pith. sign in

arxiv: 2605.20811 · v1 · pith:I67PTHOPnew · submitted 2026-05-20 · 💻 cs.RO

Demo-JEPA: Joint-Embedding Predictive Architecture for One-shot Cross-Embodiment Imitation

Jingyang He , Guangrun Li , Jieyu Zhang , Chengkai Hou , Zhengping Che , Shanghang Zhang This is my paper

Pith reviewed 2026-05-21 04:41 UTC · model grok-4.3

classification 💻 cs.RO
keywords cross-embodiment imitationjoint-embedding predictive architectureimitation learningworld modelslatent trajectoriesrobotic manipulationone-shot learninggoal inference
0
0 comments X

The pith

A JEPA world model turns visual demonstrations into latent subgoals that any robot body can plan toward.

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

The paper reframes imitation as inferring intended future states from visuals rather than copying actions that depend on body shape or controls. A JEPA-based world model trained on the target robot's own experience creates a shared space where source demonstrations become future latent trajectories. The target robot treats those trajectories as subgoals and reaches them by planning with its own forward model. This matters because it removes requirements for matching action spaces or training across many robot types at once, needing only one visual demo plus the target's self-collected data. Experiments on RLBench and real manipulation tasks show the approach matches specialized planners while succeeding on unseen tasks and body changes where earlier methods break.

Core claim

Demo-JEPA translates source visual demonstrations into target-compatible future latent trajectories in a shared predictive representation space. The target agent then uses these latent trajectories as subgoals and realizes them through planning under its own learned forward dynamics. Because Demo-JEPA avoids action-level correspondence and requires only visual demonstrations plus the target agent's own interaction experience, it supports flexible imitation across heterogeneous embodiments.

What carries the argument

The JEPA-based world model that builds a shared predictive representation space to convert visual demonstrations into future latent trajectories usable as subgoals by the target agent.

If this is right

  • Imitation succeeds without shared action spaces, retargeting, or multi-embodiment co-training.
  • The method generalizes to unseen tasks and new embodiment configurations.
  • Performance matches specialized in-domain planners on RLBench and real-world manipulation tasks.
  • Only visual demonstrations and the target agent's own experience are required.

Where Pith is reading between the lines

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

  • The same latent space might let a single world model translate goals between agents that differ even more, such as from human hands to robotic grippers.
  • Predictive models could become standard translators for goal inference in multi-robot teams with mismatched sensors.
  • Further tests with extreme morphology gaps would show where the shared representation starts to lose intent information.

Load-bearing premise

A world model trained primarily on the target agent's interactions can still produce a latent space that correctly captures the intent behind demonstrations from other embodiments.

What would settle it

Run a controlled test where the visual demonstration shows one clear goal but the translated latent trajectory leads the target robot to a different outcome; consistent failure to match the demonstrator's intent would falsify the shared-space claim.

Figures

Figures reproduced from arXiv: 2605.20811 by Chengkai Hou, Guangrun Li, Jieyu Zhang, Jingyang He, Shanghang Zhang, Zhengping Che.

Figure 1
Figure 1. Figure 1: Overview of Demo-JEPA. Demo-JEPA performs cross-embodiment imitation in a JEPA latent space, where the Dreamer Predictor infers target-compatible goals from source demonstrations for planning. We evaluate it across three increasingly shifted suites: Behavior Grounding for seen tasks, Cross-Embodiment Bridging for unseen actions, and Zero-Shot Generalization for unseen con￾figurations. Demo-JEPA achieves la… view at source ↗
Figure 2
Figure 2. Figure 2: Demo-JEPA training and inference pipeline. The top panels show the overall training and inference stages, from target world-model initialization to closed-loop planning with adaptive goal updating. The bottom panel highlights the Dreamer Predictor, which uses JEPA latents, cross￾attention, and 3D convolutional fusion to translate source demonstrations into target-compatible future latent goals. 3.1 Problem… view at source ↗
Figure 3
Figure 3. Figure 3: From demo to policy execution. Policy takes demo as input and executes action step by step. Top: real world deployment. Bottom: simulation deployment [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Real-world Tasks. We visualizes the progression of six real-world manipulation tasks. Task Definitions. To evaluate diverse facets of robotic perception and control, we define six representative real-world manipulation tasks: (i) Lift cup requires a stable rim-grasp on a pink cup followed by vertical elevation; (ii) Lift cube involves the precise picking of a red cube within the operational space; (iii) Re… view at source ↗
Figure 5
Figure 5. Figure 5: Real world experiment environment setup. Franka workspace(left) and UR5e workspace(right). As shown in [PITH_FULL_IMAGE:figures/full_fig_p019_5.png] view at source ↗
read the original abstract

Robotic imitation learning is often treated as reproducing demonstrated actions, but actions are inherently embodiment-specific. When demonstrations come from humans or robots with different morphology, kinematics, or action spaces, this action-centric view requires shared action spaces, heuristic retargeting, or large-scale multi-embodiment co-training. We instead view demonstrations as implicit specifications of future goals: the target agent should infer what state the demonstrator is trying to realize, rather than how the demonstrator executes it. We propose Demo-JEPA, a cross-embodiment imitation framework that decouples demonstration intent from embodiment-specific execution. Built on a JEPA-based world model, Demo-JEPA translates source visual demonstrations into target-compatible future latent trajectories in a shared predictive representation space. The target agent then uses these latent trajectories as subgoals and realizes them through planning under its own learned forward dynamics. Because Demo-JEPA avoids action-level correspondence and requires only visual demonstrations plus the target agent's own interaction experience, it supports flexible imitation across heterogeneous embodiments. Experiments on RLBench and real-world manipulation tasks show that Demo-JEPA matches specialized in-domain planners and generalizes to unseen tasks and embodiment configurations where prior methods fail.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 2 minor

Summary. The manuscript introduces Demo-JEPA, a cross-embodiment imitation framework built on a JEPA world model. Source visual demonstrations are encoded into a shared predictive latent space to produce future trajectories that serve as subgoals; the target agent then plans to realize these trajectories using its own learned forward dynamics. The approach requires only visual demonstrations and the target agent's interaction data, avoiding action retargeting or multi-embodiment co-training. Experiments on RLBench and real-world manipulation are reported to match in-domain planners while generalizing to unseen tasks and embodiment variations.

Significance. If the central mechanism holds, the work offers a principled way to separate intent inference from embodiment-specific execution in imitation learning. The reliance on a predictive JEPA representation rather than direct action matching is a conceptual strength, and the one-shot setting with only target self-interaction data could reduce data requirements compared with prior cross-embodiment methods. Reproducible code or explicit falsifiable predictions are not mentioned in the provided text.

major comments (2)
  1. [Method] Method section (description of JEPA training and inference): the claim that a JEPA encoder and predictor trained exclusively on target-agent interaction data produces a latent space in which source demonstration frames yield realizable target trajectories rests on an untested invariance assumption. No contrastive loss, domain-adaptation term, or explicit cross-embodiment alignment is described; therefore the translation step may map visually dissimilar source frames to latents whose predicted futures are unreachable under the target dynamics. This assumption is load-bearing for the central cross-embodiment claim.
  2. [Experiments] Experiments section (quantitative results and ablations): the abstract asserts that Demo-JEPA matches specialized in-domain planners and generalizes where prior methods fail, yet no success rates, baseline comparisons, ablation studies on the predictive component, or metrics for latent-space alignment across embodiments are supplied in the available text. Without these, the empirical support for generalization cannot be evaluated.
minor comments (2)
  1. [Method] Notation for the latent trajectory and subgoal extraction should be defined explicitly with equations rather than prose only.
  2. [Experiments] Figure captions for any qualitative rollout visualizations should include embodiment labels and camera viewpoints to clarify cross-embodiment differences.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive review and for identifying key areas where the manuscript could be strengthened. We address each major comment below with clarifications on the method and plans to improve the experimental presentation.

read point-by-point responses

  1. Referee: [Method] Method section (description of JEPA training and inference): the claim that a JEPA encoder and predictor trained exclusively on target-agent interaction data produces a latent space in which source demonstration frames yield realizable target trajectories rests on an untested invariance assumption. No contrastive loss, domain-adaptation term, or explicit cross-embodiment alignment is described; therefore the translation step may map visually dissimilar source frames to latents whose predicted futures are unreachable under the target dynamics. This assumption is load-bearing for the central cross-embodiment claim.

    Authors: We thank the referee for this observation. The JEPA world model is trained exclusively on the target agent's self-supervised interaction data using a predictive objective that learns to forecast future latent states. This training encourages the encoder to produce representations focused on task-relevant dynamics rather than embodiment-specific visual features, as the loss penalizes inaccurate future predictions under the target's own actions. Consequently, when source demonstration frames are encoded into this space, the resulting latent trajectories correspond to future states that the target can realize by planning with its learned dynamics model. No explicit alignment term is included by design, since the approach avoids requiring paired cross-embodiment data. We will revise the method section to expand on this rationale, including why the predictive (rather than reconstructive) objective supports the observed cross-embodiment generalization. revision: partial

  2. Referee: [Experiments] Experiments section (quantitative results and ablations): the abstract asserts that Demo-JEPA matches specialized in-domain planners and generalizes where prior methods fail, yet no success rates, baseline comparisons, ablation studies on the predictive component, or metrics for latent-space alignment across embodiments are supplied in the available text. Without these, the empirical support for generalization cannot be evaluated.

    Authors: We agree that the quantitative support should be presented more explicitly. While the manuscript reports results on RLBench and real-world tasks showing performance comparable to in-domain methods and superior generalization, we will revise the experiments section to include a dedicated table of success rates, direct numerical comparisons to baselines (such as behavior cloning and other cross-embodiment approaches), an ablation isolating the contribution of the predictive JEPA component, and metrics evaluating latent trajectory consistency across embodiments. These additions will make the empirical claims fully evaluable. revision: yes

Circularity Check

0 steps flagged

No significant circularity in Demo-JEPA derivation chain

full rationale

The paper introduces Demo-JEPA as a framework that trains a JEPA world model on target-agent interaction data, encodes source visual demonstrations into the resulting latent space, and uses predicted trajectories as subgoals for planning under the target's dynamics. No equations, fitting procedures, or self-citations are described that would reduce the central translation claim to a quantity defined by the same inputs or prior author work. The approach rests on an empirical assumption about representation invariance rather than any definitional or constructional equivalence between prediction and training data. The derivation from world-model training to cross-embodiment subgoal realization remains self-contained and independent of the target result.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach rests on the domain assumption that predictive latent representations learned from target-agent interactions can serve as a neutral bridge for demonstrator intent.

axioms (1)
  • domain assumption A JEPA-style world model can learn a shared predictive representation space that captures future states independently of embodiment-specific details.
    This is required for the translation step to produce usable subgoals for the target agent.

pith-pipeline@v0.9.0 · 5755 in / 1156 out tokens · 52048 ms · 2026-05-21T04:41:52.015014+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

48 extracted references · 48 canonical work pages · 9 internal anchors

  1. [1]

    Behavioral Cloning from Observation

    Faraz Torabi, Garrett Warnell, and Peter Stone. Behavioral cloning from observation.arXiv preprint arXiv:1805.01954, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  2. [2]

    Imitation learning: A survey of learning methods.ACM Computing Surveys (CSUR), 50(2):1–35, 2017

    Ahmed Hussein, Mohamed Medhat Gaber, Eyad Elyan, and Chrisina Jayne. Imitation learning: A survey of learning methods.ACM Computing Surveys (CSUR), 50(2):1–35, 2017

    work page 2017

  3. [3]

    RT-1: Robotics Transformer for Real-World Control at Scale

    Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, Joseph Dabis, Chelsea Finn, Keerthana Gopalakrishnan, Karol Hausman, Alex Herzog, Jasmine Hsu, et al. Rt-1: Robotics transformer for real-world control at scale.arXiv preprint arXiv:2212.06817, 2022

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  4. [4]

    RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control

    Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, Xi Chen, Krzysztof Choro- manski, Tianli Ding, Danny Driess, Avinava Dubey, Chelsea Finn, et al. Rt-2: Vision- language-action models transfer web knowledge to robotic control, 2023.URL https://arxiv. org/abs/2307.15818, 1:2, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  5. [5]

    π0: A vision-language-action flow model for general robot control, 2026

    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, James Tanner, Quan Vuong, Anna Walling, Haohuan Wang, and Ury Zhilinsky. π0: A visi...

    work page 2026

  6. [6]

    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 LeBlanc, Sergey Levine, Adrian Li-Bell, Mohith Mothukuri, Suraj Nair, Karl Pertsc...

    work page 2025

  7. [7]

    Xirl: Cross-embodiment inverse reinforcement learning, 2021

    Kevin Zakka, Andy Zeng, Pete Florence, Jonathan Tompson, Jeannette Bohg, and Debidatta Dwibedi. Xirl: Cross-embodiment inverse reinforcement learning, 2021

    work page 2021

  8. [8]

    Scaling cross- embodied learning: One policy for manipulation, navigation, locomotion and aviation.arXiv preprint arXiv:2408.11812, 2024

    Ria Doshi, Homer Walke, Oier Mees, Sudeep Dasari, and Sergey Levine. Scaling cross- embodied learning: One policy for manipulation, navigation, locomotion and aviation.arXiv preprint arXiv:2408.11812, 2024. 10

    work page arXiv 2024

  9. [9]

    XR-1: Towards Versatile Vision-Language-Action Models via Learning Unified Vision-Motion Representations

    Shichao Fan, Kun Wu, Zhengping Che, Xinhua Wang, Di Wu, Fei Liao, Ning Liu, Yixue Zhang, Zhen Zhao, Zhiyuan Xu, et al. Xr-1: Towards versatile vision-language-action models via learning unified vision-motion representations.arXiv preprint arXiv:2511.02776, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  10. [10]

    La- tent action diffusion for cross-embodiment manipulation

    Erik Bauer, Elvis Nava, and Robert K Katzschmann. Latent action diffusion for cross- embodiment manipulation.arXiv preprint arXiv:2506.14608, 2025

    work page arXiv 2025

  11. [11]

    arXiv preprint arXiv:2509.22199 (2025)

    Haoyun Li, Ivan Zhang, Runqi Ouyang, Xiaofeng Wang, Zheng Zhu, Zhiqin Yang, Zhentao Zhang, Boyuan Wang, Chaojun Ni, Wenkang Qin, et al. Mimicdreamer: Aligning human and robot demonstrations for scalable vla training.arXiv preprint arXiv:2509.22199, 2025

    work page arXiv 2025

  12. [12]

    Deep visual foresight for planning robot motion, 2017

    Chelsea Finn and Sergey Levine. Deep visual foresight for planning robot motion, 2017

    work page 2017

  13. [13]

    Goal-conditioned reinforcement learning: Problems and solutions, 2022

    Minghuan Liu, Menghui Zhu, and Weinan Zhang. Goal-conditioned reinforcement learning: Problems and solutions, 2022

    work page 2022

  14. [14]

    A path towards autonomous machine intelligence version 0.9

    Yann LeCun et al. A path towards autonomous machine intelligence version 0.9. 2, 2022-06-27. Open Review, 62(1):1–62, 2022

    work page 2022

  15. [15]

    Self-supervised learning from images with a joint- embedding predictive architecture

    Mahmoud Assran, Quentin Duval, Ishan Misra, Piotr Bojanowski, Pascal Vincent, Michael Rabbat, Yann LeCun, and Nicolas Ballas. Self-supervised learning from images with a joint- embedding predictive architecture. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 15619–15629, 2023

    work page 2023

  16. [16]

    Revisiting Feature Prediction for Learning Visual Representations from Video

    Adrien Bardes, Quentin Garrido, Jean Ponce, Xinlei Chen, Michael Rabbat, Yann LeCun, Mahmoud Assran, and Nicolas Ballas. Revisiting feature prediction for learning visual repre- sentations from video.arXiv preprint arXiv:2404.08471, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  17. [17]

    V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning

    Mido Assran, Adrien Bardes, David Fan, Quentin Garrido, Russell Howes, Matthew Muckley, Ammar Rizvi, Claire Roberts, Koustuv Sinha, Artem Zholus, et al. V-jepa 2: Self-supervised video models enable understanding, prediction and planning.arXiv preprint arXiv:2506.09985, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  18. [18]

    V-jepa 2.1: Unlocking dense features in video self-supervised learning, 2026

    Lorenzo Mur-Labadia, Matthew Muckley, Amir Bar, Mido Assran, Koustuv Sinha, Mike Rabbat, Yann LeCun, Nicolas Ballas, and Adrien Bardes. V-jepa 2.1: Unlocking dense features in video self-supervised learning, 2026

    work page 2026

  19. [19]

    Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio

    Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks, 2014

    work page 2014

  20. [20]

    Taming transformers for high-resolution image synthesis, 2021

    Patrick Esser, Robin Rombach, and Björn Ommer. Taming transformers for high-resolution image synthesis, 2021

    work page 2021

  21. [21]

    Denoising diffusion probabilistic models, 2020

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models, 2020

    work page 2020

  22. [22]

    Denoising diffusion implicit models, 2022

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models, 2022

    work page 2022

  23. [23]

    Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le. Flow matching for generative modeling, 2023

    work page 2023

  24. [24]

    Auto-encoding variational bayes, 2022

    Diederik P Kingma and Max Welling. Auto-encoding variational bayes, 2022

    work page 2022

  25. [25]

    Extracting and composing robust features with denoising autoencoders

    Pascal Vincent, Hugo Larochelle, Yoshua Bengio, and Pierre-Antoine Manzagol. Extracting and composing robust features with denoising autoencoders. InProceedings of the 25th International Conference on Machine Learning, ICML ’08, page 1096–1103, New York, NY , USA, 2008. Association for Computing Machinery

    work page 2008

  26. [26]

    Masked autoencoders are scalable vision learners, 2021

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, and Ross Girshick. Masked autoencoders are scalable vision learners, 2021

    work page 2021

  27. [27]

    Optimization of computer simulation models with rare events.European Journal of Operational Research, 99(1):89–112, 1997

    Reuven Y Rubinstein. Optimization of computer simulation models with rare events.European Journal of Operational Research, 99(1):89–112, 1997. 11

    work page 1997

  28. [28]

    Deep reinforcement learning in a handful of trials using probabilistic dynamics models.Advances in neural information processing systems, 31, 2018

    Kurtland Chua, Roberto Calandra, Rowan McAllister, and Sergey Levine. Deep reinforcement learning in a handful of trials using probabilistic dynamics models.Advances in neural information processing systems, 31, 2018

    work page 2018

  29. [29]

    Rlbench: The robot learning benchmark & learning environment.IEEE Robotics and Automation Letters, 5(2):3019–3026, 2020

    Stephen James, Zicong Ma, David Rovick Arrojo, and Andrew J Davison. Rlbench: The robot learning benchmark & learning environment.IEEE Robotics and Automation Letters, 5(2):3019–3026, 2020

    work page 2020

  30. [30]

    A survey of robot learning from demonstration.Robotics and autonomous systems, 57(5):469–483, 2009

    Brenna D Argall, Sonia Chernova, Manuela Veloso, and Brett Browning. A survey of robot learning from demonstration.Robotics and autonomous systems, 57(5):469–483, 2009

    work page 2009

  31. [31]

    Transfer learning for reinforcement learning domains: A survey.Journal of Machine Learning Research, 10(7), 2009

    Matthew E Taylor and Peter Stone. Transfer learning for reinforcement learning domains: A survey.Journal of Machine Learning Research, 10(7), 2009

    work page 2009

  32. [32]

    Learning Invariant Feature Spaces to Transfer Skills with Reinforcement Learning

    Abhishek Gupta, Coline Devin, YuXuan Liu, Pieter Abbeel, and Sergey Levine. Learn- ing invariant feature spaces to transfer skills with reinforcement learning.arXiv preprint arXiv:1703.02949, 2017

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  33. [33]

    Learn- ing modular neural network policies for multi-task and multi-robot transfer

    Coline Devin, Abhishek Gupta, Trevor Darrell, Pieter Abbeel, and Sergey Levine. Learn- ing modular neural network policies for multi-task and multi-robot transfer. In2017 IEEE international conference on robotics and automation (ICRA), pages 2169–2176. IEEE, 2017

    work page 2017

  34. [34]

    Universal actions for enhanced embodied foundation models, 2025

    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, 2025

    work page 2025

  35. [35]

    Unit: Toward a unified physical language for human-to-humanoid policy learning and world modeling, 2026

    Boyu Chen, Yi Chen, Lu Qiu, Jerry Bai, Yuying Ge, and Yixiao Ge. Unit: Toward a unified physical language for human-to-humanoid policy learning and world modeling, 2026

    work page 2026

  36. [36]

    Embodiment Collaboration, Abby O’Neill, Abdul Rehman, Abhinav Gupta, Abhiram Mad- dukuri, Abhishek Gupta, Abhishek Padalkar, Abraham Lee, Acorn Pooley, Agrim Gupta, Ajay Mandlekar, Ajinkya Jain, Albert Tung, Alex Bewley, Alex Herzog, Alex Irpan, Alexander Khaz- atsky, Anant Rai, Anchit Gupta, Andrew Wang, Andrey Kolobov, Anikait Singh, Animesh Garg, Aniru...

    work page 2025

  37. [37]

    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

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  38. [38]

    Xskill: Cross embodiment skill discovery, 2023

    Mengda Xu, Zhenjia Xu, Cheng Chi, Manuela Veloso, and Shuran Song. Xskill: Cross embodiment skill discovery, 2023

    work page 2023

  39. [39]

    Learning from observation: A survey of recent advances, 2025

    Returaj Burnwal, Hriday Mehta, Nirav Pravinbhai Bhatt, and Balaraman Ravindran. Learning from observation: A survey of recent advances, 2025

    work page 2025

  40. [40]

    Mirage: Cross-embodiment zero-shot policy transfer with cross-painting, 2024

    Lawrence Yunliang Chen, Kush Hari, Karthik Dharmarajan, Chenfeng Xu, Quan Vuong, and Ken Goldberg. Mirage: Cross-embodiment zero-shot policy transfer with cross-painting, 2024

    work page 2024

  41. [41]

    Latent diffusion planning for imitation learning, 2025

    Amber Xie, Oleh Rybkin, Dorsa Sadigh, and Chelsea Finn. Latent diffusion planning for imitation learning, 2025

    work page 2025

  42. [42]

    One-shot imitation under mismatched execution, 2025

    Kushal Kedia, Prithwish Dan, Angela Chao, Maximus Adrian Pace, and Sanjiban Choudhury. One-shot imitation under mismatched execution, 2025

    work page 2025

  43. [43]

    Video prediction policy: A generalist robot policy with predictive visual representations, 2025

    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, 2025

    work page 2025

  44. [44]

    Dream to control: Learning behaviors by latent imagination, 2020

    Danijar Hafner, Timothy Lillicrap, Jimmy Ba, and Mohammad Norouzi. Dream to control: Learning behaviors by latent imagination, 2020

    work page 2020

  45. [45]

    World Models

    David Ha and Jürgen Schmidhuber. World models.arXiv preprint arXiv:1803.10122, 2(3):440, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  46. [46]

    World action models are zero-shot policies, 2026

    Seonghyeon Ye, Yunhao Ge, Kaiyuan Zheng, Shenyuan Gao, Sihyun Yu, George Kurian, Suneel Indupuru, You Liang Tan, Chuning Zhu, Jiannan Xiang, Ayaan Malik, Kyungmin Lee, William Liang, Nadun Ranawaka, Jiasheng Gu, Yinzhen Xu, Guanzhi Wang, Fengyuan Hu, Avnish Narayan, Johan Bjorck, Jing Wang, Gwanghyun Kim, Dantong Niu, Ruijie Zheng, Yuqi Xie, Jimmy Wu, Qi ...

    work page 2026

  47. [47]

    Learning to predict activity progress by self-supervised video alignment

    Gerard Donahue and Ehsan Elhamifar. Learning to predict activity progress by self-supervised video alignment. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18667–18677, 2024

    work page 2024

  48. [48]

    Diffusion policy: Visuomotor policy learning via action diffusion

    Cheng Chi, Zhenjia Xu, Siyuan Feng, Eric Cousineau, Yilun Du, Benjamin Burchfiel, Russ Tedrake, and Shuran Song. Diffusion policy: Visuomotor policy learning via action diffusion. The International Journal of Robotics Research, 44(10-11):1684–1704, 2025. 13 A Preliminary We briefly review the key building blocks of our approach: action-conditioned world m...

    work page 2025