) for two examples containing the same starting state (visual observation) but different action goals (textual command). LangToMo generates meaningful motions for sce…
Figure 7
Figure 7. Figure 7: Real World Tasks: We illustrate the four real-world tasks following LLaRA Li et al. (2024). Start and end states are shown in the first and last columns, with predicted pixel motion (color indicates motion direction) overlaid on intermediate states. LangToMo performs t…
Figure 8
Figure 8. Figure 8: Histogram Comparisons for RGB and Pixel Motion Distributions of Robot vs Human Demon￾strations: We illustrate two histograms that compare the aggregate pixel value distributions of 40 human and 40 robot demonstrations. For the RGB distributions (left), the high degree …

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. DynamicWAM: Dual-Path Motion Conditioning for World-Action Models in Dynamic Manipulation

    cs.RO 2026-08 conditional novelty 6.0 of 10

    A world-action model that conditions future prediction on optical-flow history and action generation on kinematic descriptors improves dynamic manipulation success rates in simulation and on a real robot.

Reference graph

Works this paper leans on

99 extracted references · 39 canonical work pages · cited by 1 Pith paper

  1. [1]

    write newline

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

  2. [2]

    Learning optical flow from still images

    Filippo Aleotti, Matteo Poggi, and Stefano Mattoccia. Learning optical flow from still images. In CVPR, pp.\ 15196--15206, 2021

  3. [3]

    Flowcontrol: Optical flow based visual servoing

    Max Argus, Luk \'a s Hermann, Jon Long, and Thomas Brox. Flowcontrol: Optical flow based visual servoing. 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp.\ 7534--7541, 2020. URL https://api.semanticscholar.org/CorpusID:220280145

  4. [4]

    Rheotropism in fishes

    Geoff Arnold. Rheotropism in fishes. Biological Reviews, 49, 1974. URL https://api.semanticscholar.org/CorpusID:30755969

  5. [5]

    Human-to-robot imitation in the wild

    Shikhar Bahl, Abhinav Gupta, and Deepak Pathak. Human-to-robot imitation in the wild. In Robotics: Science and Systems, 2022

  6. [6]

    Srinivasan

    Emily Baird, Norbert Boeddeker, and Mandyam V. Srinivasan. The effect of optic flow cues on honeybee flight control in wind. Proceedings of the Royal Society B, 288, 2021. URL https://api.semanticscholar.org/CorpusID:231643236

  7. [7]

    Rt-h: Action hierarchies using language

    Suneel Belkhale, Tianli Ding, Ted Xiao, Pierre Sermanet, Quon Vuong, Jonathan Tompson, Yevgen Chebotar, Debidatta Dwibedi, and Dorsa Sadigh. Rt-h: Action hierarchies using language. ArXiv, abs/2403.01823, 2024. URL https://api.semanticscholar.org/CorpusID:268249108

  8. [8]

    Gen2act: Human video generation in novel scenarios enables generalizable robot manipulation

    Homanga Bharadhwaj, Debidatta Dwibedi, Abhinav Gupta, Shubham Tulsiani, Carl Doersch, Ted Xiao, Dhruv Shah, Fei Xia, Dorsa Sadigh, and Sean Kirmani. Gen2act: Human video generation in novel scenarios enables generalizable robot manipulation. ArXiv, 2024 a

Show all 99 references
  1. [9]

    Track2act: Predicting point tracks from internet videos enables diverse zero-shot robot manipulation

    Homanga Bharadhwaj, Roozbeh Mottaghi, Abhinav Gupta, and Shubham Tulsiani. Track2act: Predicting point tracks from internet videos enables diverse zero-shot robot manipulation. ECCV, 2024 b

  2. [10]

    Zero-shot robotic manipulation with pretrained image-editing diffusion models

    Kevin Black, Mitsuhiko Nakamoto, Pranav Atreya, Homer Rich Walke, Chelsea Finn, Aviral Kumar, and Sergey Levine. Zero-shot robotic manipulation with pretrained image-editing diffusion models. ArXiv, abs/2310.10639, 2023. URL https://api.semanticscholar.org/CorpusID:264172455

  3. [11]

    0: A vision-language-action flow model for general robot control

    Kevin Black, Noah Brown, Danny Driess, Adnan Esmail, Michael Equi, Chelsea Finn, Niccolo Fusai, Lachy Groom, Karol Hausman, Brian Ichter, Szymon Jakubczak, Tim Jones, Liyiming Ke, Sergey Levine, Adrian Li-Bell, Mohith Mothukuri, Suraj Nair, Karl Pertsch, Lucy Xiaoyang Shi, Jam...

  4. [12]

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

    Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, Xi Chen, Krzysztof Choromanski, Tianli Ding, Danny Driess, Avinava Dubey, Chelsea Finn, et al. Rt-2: Vision-language-action models transfer web knowledge to robotic control. arXiv preprint arXiv:2307.15818, 2023 a

  5. [13]

    Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, Joseph Dabis, Chelsea Finn, Keerthana Gopalakrishnan, Karol Hausman, Alex Herzog, Jasmine Hsu, Julian Ibarz, Brian Ichter, Alex Irpan, Tomas Jackson, Sally Jesmonth, Nikhil J Joshi, Ryan Julian, Dmitry Kalashnikov,...

  6. [14]

    John, Noah Constant, Mario Guajardo-Cespedes, Steve Yuan, Chris Tar, Yun-Hsuan Sung, Brian Strope, and Ray Kurzweil

    Daniel Matthew Cer, Yinfei Yang, Sheng yi Kong, Nan Hua, Nicole Limtiaco, Rhomni St. John, Noah Constant, Mario Guajardo-Cespedes, Steve Yuan, Chris Tar, Yun-Hsuan Sung, Brian Strope, and Ray Kurzweil. Universal sentence encoder. ArXiv, 2018

  7. [15]

    Gr-3 technical report

    Chi-Lam Cheang, Sijin Chen, Zhongren Cui, Yingdong Hu, Liqun Huang, Tao Kong, Hang Li, Yifeng Li, Yuxiao Liu, Xiao Ma, Hao Niu, Wenxuan Ou, Wanli Peng, Zeyu Ren, Haixin Shi, Jiawen Tian, Hongtao Wu, Xin Xiao, Yuyang Xiao, Jiafeng Xu, and Yichu Yang. Gr-3 technical report. ArXi...

  8. [16]

    In-The-Wild

    Annie S Chen, Suraj Nair, and Chelsea Finn. Learning Generalizable Robotic Reward Functions from "In-The-Wild" Human Videos . In Robotics: Science and Systems, 2021

  9. [17]

    Moddm: Text-to-motion synthesis using discrete diffusion model

    Xin Chen, Yanchao Li, Zhen Li, Zhen Wang, Li Wang, and Chen Qian. Moddm: Text-to-motion synthesis using discrete diffusion model. arXiv preprint arXiv:2308.06240, 2023

  10. [18]

    Diffusion policy: Visuomotor policy learning via action diffusion

    Cheng Chi, Siyuan Feng, Yilun Du, Zhenjia Xu, Eric Cousineau, Benjamin Burchfiel, and Shuran Song. Diffusion policy: Visuomotor policy learning via action diffusion. ArXiv, abs/2303.04137, 2023

  11. [19]

    de Croon, Christophe de Wagter, and Tobias Seidl

    Guido C.H.E. de Croon, Christophe de Wagter, and Tobias Seidl. Enhancing optical-flow-based control by learning visual appearance cues for flying robots. Nature Machine Intelligence, 3: 0 33 -- 41, 2021. URL https://api.semanticscholar.org/CorpusID:231655448

  12. [20]

    Diffusion Models Beat GANs on Image Synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion Models Beat GANs on Image Synthesis . In Neural Information Processing Systems, 2021

  13. [21]

    Palm-e: An embodied multimodal language model

    Danny Driess, Fei Xia, Mehdi SM Sajjadi, Corey Lynch, Aakanksha Chowdhery, Brian Ichter, Ayzaan Wahid, Jonathan Tompson, Quan Vuong, Tianhe Yu, et al. Palm-e: An embodied multimodal language model. arXiv preprint arXiv:2303.03378, 2023

  14. [22]

    Reduce, Reuse, Recycle: Compositional Generation with Energy-Based Diffusion Models and MCMC

    Yilun Du, Conor Durkan, Robin Strudel, Joshua B Tenenbaum, Sander Dieleman, Rob Fergus, Jascha Sohl-Dickstein, Arnaud Doucet, and Will Grathwohl. Reduce, Reuse, Recycle: Compositional Generation with Energy-Based Diffusion Models and MCMC . In International Conference on Machi...

  15. [23]

    Learning Universal Policies via Text-Guided Video Generation

    Yilun Du, Mengjiao Yang, Bo Dai, Hanjun Dai, Ofir Nachum, Joshua B Tenenbaum, Dale Schuurmans, and Pieter Abbeel. Learning Universal Policies via Text-Guided Video Generation . arXiv:2302.00111, 2023 b

  16. [24]

    Deep Visual Foresight for Planning Robot Motion

    Chelsea Finn and Sergey Levine. Deep Visual Foresight for Planning Robot Motion . In IEEE International Conference on Robotics and Automation, 2017

  17. [25]

    Flip: Flow-centric generative planning as general-purpose manipulation world model

    Chongkai Gao, Haozhuo Zhang, Zhixuan Xu, Zhehao Cai, and Lin Shao. Flip: Flow-centric generative planning as general-purpose manipulation world model. In ICLR, 2025

  18. [26]

    Im2flow: Motion hallucination from static images for action recognition

    Ruohan Gao, Bo Xiong, and Kristen Grauman. Im2flow: Motion hallucination from static images for action recognition. In CVPR, pp.\ 5937--5947, 2018

  19. [27]

    Long video generation with time-agnostic vqgan and time-sensitive transformer, 2022

    Songwei Ge, Thomas Hayes, Harry Yang, Xi Yin, Guan Pang, David Jacobs, Jia-Bin Huang, and Devi Parikh. Long video generation with time-agnostic vqgan and time-sensitive transformer, 2022

  20. [28]

    Flight control in drosophila by visual perception of motion

    Karl Georg G \"o tz. Flight control in drosophila by visual perception of motion. Kybernetik, 4: 0 199--208, 1968. URL https://api.semanticscholar.org/CorpusID:24070951

  21. [29]

    Seer: Language instructed video prediction with latent diffusion models

    Xianfan Gu, Chuan Wen, Jiaming Song, and Yang Gao. Seer: Language instructed video prediction with latent diffusion models. ArXiv, abs/2303.14897, 2023. URL https://api.semanticscholar.org/CorpusID:257766959

  22. [30]

    Self-supervised co-training for video representation learning

    Tengda Han, Weidi Xie, and Andrew Zisserman. Self-supervised co-training for video representation learning. ArXiv, abs/2010.09709, 2020. URL https://api.semanticscholar.org/CorpusID:224703413

  23. [31]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33: 0 6840--6851, 2020

  24. [32]

    Video Diffusion Models

    Jonathan Ho, Tim Salimans, Alexey Gritsenko, William Chan, Mohammad Norouzi, and David J Fleet. Video Diffusion Models . In Neural Information Processing Systems, 2022

  25. [33]

    Seeing through pixel motion: Learning obstacle avoidance from optical flow with one camera

    Yu Hu, Yuang Zhang, Yunlong Song, Yang Deng, Feng Yu, Linzuo Zhang, Weiyao Lin, Danping Zou, and Wenxian Yu. Seeing through pixel motion: Learning obstacle avoidance from optical flow with one camera. ArXiv, abs/2411.04413, 2024. URL https://api.semanticscholar.org/CorpusID:273877940

  26. [34]

    Video prediction policy: A generalist robot policy with predictive visual representations

    Yucheng Hu, Yanjiang Guo, Pengchao Wang, Xiaoyu Chen, Yen-Jen Wang, Jianke Zhang, Koushil Sreenath, Chaochao Lu, and Jianyu Chen. Video prediction policy: A generalist robot policy with predictive visual representations. ICML, 2025

  27. [35]

    Rekep: Spatio-temporal reasoning of relational keypoint constraints for robotic manipulation

    Wenlong Huang, Chen Wang, Yunzhu Li, Ruohan Zhang, and Fei-Fei Li. Rekep: Spatio-temporal reasoning of relational keypoint constraints for robotic manipulation. ArXiv, 2024

  28. [36]

    Physical Intelligence, Kevin Black, Noah Brown, James Darpinian, Karan Dhabalia, Danny Driess, Adnan Esmail, Michael Equi, Chelsea Finn, Niccolo Fusai, Manuel Y. Galliker, Dibya Ghosh, Lachy Groom, Karol Hausman, Brian Ichter, Szymon Jakubczak, Tim Jones, Liyiming Ke, Devin Le...

  29. [37]

    Tenenbaum, and Sergey Levine

    Michael Janner, Yilun Du, Joshua B. Tenenbaum, and Sergey Levine. Planning with Diffusion for Flexible Behavior Synthesis . In International Conference on Machine Learning, 2022

  30. [38]

    Object-centric world model for language-guided manipulation

    Youngjoon Jeong, Junha Chun, Soonwoo Cha, and Taesup Kim. Object-centric world model for language-guided manipulation. ArXiv, abs/2503.06170, 2025. URL https://api.semanticscholar.org/CorpusID:276903201

  31. [39]

    Thinking, fast and slow, 2011

    Daniel Kahneman. Thinking, fast and slow, 2011

  32. [40]

    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

  33. [41]

    Learning to act from actionless videos through dense correspondences

    Po-Chen Ko, Jiayuan Mao, Yilun Du, Shao-Hua Sun, and Josh Tenenbaum. Learning to act from actionless videos through dense correspondences. ArXiv, abs/2310.08576, 2023

  34. [42]

    Onlyflow: Optical flow based motion conditioning for video diffusion models, 2024

    Mathis Koroglu, Hugo Caselles-Dupr'e, Guillaume Jeanneret Sanmiguel, and Matthieu Cord. Onlyflow: Optical flow based motion conditioning for video diffusion models, 2024

  35. [43]

    Multi-concept customization of text-to-image diffusion, 2023

    Nupur Kumari, Bingliang Zhang, Richard Zhang, Eli Shechtman, and Jun-Yan Zhu. Multi-concept customization of text-to-image diffusion, 2023

  36. [44]

    Learning Plannable Representations with Causal InfoGAN

    Thanard Kurutach, Aviv Tamar, Ge Yang, Stuart J Russell, and Pieter Abbeel. Learning Plannable Representations with Causal InfoGAN . In Neural Information Processing Systems, 2018

  37. [45]

    Learning Robot Activities from First-Person Human Videos Using Convolutional Future Regression

    Jangwon Lee and Michael S Ryoo. Learning Robot Activities from First-Person Human Videos Using Convolutional Future Regression . In CVPRW, 2017

  38. [46]

    Molmoact: Action reasoning models that can reason in space

    Jason Lee, Jiafei Duan, Haoquan Fang, Yuquan Deng, Shuo Liu, Boyang Li, Bohan Fang, Jieyu Zhang, Yi Ru Wang, Sangho Lee, Winson Han, Wilbert Pumacay, Angelica Wu, Rose Hendrix, Karen Farley, Eli VanderBilt, Ali Farhadi, Dieter Fox, and Ranjay Krishna. Molmoact: Action reasonin...

  39. [47]

    Theodorou

    Keuntaek Lee, Jason Gibson, and Evangelos A. Theodorou. Aggressive perception-aware navigation using deep optical flow dynamics and pixelmpc. IEEE Robotics and Automation Letters, 5: 0 1207--1214, 2020. URL https://api.semanticscholar.org/CorpusID:210064565

  40. [48]

    Xiang Li, Cristina Mata, Jong Sung Park, Kumara Kahatapitiya, Yoo Sung Jang, Jinghuan Shang, Kanchana Ranasinghe, Ryan Burgert, Mu Cai, Yong Jae Lee, and Michael S. Ryoo. Llara: Supercharging robot learning data for vision-language policy. ArXiv, abs/2406.20095, 2024

  41. [49]

    Movideo: Motion-aware video generation with diffusion model

    Jingyun Liang, Yuchen Fan, Kai Zhang, Radu Timofte, Luc van Gool, and Rakesh Ranjan. Movideo: Motion-aware video generation with diffusion model. In European Conference on Computer Vision, 2024. URL https://api.semanticscholar.org/CorpusID:273232410

  42. [50]

    Selflow: Self-supervised learning of optical flow

    Pengpeng Liu, Michael Lyu, Irwin King, and Jia Xu. Selflow: Self-supervised learning of optical flow. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 4571--4580, 2019

  43. [51]

    StructDiffusion: Language-Guided Creation of Physically-Valid Structures using Unseen Objects

    Weiyu Liu, Yilun Du, Tucker Hermans, Sonia Chernova, and Chris Paxton. StructDiffusion: Language-Guided Creation of Physically-Valid Structures using Unseen Objects . In Robotics: Science and Systems, 2023

  44. [52]

    Flowdiffuser: Advancing optical flow estimation with diffusion models

    Ao Luo, Xin Li, Fan Yang, Jiangyu Liu, Haoqiang Fan, and Shuaicheng Liu. Flowdiffuser: Advancing optical flow estimation with diffusion models. 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.\ 19167--19176, 2024

  45. [53]

    Calvin: A benchmark for language-conditioned policy learning for long-horizon robot manipulation tasks

    Oier Mees, Luk \'a s Hermann, Erick Rosete-Beas, and Wolfram Burgard. Calvin: A benchmark for language-conditioned policy learning for long-horizon robot manipulation tasks. IEEE Robotics and Automation Letters, 7: 0 7327--7334, 2021

  46. [54]

    R3M: A Universal Visual Representation for Robot Manipulation

    Suraj Nair, Aravind Rajeswaran, Vikash Kumar, Chelsea Finn, and Abhinav Gupta. R3M: A Universal Visual Representation for Robot Manipulation . In Conference on Robot Learning, 2022

  47. [55]

    Llarva: Vision-action instruction tuning enhances robot learning

    Dantong Niu, Yuvan Sharma, Giscard Biamby, Jerome Quenum, Yutong Bai, Baifeng Shi, Trevor Darrell, and Roei Herzig. Llarva: Vision-action instruction tuning enhances robot learning. arXiv preprint arXiv:2406.11815, 2024

  48. [56]

    Gr00t n1: An open foundation model for generalist humanoid robots

    Nvidia, Johan Bjorck, et al. Gr00t n1: An open foundation model for generalist humanoid robots. ArXiv, abs/2503.14734, 2025

  49. [57]

    Octo: An open-source generalist robot policy

    Octo Model Team , Dibya Ghosh, Homer Walke, Karl Pertsch, Kevin Black, Oier Mees, Sudeep Dasari, Joey Hejna, Charles Xu, Jianlan Luo, Tobias Kreiman, You Liang Tan, Lawrence Yunliang Chen, Pannag Sanketi, Quan Vuong, Ted Xiao, Dorsa Sadigh, Chelsea Finn, and Sergey Levine. Oct...

  50. [58]

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

    Abhishek Padalkar, Acorn Pooley, Ajinkya Jain, Alex Bewley, Alex Herzog, Alex Irpan, Alexander Khazatsky, Anant Rai, Anikait Singh, Anthony Brohan, et al. Open x-embodiment: Robotic learning datasets and rt-x models. arXiv preprint arXiv:2310.08864, 2023

  51. [59]

    The Surprising Effectiveness of Representation Learning for Visual Imitation

    Jyothish Pari, Nur Muhammad Shafiullah, Sridhar Pandian Arunachalam, and Lerrel Pinto. The Surprising Effectiveness of Representation Learning for Visual Imitation . In Robotics: Science and Systems, 2022

  52. [60]

    Language models are unsupervised multitask learners, 2019

    Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. Language models are unsupervised multitask learners, 2019

  53. [61]

    Hierarchical text-conditional image generation with clip latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image generation with clip latents. preprint, 2022. [arxiv:2204.06125]

  54. [62]

    A generalist agent

    Scott Reed, Konrad Zolna, Emilio Parisotto, Sergio Gomez Colmenarejo, Alexander Novikov, Gabriel Barth-Maron, Mai Gimenez, Yury Sulsky, Jackie Kay, Jost Tobias Springenberg, Tom Eccles, Jake Bruce, Ali Razavi, Ashley Edwards, Nicolas Heess, Yutian Chen, Raia Hadsell, Oriol Vin...

  55. [63]

    Motion tracks: A unified representation for human-robot transfer in few-shot imitation learning

    Juntao Ren, Priya Sundaresan, Dorsa Sadigh, Sanjiban Choudhury, and Jeannette Bohg. Motion tracks: A unified representation for human-robot transfer in few-shot imitation learning. ArXiv, abs/2501.06994, 2025. URL https://api.semanticscholar.org/CorpusID:275471722

  56. [64]

    Diffusion motion: Generate text-guided 3d human motion by diffusion model

    Zhiyuan Ren, Zhihong Pan, Xin Zhou, and Le Kang. Diffusion motion: Generate text-guided 3d human motion by diffusion model. arXiv preprint arXiv:2210.12315, 2022

  57. [65]

    U-net: Convolutional networks for biomedical image segmentation, 2015

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation, 2015

  58. [66]

    Ros and Andrew A

    Ivo G. Ros and Andrew A. Biewener. Optic flow stabilizes flight in ruby-throated hummingbirds. Journal of Experimental Biology, 219: 0 2443 -- 2448, 2016. URL https://api.semanticscholar.org/CorpusID:11106817

  59. [67]

    Saurabh Saxena, Charles Herrmann, Junhwa Hur, Abhishek Kar, Mohammad Norouzi, Deqing Sun, and David J. Fleet. The surprising effectiveness of diffusion models for optical flow and monocular depth estimation. ArXiv, abs/2306.01923, 2023

  60. [68]

    Concept2Robot: Learning Manipulation Concepts from Instructions and Human Demonstrations

    Lin Shao, Toki Migimatsu, Qiang Zhang, Karen Yang, and Jeannette Bohg. Concept2Robot: Learning Manipulation Concepts from Instructions and Human Demonstrations . IJRR, 2021

  61. [69]

    Third-person visual imitation learning via decoupled hierarchical controller

    Pratyusha Sharma, Deepak Pathak, and Abhinav Gupta. Third-person visual imitation learning via decoupled hierarchical controller. In Neural Information Processing Systems, 2019

  62. [70]

    Pixel-level correspondence for self-supervised learning from video

    Yash Sharma, Yi Zhu, Chris Russell, and Thomas Brox. Pixel-level correspondence for self-supervised learning from video. ArXiv, abs/2207.03866, 2022. URL https://api.semanticscholar.org/CorpusID:250407930

  63. [71]

    Zeromimic: Distilling robotic manipulation skills from web videos, 2025 a

    Junyao Shi, Zhuolun Zhao, Tianyou Wang, Ian Pedroza, Amy Luo, Jie Wang, Jason Ma, and Dinesh Jayaraman. Zeromimic: Distilling robotic manipulation skills from web videos, 2025 a

  64. [72]

    Hi robot: Open-ended instruction following with hierarchical vision-language-action models

    Lucy Xiaoyang Shi, Brian Ichter, Michael Equi, Liyiming Ke, Karl Pertsch, Quan Vuong, James Tanner, Anna Walling, Haohuan Wang, Niccolo Fusai, Adrian Li-Bell, Danny Driess, Lachy Groom, Sergey Levine, and Chelsea Finn. Hi robot: Open-ended instruction following with hierarchic...

  65. [73]

    Generative image as action models

    Mohit Shridhar, Yat Long Lo, and Stephen James. Generative image as action models. ArXiv, abs/2407.07875, 2024

  66. [74]

    Make-a-video: Text-to-video generation without text-video data, 2022

    Uriel Singer, Adam Polyak, Thomas Hayes, Xi Yin, Jie An, Songyang Zhang, Qiyuan Hu, Harry Yang, Oron Ashual, Oran Gafni, Devi Parikh, Sonal Gupta, and Yaniv Taigman. Make-a-video: Text-to-video generation without text-video data, 2022

  67. [75]

    Robotic Telekinesis: Learning a Robotic Hand Imitator by Watching Humans on Youtube

    Aravind Sivakumar, Kenneth Shaw, and Deepak Pathak. Robotic Telekinesis: Learning a Robotic Hand Imitator by Watching Humans on Youtube . In Robotics: Science and Systems, 2022

  68. [76]

    History-guided video diffusion, 2025

    Kiwhan Song, Boyuan Chen, Max Simchowitz, Yilun Du, Russ Tedrake, and Vincent Sitzmann. History-guided video diffusion, 2025

  69. [77]

    Controlling the world by sleight of hand

    Sruthi Sudhakar, Ruoshi Liu, Basile Van Hoorick, Carl Vondrick, and Richard Zemel. Controlling the world by sleight of hand. ArXiv, abs/2408.07147, 2024

  70. [78]

    Neural program synthesis from diverse demonstration videos

    Shao-Hua Sun, Hyeonwoo Noh, Sriram Somasundaram, and Joseph Lim. Neural program synthesis from diverse demonstration videos. In International Conference on Machine Learning, 2018

  71. [79]

    Raft: Recurrent all-pairs field transforms for optical flow

    Zachary Teed and Jia Deng. Raft: Recurrent all-pairs field transforms for optical flow. In European Conference on Computer Vision, pp.\ 402--419. Springer, 2020

  72. [80]

    Predictive inverse dynamics models are scalable learners for robotic manipulation

    Yang Tian, Sizhe Yang, Jia Zeng, Ping Wang, Dahua Lin, Hao Dong, and Jiangmiao Pang. Predictive inverse dynamics models are scalable learners for robotic manipulation. ArXiv, abs/2412.15109, 2024. URL https://api.semanticscholar.org/CorpusID:274859727

  73. [81]

    Phenaki: Variable length video generation from open domain textual description, 2022

    Ruben Villegas, Mohammad Babaeizadeh, Pieter-Jan Kindermans, Hernan Moraldo, Han Zhang, Mohammad Taghi Saffar, Santiago Castro, Julius Kunze, and Dumitru Erhan. Phenaki: Variable length video generation from open domain textual description, 2022

  74. [82]

    Dense optical flow prediction from a static image

    Jacob Walker, Abhinav Kumar Gupta, and Martial Hebert. Dense optical flow prediction from a static image. In ICCV, pp.\ 2443--2451, 2015

  75. [83]

    Diffusion Model-Augmented Behavioral Cloning

    Hsiang-Chun Wang, Shang-Fu Chen, and Shao-Hua Sun. Diffusion Model-Augmented Behavioral Cloning . arXiv:2302.13335, 2023

  76. [84]

    Any-point trajectory modeling for policy learning

    Chuan Wen, Xingyu Lin, John So, Kai Chen, Qi Dou, Yang Gao, and Pieter Abbeel. Any-point trajectory modeling for policy learning. ArXiv, 2023

  77. [85]

    Unleashing large-scale video generative pre-training for visual robot manipulation

    Hongtao Wu, Ya Jing, Chilam Cheang, Guangzeng Chen, Jiafeng Xu, Xinghang Li, Minghuan Liu, Hang Li, and Tao Kong. Unleashing large-scale video generative pre-training for visual robot manipulation. arXiv preprint arXiv:2312.13139, 2023

  78. [86]

    GMFlow: Learning Optical Flow via Global Matching

    Haofei Xu, Jing Zhang, Jianfei Cai, Hamid Rezatofighi, and Dacheng Tao. GMFlow: Learning Optical Flow via Global Matching . In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022

  79. [87]

    Flow as the cross-domain manipulation interface

    Mengda Xu, Zhenjia Xu, Yinghao Xu, Cheng Chi, Gordon Wetzstein, Manuela Veloso, and Shuran Song. Flow as the cross-domain manipulation interface. In Conference on Robot Learning, 2024

  80. [88]

    Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning

    Tianhe Yu, Deirdre Quillen, Zhanpeng He, Ryan Julian, Karol Hausman, Chelsea Finn, and Sergey Levine. Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning . In Conference on Robot Learning, 2019

  81. [89]

    General flow as foundation affordance for scalable robot learning

    Chengbo Yuan, Chuan Wen, Tong Zhang, and Yang Gao. General flow as foundation affordance for scalable robot learning. arXiv preprint arXiv:2401.11439, 2024 a

  82. [90]

    Robopoint: A vision-language model for spatial affordance prediction for robotics

    Wentao Yuan, Jiafei Duan, Valts Blukis, Wilbert Pumacay, Ranjay Krishna, Adithyavairavan Murali, Arsalan Mousavian, and Dieter Fox. Robopoint: A vision-language model for spatial affordance prediction for robotics. arXiv preprint arXiv:2406.10721, 2024 b

  83. [91]

    Robotic control via embodied chain-of-thought reasoning

    Michal Zawalski, William Chen, Karl Pertsch, Oier Mees, Chelsea Finn, and Sergey Levine. Robotic control via embodied chain-of-thought reasoning. In Conference on Robot Learning, 2024. URL https://api.semanticscholar.org/CorpusID:271097636

  84. [92]

    Adding conditional control to text-to-image diffusion models, 2023

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models, 2023

  85. [93]

    Motiondiffuse: Text-driven human motion generation with diffusion model

    Mingyuan Zhang, Zhongang Cai, Liang Pan, Fangzhou Hong, Xinying Guo, Lei Yang, and Ziwei Liu. Motiondiffuse: Text-driven human motion generation with diffusion model. arXiv preprint arXiv:2208.15001, 2022

  86. [94]

    Dreamvla: A vision-language-action model dreamed with comprehensive world knowledge

    Wenyao Zhang, Hongsi Liu, Zekun Qi, Yunnan Wang, Xinqiang Yu, Jiazhao Zhang, Runpei Dong, Jiawei He, He Wang, Zhizheng Zhang, Li Yi, Wenjun Zeng, and Xin Jin. Dreamvla: A vision-language-action model dreamed with comprehensive world knowledge. ArXiv, abs/2507.04447, 2025. URL ...

  87. [95]

    Universal actions for enhanced embodied foundation models

    Jinliang Zheng, Jianxiong Li, Dongxiu Liu, Yinan Zheng, Zhihao Wang, Zhonghong Ou, Yu Liu, Jingjing Liu, Ya-Qin Zhang, and Xianyuan Zhan. Universal actions for enhanced embodied foundation models. ArXiv, abs/2501.10105, 2025. URL https://api.semanticscholar.org/CorpusID:275606605

  88. [96]

    Tracevla: Visual trace prompting enhances spatial-temporal awareness for generalist robotic policies

    Ruijie Zheng, Yongyuan Liang, Shuaiyi Huang, Jianfeng Gao, Hal Daum \'e III, Andrey Kolobov, Furong Huang, and Jianwei Yang. Tracevla: Visual trace prompting enhances spatial-temporal awareness for generalist robotic policies. arXiv preprint arXiv:2412.10345, 2024

  89. [97]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  90. [98]

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

  91. [99]

    ) leads to a modest performance drop (from 53.6\ Removing the previous optical flow input (``Prev Flow

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.