pith. machine review for the scientific record. sign in

arxiv: 2512.01773 · v2 · submitted 2025-12-01 · 💻 cs.RO

IGen: Scalable Data Generation for Robot Learning from Open-World Images

Chenghao Gu , Haolan Kang , Junchao Lin , Jinghe Wang , Duo Wu , Shuzhao Xie , Fanding Huang , Junchen Ge
show 5 more authors
Ziyang Gong Letian Li Hongying Zheng Changwei Lv Zhi Wang
This is my paper

Pith reviewed 2026-05-17 02:54 UTC · model grok-4.3

classification 💻 cs.RO
keywords robot learningdata generationopen-world imagesvision-language modelsvisuomotor dataSE(3) posesscalable traininggeneralist policies
0
0 comments X

The pith

Open-world images can be converted into scalable, executable robot training data that produces policies matching real-world performance.

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

The paper sets out to demonstrate that the vast supply of unstructured open-world images can be turned into large quantities of realistic visuomotor training data without requiring expensive on-robot collection. IGen does this by first lifting 2D images into 3D scene representations, then using vision-language models to create both high-level task plans and precise low-level SE(3) end-effector pose sequences. These poses drive a synthesis step that generates plausible scene dynamics and renders temporally coherent image sequences. A sympathetic reader would care because current generalist robot policies are limited by the cost and narrowness of real-world data collection; if the method works, training can draw on essentially unlimited photo collections. Experiments in the paper show that policies trained exclusively on the generated data reach performance levels comparable to those trained on actual robot trajectories.

Core claim

IGen converts unstructured open-world images into structured 3D scene representations, then applies vision-language models to produce high-level plans and low-level SE(3) end-effector pose sequences; these poses are used to synthesize dynamic scene evolution and render temporally coherent visual observations, yielding visuomotor data whose quality is high enough that policies trained solely on it achieve performance comparable to policies trained on real-world robot data.

What carries the argument

The IGen pipeline that lifts 2D pixels into 3D representations and uses vision-language models to generate high-level plans together with low-level SE(3) end-effector pose sequences for synthesizing realistic dynamic trajectories.

If this is right

  • Robot training data can be generated at large scale from any collection of open-world images without physical robot runs.
  • Policies gain exposure to far greater scene diversity than is feasible with conventional on-robot collection.
  • Generated actions are specified as executable SE(3) pose sequences that can be transferred directly to real robots.
  • Temporally coherent rendered observations support training of policies that must act over sequences of frames.
  • The overall need for labor-intensive real-world data collection for generalist policies is reduced.

Where Pith is reading between the lines

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

  • If IGen data matches real performance, mixing small amounts of real data with large IGen sets could improve robustness at lower cost.
  • The same image-to-3D-to-action pipeline could be tested on navigation or mobile manipulation by adapting the pose generation step.
  • Internet-scale photo collections could become a primary resource for robot datasets if the synthesis quality generalizes across domains.
  • Future experiments could measure how much additional real data is still needed to close any remaining performance gap.

Load-bearing premise

The vision-language model outputs and subsequent rendering steps produce actions and scene changes that remain close enough to real robot execution to avoid large distribution shifts.

What would settle it

Train identical policy architectures on IGen-generated data versus real robot trajectories for the same manipulation tasks in matched environments and measure whether the IGen-trained policy reaches at least 80 percent of the real-data policy's success rate.

Figures

Figures reproduced from arXiv: 2512.01773 by Changwei Lv, Chenghao Gu, Duo Wu, Fanding Huang, Haolan Kang, Hongying Zheng, Jinghe Wang, Junchao Lin, Junchen Ge, Letian Li, Shuzhao Xie, Zhi Wang, Ziyang Gong.

Figure 1
Figure 1. Figure 1: We propose IGen, a data generation framework that converts open-world images into grounded visuomotor data, enabling scalable data synthesis for robot learning. From a single image, IGen generates large-scale realistic observations and reliable actions. The policies trained solely on IGen-generated data can effectively generalize to real-world scenes and successfully perform manipulation tasks. Abstract Th… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of IGen. Given an open-world image and a task description, IGen first reconstructs the environment and objects as point clouds via Foundation Vision Models. After spatial keypoint extraction, VLM maps the task description to high-level plans and low-level control commands. During the robot’s execution in simulation, a virtual depth camera captures the motion point cloud sequences. The resulting en… view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative comparison of robotic behavior generation using IGen. Given a single captured image and a natural-language manipulation instruction, TesserAct [76], Cosmos [2], and our IGen generate behavior observations. IGen produces more instruction￾consistent and physically coherent object motions, closely matching the intended tasks. The green box represents action observations that adhere to physical law… view at source ↗
Figure 5
Figure 5. Figure 5: Evaluation of IGen’s computational effi￾ciency. We compare the video generation time and GPU memory consumption of IGen and baselines under iden￾tical input images and task instructions. The average computation time refers to the time required to generate one robot behavior video. object, the world pose the object evolves by rigidly following the end-effector. The manipulated object, represented as a point… view at source ↗
Figure 6
Figure 6. Figure 6: Experimental setup. Starting from a captured real-world scene image, IGen automatically generates 1,000 task demonstrations with spatial randomization. The resulting data are used to train a visuomotor policy, which is later deployed and evaluated in the real world [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Real-world robot evaluation results. We assess policy performance on real-world tasks, comparing task success rates un￾der four different conditions: zero-shot, fine-tuned with 10 human teleoperation data, fine-tuned with 100 human teleoperation data, and fine-tuned with 1,000 IGen-synthesized data. assess the perceptual discrepancy between generated scenes and the original images, we observe that IGen ach… view at source ↗
Figure 10
Figure 10. Figure 10: Robot and Camera Placement in Simulation. In simulation platforms such as IsaacSim, the virtual camera is placed at the position (0, 0, 0), while the robotic arm base is positioned at the corresponding point in the point cloud, denoted as (xr, yr, zr). RGB and depth data are collected during the robotic arm’s motion. C. Manipulation Synthesis Details We divide the point cloud sequence into three component… view at source ↗
Figure 9
Figure 9. Figure 9: Single-Image Scene Reconstruction Pipeline. B. Simulation Environment Details This section describes the details of building the robotic manipulation platform in simulation. We adopt Isaac Sim as the simulation environment and deploy both the Franka Emika Panda and Franka Research 3 robotic arms within it. For motion planning, we utilize Curobo as the solver, which computes feasible trajectories given the … view at source ↗
Figure 11
Figure 11. Figure 11: Point Cloud Synthesis during Manipulation. At time tg, the object is grasped. The gripper width is calculated based on the point cloud, and the transformation of the object point cloud at time t is computed according to the end-effector’s pose. 1 [PITH_FULL_IMAGE:figures/full_fig_p013_11.png] view at source ↗
Figure 13
Figure 13. Figure 13: Spatial randomization of real-world data and IGen [PITH_FULL_IMAGE:figures/full_fig_p014_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Spatial randomization of real-world data and IGen [PITH_FULL_IMAGE:figures/full_fig_p014_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Spatial randomization of real-world data and IGen [PITH_FULL_IMAGE:figures/full_fig_p014_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Real-World Deployment of Policy trained with IGen-Generated Data. The task instructions are as follows: Grab the watering can and water the flowers. Pick up the bottle and place it into the basket. Use the hammer to hit the cardboard box. 3 [PITH_FULL_IMAGE:figures/full_fig_p015_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Real-World Deployment of Policy trained with IGen-Generated Data. The task instructions are as follows: Grasp the toy and put it into the bin. Use the watering can on the cabinet to water the flowers. Pour water from the plastic bottle into the container. 4 [PITH_FULL_IMAGE:figures/full_fig_p016_17.png] view at source ↗
read the original abstract

The rise of generalist robotic policies has created an exponential demand for large-scale training data. However, on-robot data collection is labor-intensive and often limited to specific environments. In contrast, open-world images capture a vast diversity of real-world scenes that naturally align with robotic manipulation tasks, offering a promising avenue for low-cost, large-scale robot data acquisition. Despite this potential, the lack of associated robot actions hinders the practical use of open-world images for robot learning, leaving this rich visual resource largely unexploited. To bridge this gap, we propose IGen, a framework that scalably generates realistic visual observations and executable actions from open-world images. IGen first converts unstructured 2D pixels into structured 3D scene representations suitable for scene understanding and manipulation. It then leverages the reasoning capabilities of vision-language models to transform scene-specific task instructions into high-level plans and generate low-level actions as SE(3) end-effector pose sequences. From these poses, it synthesizes dynamic scene evolution and renders temporally coherent visual observations. Experiments validate the high quality of visuomotor data generated by IGen, and show that policies trained solely on IGen-synthesized data achieve performance comparable to those trained on real-world data. This highlights the potential of IGen to support scalable data generation from open-world images for generalist robotic policy training.

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 IGen, a pipeline that converts open-world 2D images into structured 3D scene representations, employs vision-language models to derive high-level task plans and low-level SE(3) end-effector pose sequences from scene-specific instructions, synthesizes dynamic scene evolution from those poses, and renders temporally coherent visual observations. The central claim is that the resulting visuomotor data is high-quality and that policies trained exclusively on IGen-generated data achieve performance comparable to policies trained on real-world robot data.

Significance. If the comparability claim is substantiated with quantitative evidence, IGen would provide a scalable route to leverage abundant open-world images for robot learning, substantially lowering the cost and environmental constraints of on-robot data collection for generalist policies.

major comments (2)
  1. [Abstract] Abstract: the statement that 'experiments validate the high quality of visuomotor data generated by IGen, and show that policies trained solely on IGen-synthesized data achieve performance comparable to those trained on real-world data' is unsupported by any reported metrics, baselines, success rates, task suite definitions, trial counts, or controls for distribution shift. This directly undermines the central claim and requires explicit quantitative comparison (e.g., success-rate deltas and ablation removing dynamics synthesis).
  2. [Method / Experiments] Method and Experiments: the pipeline assumes that VLM-generated SE(3) pose sequences, when used to drive 3D scene evolution and rendering, produce observation-action pairs whose distribution matches real robot execution closely enough for zero-shot policy transfer. No ablation or control (e.g., rendered vs. real images under identical policies, or contact/occlusion statistics) is described to test this assumption, which is load-bearing for the sim-to-real transfer result.
minor comments (2)
  1. [Method] Clarify the precise algorithm or parameters used for synthesizing dynamic scene evolution from the SE(3) pose sequences (e.g., interpolation method, physics model).
  2. [Experiments] Add a table or figure summarizing the exact policy architectures, training hyperparameters, and evaluation environments used in the comparability experiments.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. We address each of the major comments below and outline the revisions we will make to the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the statement that 'experiments validate the high quality of visuomotor data generated by IGen, and show that policies trained solely on IGen-synthesized data achieve performance comparable to those trained on real-world data' is unsupported by any reported metrics, baselines, success rates, task suite definitions, trial counts, or controls for distribution shift. This directly undermines the central claim and requires explicit quantitative comparison (e.g., success-rate deltas and ablation removing dynamics synthesis).

    Authors: We agree with the referee that the abstract's claim would be strengthened by explicit quantitative evidence. We will revise the abstract and add to the experiments section detailed metrics including success rates, baselines, task suite definitions, trial counts, and an ablation removing the dynamics synthesis component to support the comparability to real-world data. revision: yes

  2. Referee: [Method / Experiments] Method and Experiments: the pipeline assumes that VLM-generated SE(3) pose sequences, when used to drive 3D scene evolution and rendering, produce observation-action pairs whose distribution matches real robot execution closely enough for zero-shot policy transfer. No ablation or control (e.g., rendered vs. real images under identical policies, or contact/occlusion statistics) is described to test this assumption, which is load-bearing for the sim-to-real transfer result.

    Authors: We agree that additional controls are necessary to substantiate the key assumption in our pipeline. We will revise the experiments section to include ablations such as training identical policies on rendered versus real images and reporting contact and occlusion statistics. This will help demonstrate the closeness of the generated data distribution to real robot executions. revision: yes

Circularity Check

0 steps flagged

No circularity detected; forward pipeline from images to data with empirical validation

full rationale

The paper describes IGen as a sequential, forward pipeline: 2D-to-3D conversion, VLM-based high-level planning and SE(3) pose sequence generation, dynamic scene synthesis, and rendering of observations. The claim that policies trained on IGen data achieve comparable performance is presented as an empirical experimental outcome rather than a mathematical derivation. No equations, fitted parameters, or steps reduce by construction to the inputs or to self-citations; the method does not invoke uniqueness theorems, ansatzes smuggled via prior work, or renamings of known results in a load-bearing way. The derivation chain remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The framework rests on the assumption that current vision-language models can reliably translate scene-specific instructions into correct high-level plans and low-level SE(3) actions, and that 3D reconstruction from single images yields sufficiently accurate geometry for manipulation planning.

axioms (2)
  • domain assumption Vision-language models can produce executable robot plans from scene descriptions
    Invoked when the method uses VLMs to transform task instructions into high-level plans and low-level actions
  • domain assumption 3D scene representations extracted from 2D images are accurate enough for manipulation
    Central to the first step of converting pixels into structured 3D representations suitable for scene understanding and manipulation

pith-pipeline@v0.9.0 · 5582 in / 1315 out tokens · 29994 ms · 2026-05-17T02:54:43.638635+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 2 Pith papers

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

  1. MesonGS++: Post-training Compression of 3D Gaussian Splatting with Hyperparameter Searching

    cs.CV 2026-04 unverdicted novelty 6.0

    MesonGS++ achieves over 34x compression of 3D Gaussian Splatting models with preserved or improved PSNR by using size-aware joint optimization of pruning and quantization hyperparameters via discrete sampling and 0-1 ...

  2. MesonGS++: Post-training Compression of 3D Gaussian Splatting with Hyperparameter Searching

    cs.CV 2026-04 unverdicted novelty 5.0

    MesonGS++ achieves over 34x compression of 3D Gaussian Splatting models post-training while preserving or exceeding original rendering quality through size-aware hyperparameter optimization.

Reference graph

Works this paper leans on

80 extracted references · 80 canonical work pages · cited by 1 Pith paper · 27 internal anchors

  1. [1]

    GPT-4 Technical Report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report.arXiv preprint arXiv:2303.08774, 2023. 2, 3, 4, 6, 8

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  2. [2]

    Cosmos World Foundation Model Platform for Physical AI

    Niket Agarwal, Arslan Ali, Maciej Bala, Yogesh Balaji, Erik Barker, Tiffany Cai, Prithvijit Chattopadhyay, Yongxin Chen, Yin Cui, Yifan Ding, et al. Cosmos world foundation model platform for physical ai.arXiv preprint arXiv:2501.03575,

    work page internal anchor Pith review Pith/arXiv arXiv

  3. [3]

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, et al. Qwen2. 5-vl technical report.arXiv preprint arXiv:2502.13923, 2025. 2, 3, 4, 6, 8

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  4. [4]

    GR00T N1: An Open Foundation Model for Generalist Humanoid Robots

    Johan Bjorck, Fernando Casta ˜neda, Nikita Cherniadev, Xingye Da, Runyu Ding, Linxi Fan, Yu Fang, Dieter Fox, Fengyuan Hu, Spencer Huang, et al. Gr00t n1: An open foun- dation model for generalist humanoid robots.arXiv preprint arXiv:2503.14734, 2025. 2, 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  5. [5]

    $\pi_0$: 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, et al. pi 0: A vision-language- action flow model for general robot control.arXiv preprint arXiv:2410.24164, 2024. 2, 3, 8

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  6. [6]

    Genie: Generative interactive environments

    Jake Bruce, Michael D Dennis, Ashley Edwards, Jack Parker- Holder, Yuge Shi, Edward Hughes, Matthew Lai, Aditi Mavalankar, Richie Steigerwald, Chris Apps, et al. Genie: Generative interactive environments. InForty-first Interna- tional Conference on Machine Learning, 2024. 2

    work page 2024

  7. [7]

    Fast-in-slow: a dual-system founda- tion model unifying fast manipulation within slow reasoning,

    Hao Chen, Jiaming Liu, Chenyang Gu, Zhuoyang Liu, Ren- rui Zhang, Xiaoqi Li, Xiao He, Yandong Guo, Chi-Wing Fu, Shanghang Zhang, et al. Fast-in-slow: A dual-system founda- tion model unifying fast manipulation within slow reasoning. arXiv preprint arXiv:2506.01953, 2025. 3

    work page arXiv 2025

  8. [8]

    Vidbot: Learning generalizable 3d actions from in-the-wild 2d human videos for zero-shot robotic manipulation

    Hanzhi Chen, Boyang Sun, Anran Zhang, Marc Pollefeys, and Stefan Leutenegger. Vidbot: Learning generalizable 3d actions from in-the-wild 2d human videos for zero-shot robotic manipulation. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 27661–27672,

    work page

  9. [9]

    Object-centric dexterous manipulation from human motion data.arXiv preprint arXiv:2411.04005, 2024

    Yuanpei Chen, Chen Wang, Yaodong Yang, and C Karen Liu. Object-centric dexterous manipulation from human motion data.arXiv preprint arXiv:2411.04005, 2024. 3

    work page arXiv 2024

  10. [10]

    Open-television: Teleoperation with immersive active visual feedback.arXiv preprint arXiv:2407.01512, 2024

    Xuxin Cheng, Jialong Li, Shiqi Yang, Ge Yang, and Xiaolong Wang. Open-television: Teleoperation with immersive active visual feedback.arXiv preprint arXiv:2407.01512, 2024. 3

    work page arXiv 2024

  11. [11]

    Diffusion policy: Visuomotor policy learning via action diffu- sion.The International Journal of Robotics Research, page 02783649241273668, 2023

    Cheng Chi, Zhenjia Xu, Siyuan Feng, Eric Cousineau, Yilun Du, Benjamin Burchfiel, Russ Tedrake, and Shuran Song. Diffusion policy: Visuomotor policy learning via action diffu- sion.The International Journal of Robotics Research, page 02783649241273668, 2023. 2

    work page 2023

  12. [12]

    Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots

    Cheng Chi, Zhenjia Xu, Chuer Pan, Eric Cousineau, Ben- jamin Burchfiel, Siyuan Feng, Russ Tedrake, and Shu- ran Song. Universal manipulation interface: In-the-wild robot teaching without in-the-wild robots.arXiv preprint arXiv:2402.10329, 2024. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  13. [13]

    Auto- mated creation of digital cousins for robust policy learning

    Tianyuan Dai, Josiah Wong, Yunfan Jiang, Chen Wang, Cem Gokmen, Ruohan Zhang, Jiajun Wu, and Li Fei-Fei. Auto- mated creation of digital cousins for robust policy learning. arXiv preprint arXiv:2410.07408, 2024. 3

    work page arXiv 2024

  14. [14]

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

    Runyu Ding, Yuzhe Qin, Jiyue Zhu, Chengzhe Jia, Shiqi Yang, Ruihan Yang, Xiaojuan Qi, and Xiaolong Wang. Bunny- visionpro: Real-time bimanual dexterous teleoperation for imitation learning.arXiv preprint arXiv:2407.03162, 2024. 3

    work page arXiv 2024

  15. [15]

    Ar2-d2: Training a robot without a robot.arXiv preprint arXiv:2306.13818,

    Jiafei Duan, Yi Ru Wang, Mohit Shridhar, Dieter Fox, and Ranjay Krishna. Ar2-d2: Training a robot without a robot. arXiv preprint arXiv:2306.13818, 2023. 3

    work page arXiv 2023

  16. [16]

    Bridge Data: Boosting Generalization of Robotic Skills with Cross-Domain Datasets

    Frederik Ebert, Yanlai Yang, Karl Schmeckpeper, Bernadette Bucher, Georgios Georgakis, Kostas Daniilidis, Chelsea Finn, and Sergey Levine. Bridge data: Boosting generalization of robotic skills with cross-domain datasets.arXiv preprint arXiv:2109.13396, 2021. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  17. [17]

    Graspnet-1billion: A large-scale benchmark for general object grasping

    Hao-Shu Fang, Chenxi Wang, Minghao Gou, and Cewu Lu. Graspnet-1billion: A large-scale benchmark for general object grasping. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11444–11453,

    work page

  18. [18]

    Rh20t: A comprehensive robotic dataset for learning diverse skills in one-shot.arXiv preprint arXiv:2307.00595, 2023

    Hao-Shu Fang, Hongjie Fang, Zhenyu Tang, Jirong Liu, Chenxi Wang, Junbo Wang, Haoyi Zhu, and Cewu Lu. Rh20t: A comprehensive robotic dataset for learning diverse skills in one-shot.arXiv preprint arXiv:2307.00595, 2023. 3

    work page arXiv 2023

  19. [19]

    Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation

    Zipeng Fu, Tony Z Zhao, and Chelsea Finn. Mobile aloha: Learning bimanual mobile manipulation with low-cost whole- body teleoperation.arXiv preprint arXiv:2401.02117, 2024. 2

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  20. [20]

    Efficient data collection for robotic manip- ulation via compositional generalization.arXiv preprint arXiv:2403.05110, 2024

    Jensen Gao, Annie Xie, Ted Xiao, Chelsea Finn, and Dorsa Sadigh. Efficient data collection for robotic manip- ulation via compositional generalization.arXiv preprint arXiv:2403.05110, 2024. 3

    work page arXiv 2024

  21. [21]

    On pre-training for visuo-motor control: Re- visiting a learning-from-scratch baseline.arXiv preprint arXiv:2212.05749, 2022

    Nicklas Hansen, Zhecheng Yuan, Yanjie Ze, Tongzhou Mu, Aravind Rajeswaran, Hao Su, Huazhe Xu, and Xiao- long Wang. On pre-training for visuo-motor control: Re- visiting a learning-from-scratch baseline.arXiv preprint arXiv:2212.05749, 2022. 2

    work page arXiv 2022

  22. [22]

    LoRA: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen- Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. LoRA: Low-rank adaptation of large language models. In International Conference on Learning Representations, 2022. 8, 2

    work page 2022

  23. [23]

    Mu Hu, Wei Yin, Chi Zhang, Zhipeng Cai, Xiaoxiao Long, Hao Chen, Kaixuan Wang, Gang Yu, Chunhua Shen, and Shaojie Shen. Metric3d v2: A versatile monocular geometric foundation model for zero-shot metric depth and surface nor- mal estimation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024. 3, 1

    work page 2024

  24. [24]

    Otter: A vision-language-action model with text-aware visual feature extraction.arXiv preprint arXiv:2503.03734, 2025

    Huang Huang, Fangchen Liu, Letian Fu, Tingfan Wu, Mustafa Mukadam, Jitendra Malik, Ken Goldberg, and Pieter Abbeel. Otter: A vision-language-action model with text-aware visual feature extraction.arXiv preprint arXiv:2503.03734, 2025. 3

    work page arXiv 2025

  25. [25]

    VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models

    Wenlong Huang, Chen Wang, Ruohan Zhang, Yunzhu Li, Jiajun Wu, and Li Fei-Fei. V oxposer: Composable 3d value maps for robotic manipulation with language models.arXiv preprint arXiv:2307.05973, 2023. 3 9

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  26. [26]

    ReKep: Spatio-Temporal Reasoning of Relational Keypoint Constraints for Robotic Manipulation

    Wenlong Huang, Chen Wang, Yunzhu Li, Ruohan Zhang, and Li Fei-Fei. Rekep: Spatio-temporal reasoning of relational keypoint constraints for robotic manipulation.arXiv preprint arXiv:2409.01652, 2024. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  27. [27]

    SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size

    Forrest N Iandola, Song Han, Matthew W Moskewicz, Khalid Ashraf, William J Dally, and Kurt Keutzer. Squeezenet: Alexnet-level accuracy with 50x fewer parameters and¡ 0.5 mb model size.arXiv preprint arXiv:1602.07360, 2016. 6

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  28. [28]

    $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization

    Physical Intelligence, Kevin Black, Noah Brown, James Darpinian, Karan Dhabalia, Danny Driess, Adnan Esmail, Michael Equi, Chelsea Finn, Niccolo Fusai, et al. pi 0.5: a vision-language-action model with open-world generalization. arXiv preprint arXiv:2504.16054, 2025. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  29. [29]

    Dreamgen: Unlocking gener- alization in robot learning through neural trajectories.arXiv e-prints, pages arXiv–2505, 2025

    Joel Jang, Seonghyeon Ye, Zongyu Lin, Jiannan Xiang, Jo- han Bjorck, Yu Fang, Fengyuan Hu, Spencer Huang, Kaushil Kundalia, Yen-Chen Lin, et al. Dreamgen: Unlocking gener- alization in robot learning through neural trajectories.arXiv e-prints, pages arXiv–2505, 2025. 2, 3, 6, 7

    work page 2025

  30. [30]

    Discoverse: Efficient robot simula- tion in complex high-fidelity environments.arXiv preprint arXiv:2507.21981, 2025

    Yufei Jia, Guangyu Wang, Yuhang Dong, Junzhe Wu, Yu- pei Zeng, Haonan Lin, Zifan Wang, Haizhou Ge, Weibin Gu, Kairui Ding, et al. Discoverse: Efficient robot simula- tion in complex high-fidelity environments.arXiv preprint arXiv:2507.21981, 2025. 2, 3

    work page arXiv 2025

  31. [31]

    Ditto: Build- ing digital twins of articulated objects from interaction

    Zhenyu Jiang, Cheng-Chun Hsu, and Yuke Zhu. Ditto: Build- ing digital twins of articulated objects from interaction. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 5616–5626, 2022. 3

    work page 2022

  32. [32]

    Dexmimicgen: Automated data generation for bimanual dex- terous manipulation via imitation learning

    Zhenyu Jiang, Yuqi Xie, Kevin Lin, Zhenjia Xu, Weikang Wan, Ajay Mandlekar, Linxi Jim Fan, and Yuke Zhu. Dexmimicgen: Automated data generation for bimanual dex- terous manipulation via imitation learning. In2025 IEEE In- ternational Conference on Robotics and Automation (ICRA), pages 16923–16930. IEEE, 2025. 3

    work page 2025

  33. [33]

    OpenVLA: An Open-Source Vision-Language-Action Model

    Moo Jin Kim, Karl Pertsch, Siddharth Karamcheti, Ted Xiao, Ashwin Balakrishna, Suraj Nair, Rafael Rafailov, Ethan Foster, Grace Lam, Pannag Sanketi, et al. Openvla: An open-source vision-language-action model.arXiv preprint arXiv:2406.09246, 2024. 2, 3

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  34. [34]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. InProceedings of the IEEE/CVF international confer- ence on computer vision, pages 4015–4026, 2023. 3, 1

    work page 2023

  35. [35]

    Learning to act from actionless videos through dense correspondences.arXiv preprint arXiv:2310.08576, 2023

    Po-Chen Ko, Jiayuan Mao, Yilun Du, Shao-Hua Sun, and Joshua B Tenenbaum. Learning to act from actionless videos through dense correspondences.arXiv preprint arXiv:2310.08576, 2023. 3

    work page arXiv 2023

  36. [36]

    Im- agenet classification with deep convolutional neural networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Im- agenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 2012. 6

    work page 2012

  37. [37]

    What matters when building vision-language mod- els?Advances in Neural Information Processing Systems, 37: 87874–87907, 2024

    Hugo Laurenc ¸on, L´eo Tronchon, Matthieu Cord, and Victor Sanh. What matters when building vision-language mod- els?Advances in Neural Information Processing Systems, 37: 87874–87907, 2024. 2

    work page 2024

  38. [38]

    Any6d: Model-free 6d pose estimation of novel objects

    Taeyeop Lee, Bowen Wen, Minjun Kang, Gyuree Kang, In So Kweon, and Kuk-Jin Yoon. Any6d: Model-free 6d pose estimation of novel objects. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 11633– 11643, 2025. 4

    work page 2025

  39. [39]

    Phantom: Training robots without robots using only human videos.arXiv preprint arXiv:2503.00779,

    Marion Lepert, Jiaying Fang, and Jeannette Bohg. Phan- tom: Training robots without robots using only human videos. arXiv preprint arXiv:2503.00779, 2025. 3

    work page arXiv 2025

  40. [40]

    Evaluating Real-World Robot Manipulation Policies in Simulation

    Xuanlin Li, Kyle Hsu, Jiayuan Gu, Karl Pertsch, Oier Mees, Homer Rich Walke, Chuyuan Fu, Ishikaa Lunawat, Isabel Sieh, Sean Kirmani, et al. Evaluating real-world robot manipulation policies in simulation.arXiv preprint arXiv:2405.05941, 2024. 3, 6

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  41. [41]

    Robogsim: A real2sim2real robotic gaussian splatting simu- lator.arXiv preprint arXiv:2411.11839, 2024

    Xinhai Li, Jialin Li, Ziheng Zhang, Rui Zhang, Fan Jia, Tian- cai Wang, Haoqiang Fan, Kuo-Kun Tseng, and Ruiping Wang. Robogsim: A real2sim2real robotic gaussian splatting simu- lator.arXiv preprint arXiv:2411.11839, 2024. 2, 3

    work page arXiv 2024

  42. [42]

    HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model

    Jiaming Liu, Hao Chen, Pengju An, Zhuoyang Liu, Renrui Zhang, Chenyang Gu, Xiaoqi Li, Ziyu Guo, Sixiang Chen, Mengzhen Liu, et al. Hybridvla: Collaborative diffusion and autoregression in a unified vision-language-action model. arXiv preprint arXiv:2503.10631, 2025. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  43. [43]

    Robo-gs: A physics consistent spatial- temporal model for robotic arm with hybrid representation

    Haozhe Lou, Yurong Liu, Yike Pan, Yiran Geng, Jianteng Chen, Wenlong Ma, Chenglong Li, Lin Wang, Hengzhen Feng, Lu Shi, et al. Robo-gs: A physics consistent spatial- temporal model for robotic arm with hybrid representation. In2025 IEEE International Conference on Robotics and Au- tomation (ICRA), pages 15379–15386. IEEE, 2025. 2, 3

    work page 2025

  44. [44]

    DeepSeek-VL: Towards Real-World Vision-Language Understanding

    Haoyu Lu, Wen Liu, Bo Zhang, Bingxuan Wang, Kai Dong, Bo Liu, Jingxiang Sun, Tongzheng Ren, Zhuoshu Li, Hao Yang, et al. Deepseek-vl: towards real-world vision-language understanding.arXiv preprint arXiv:2403.05525, 2024. 2, 3

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  45. [45]

    MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations

    Ajay Mandlekar, Soroush Nasiriany, Bowen Wen, Iretiayo Akinola, Yashraj Narang, Linxi Fan, Yuke Zhu, and Dieter Fox. Mimicgen: A data generation system for scalable robot learning using human demonstrations.arXiv preprint arXiv:2310.17596, 2023. 3

    work page internal anchor Pith review arXiv 2023

  46. [46]

    So you think you can scale up autonomous robot data collection?arXiv preprint arXiv:2411.01813, 2024

    Suvir Mirchandani, Suneel Belkhale, Joey Hejna, Evelyn Choi, Md Sazzad Islam, and Dorsa Sadigh. So you think you can scale up autonomous robot data collection?arXiv preprint arXiv:2411.01813, 2024. 2, 3

    work page arXiv 2024

  47. [47]

    Graspgen: A diffusion-based framework for 6-dof grasping with on-generator training.arXiv preprint arXiv:2507.13097,

    Adithyavairavan Murali, Balakumar Sundaralingam, Yu-Wei Chao, Wentao Yuan, Jun Yamada, Mark Carlson, Fabio Ramos, Stan Birchfield, Dieter Fox, and Clemens Eppner. Graspgen: A diffusion-based framework for 6-dof grasping with on-generator training.arXiv preprint arXiv:2507.13097,

    work page arXiv

  48. [48]

    DINOv2: Learning Robust Visual Features without Supervision

    Maxime Oquab, Timoth´ee Darcet, Th´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision.arXiv preprint arXiv:2304.07193, 2023. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  49. [49]

    Open x-embodiment: Robotic learning datasets and rt-x models: Open x-embodiment collaboration 0

    Abby O’Neill, Abdul Rehman, Abhiram Maddukuri, Ab- hishek Gupta, Abhishek Padalkar, Abraham Lee, Acorn Poo- ley, Agrim Gupta, Ajay Mandlekar, Ajinkya Jain, et al. Open x-embodiment: Robotic learning datasets and rt-x models: Open x-embodiment collaboration 0. In2024 IEEE Interna- tional Conference on Robotics and Automation (ICRA), pages 6892–6903. IEEE, ...

    work page 2024

  50. [50]

    Omnimanip: Towards general robotic manipulation via object-centric interaction primitives as spatial constraints

    Mingjie Pan, Jiyao Zhang, Tianshu Wu, Yinghao Zhao, Wen- long Gao, and Hao Dong. Omnimanip: Towards general robotic manipulation via object-centric interaction primitives as spatial constraints. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 17359–17369,

    work page

  51. [51]

    A real-to-sim-to-real approach to robotic manipu- lation with vlm-generated iterative keypoint rewards.arXiv preprint arXiv:2502.08643, 2025

    Shivansh Patel, Xinchen Yin, Wenlong Huang, Shubham Garg, Hooshang Nayyeri, Li Fei-Fei, Svetlana Lazebnik, and Yunzhu Li. A real-to-sim-to-real approach to robotic manipu- lation with vlm-generated iterative keypoint rewards.arXiv preprint arXiv:2502.08643, 2025. 3

    work page arXiv 2025

  52. [52]

    Splatsim: Zero- shot sim2real transfer of rgb manipulation policies using gaus- sian splatting

    M Nomaan Qureshi, Sparsh Garg, Francisco Yandun, David Held, George Kantor, and Abhisesh Silwal. Splatsim: Zero- shot sim2real transfer of rgb manipulation policies using gaus- sian splatting. In2025 IEEE International Conference on Robotics and Automation (ICRA), pages 6502–6509. IEEE,

    work page

  53. [53]

    Very Deep Convolutional Networks for Large-Scale Image Recognition

    Karen Simonyan and Andrew Zisserman. Very deep convo- lutional networks for large-scale image recognition.arXiv preprint arXiv:1409.1556, 2014. 6

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  54. [54]

    Hand-object interaction pretraining from videos

    Himanshu Gaurav Singh, Antonio Loquercio, Carmelo Sfer- razza, Jane Wu, Haozhi Qi, Pieter Abbeel, and Jitendra Malik. Hand-object interaction pretraining from videos. In2025 IEEE International Conference on Robotics and Automation (ICRA), pages 3352–3360. IEEE, 2025. 3

    work page 2025

  55. [55]

    Gemini Robotics: Bringing AI into the Physical World

    Gemini Robotics Team, Saminda Abeyruwan, Joshua Ainslie, Jean-Baptiste Alayrac, Montserrat Gonzalez Arenas, Travis Armstrong, Ashwin Balakrishna, Robert Baruch, Maria Bauza, Michiel Blokzijl, et al. Gemini robotics: Bringing ai into the physical world.arXiv preprint arXiv:2503.20020,

    work page internal anchor Pith review Pith/arXiv arXiv

  56. [56]

    Octo: An Open-Source Generalist Robot Policy

    Octo Model Team, Dibya Ghosh, Homer Walke, Karl Pertsch, Kevin Black, Oier Mees, Sudeep Dasari, Joey Hejna, Tobias Kreiman, Charles Xu, et al. Octo: An open-source generalist robot policy.arXiv preprint arXiv:2405.12213, 2024. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  57. [57]

    Droid-slam: Deep visual slam for monocular, stereo, and rgb-d cameras.Advances in neural information processing systems, 34:16558–16569, 2021

    Zachary Teed and Jia Deng. Droid-slam: Deep visual slam for monocular, stereo, and rgb-d cameras.Advances in neural information processing systems, 34:16558–16569, 2021. 3

    work page 2021

  58. [58]

    Robot learning with super-linear scaling.arXiv preprint arXiv:2412.01770, 2024

    Marcel Torne, Arhan Jain, Jiayi Yuan, Vidaaranya Macha, Lars Ankile, Anthony Simeonov, Pulkit Agrawal, and Ab- hishek Gupta. Robot learning with super-linear scaling.arXiv preprint arXiv:2412.01770, 2024. 2, 3

    work page arXiv 2024

  59. [59]

    Gensim: Generating robotic simulation tasks via large language models.arXiv preprint arXiv:2310.01361,

    Lirui Wang, Yiyang Ling, Zhecheng Yuan, Mohit Shridhar, Chen Bao, Yuzhe Qin, Bailin Wang, Huazhe Xu, and Xiao- long Wang. Gensim: Generating robotic simulation tasks via large language models.arXiv preprint arXiv:2310.01361,

    work page arXiv

  60. [60]

    The unreasonable effectiveness of mathematics in the natural sciences.Mathematics and science, 13:1–14, 1990

    Eugene P Wigner et al. The unreasonable effectiveness of mathematics in the natural sciences.Mathematics and science, 13:1–14, 1990. 6

    work page 1990

  61. [61]

    Structured 3d latents for scalable and versatile 3d gen- eration

    Jianfeng Xiang, Zelong Lv, Sicheng Xu, Yu Deng, Ruicheng Wang, Bowen Zhang, Dong Chen, Xin Tong, and Jiaolong Yang. Structured 3d latents for scalable and versatile 3d gen- eration. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 21469–21480, 2025. 3, 1

    work page 2025

  62. [62]

    Demogen: Synthetic demonstration generation for data-efficient visuomotor policy learning.arXiv preprint arXiv:2502.16932,

    Zhengrong Xue, Shuying Deng, Zhenyang Chen, Yixuan Wang, Zhecheng Yuan, and Huazhe Xu. Demogen: Synthetic demonstration generation for data-efficient visuomotor policy learning.arXiv preprint arXiv:2502.16932, 2025. 3

    work page arXiv 2025

  63. [63]

    Learning Interactive Real-World Simulators

    Mengjiao Yang, Yilun Du, Kamyar Ghasemipour, Jonathan Tompson, Dale Schuurmans, and Pieter Abbeel. Learn- ing interactive real-world simulators.arXiv preprint arXiv:2310.06114, 1(2):6, 2023. 2, 3

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  64. [64]

    CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer

    Zhuoyi Yang, Jiayan Teng, Wendi Zheng, Ming Ding, Shiyu Huang, Jiazheng Xu, Yuanming Yang, Wenyi Hong, Xiao- han Zhang, Guanyu Feng, et al. Cogvideox: Text-to-video diffusion models with an expert transformer.arXiv preprint arXiv:2408.06072, 2024. 7

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  65. [65]

    Video2policy: Scaling up manip- ulation tasks in simulation through internet videos.arXiv preprint arXiv:2502.09886, 2025

    Weirui Ye, Fangchen Liu, Zheng Ding, Yang Gao, Oleh Ry- bkin, and Pieter Abbeel. Video2policy: Scaling up manip- ulation tasks in simulation through internet videos.arXiv preprint arXiv:2502.09886, 2025. 3

    work page arXiv 2025

  66. [66]

    Inpaint anything: Segment anything meets image inpainting

    Tao Yu, Runseng Feng, Ruoyu Feng, Jinming Liu, Xin Jin, Wenjun Zeng, and Zhibo Chen. Inpaint anything: Seg- ment anything meets image inpainting.arXiv preprint arXiv:2304.06790, 2023. 3

    work page arXiv 2023

  67. [67]

    Hermes: Human-to-robot embodied learning from multi-source mo- tion data for mobile dexterous manipulation.arXiv preprint arXiv:2508.20085, 2025

    Zhecheng Yuan, Tianming Wei, Langzhe Gu, Pu Hua, Tianhai Liang, Yuanpei Chen, and Huazhe Xu. Hermes: Human-to-robot embodied learning from multi-source mo- tion data for mobile dexterous manipulation.arXiv preprint arXiv:2508.20085, 2025. 3

    work page arXiv 2025

  68. [68]

    3D Diffusion Policy: Generalizable Visuomotor Policy Learning via Simple 3D Representations

    Yanjie Ze, Gu Zhang, Kangning Zhang, Chenyuan Hu, Muhan Wang, and Huazhe Xu. 3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations. arXiv preprint arXiv:2403.03954, 2024. 2

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  69. [69]

    GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models

    Aohan Zeng, Xin Lv, Qinkai Zheng, Zhenyu Hou, Bin Chen, Chengxing Xie, Cunxiang Wang, Da Yin, Hao Zeng, Jiajie Zhang, et al. Glm-4.5: Agentic, reasoning, and coding (arc) foundation models.arXiv preprint arXiv:2508.06471, 2025. 6, 8

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  70. [70]

    Activeumi: Robotic manipulation with active perception from robot-free human demonstrations.arXiv preprint arXiv:2510.01607,

    Qiyuan Zeng, Chengmeng Li, Jude St John, Zhongyi Zhou, Junjie Wen, Guorui Feng, Yichen Zhu, and Yi Xu. Activeumi: Robotic manipulation with active perception from robot-free human demonstrations.arXiv preprint arXiv:2510.01607,

    work page arXiv

  71. [71]

    Combo: compositional world models for embodied multi-agent cooperation.arXiv preprint arXiv:2404.10775, 2024

    Hongxin Zhang, Zeyuan Wang, Qiushi Lyu, Zheyuan Zhang, Sunli Chen, Tianmin Shu, Behzad Dariush, Kwonjoon Lee, Yilun Du, and Chuang Gan. Combo: compositional world models for embodied multi-agent cooperation.arXiv preprint arXiv:2404.10775, 2024. 2

    work page arXiv 2024

  72. [72]

    Doglove: Dexterous manipulation with a low-cost open-source haptic force feedback glove.arXiv preprint arXiv:2502.07730, 2025

    Han Zhang, Songbo Hu, Zhecheng Yuan, and Huazhe Xu. Doglove: Dexterous manipulation with a low-cost open-source haptic force feedback glove.arXiv preprint arXiv:2502.07730, 2025. 3

    work page arXiv 2025

  73. [73]

    Vision-language models for vision tasks: A survey.IEEE transactions on pattern analysis and machine intelligence, 46 (8):5625–5644, 2024

    Jingyi Zhang, Jiaxing Huang, Sheng Jin, and Shijian Lu. Vision-language models for vision tasks: A survey.IEEE transactions on pattern analysis and machine intelligence, 46 (8):5625–5644, 2024. 2

    work page 2024

  74. [74]

    Robot learning from any images

    Siheng Zhao, Jiageng Mao, Wei Chow, Zeyu Shangguan, Tianheng Shi, Rong Xue, Yuxi Zheng, Yijia Weng, Yang You, Daniel Seita, et al. Robot learning from any images. In Conference on Robot Learning, pages 4226–4245. PMLR,

    work page

  75. [75]

    Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware

    Tony Z Zhao, Vikash Kumar, Sergey Levine, and Chelsea Finn. Learning fine-grained bimanual manipulation with low- cost hardware.arXiv preprint arXiv:2304.13705, 2023. 2

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  76. [76]

    Tesseract: learning 4d embodied world models.arXiv preprint arXiv:2504.20995,

    Haoyu Zhen, Qiao Sun, Hongxin Zhang, Junyan Li, Siyuan Zhou, Yilun Du, and Chuang Gan. Tesseract: learning 4d embodied world models.arXiv preprint arXiv:2504.20995,

    work page arXiv

  77. [77]

    Extraneousness-aware imitation learning

    Ray Chen Zheng, Kaizhe Hu, Zhecheng Yuan, Boyuan Chen, and Huazhe Xu. Extraneousness-aware imitation learning. arXiv preprint arXiv:2210.01379, 2022. 2

    work page arXiv 2022

  78. [78]

    RoboDreamer: Learning Compositional World Models for Robot Imagination

    Siyuan Zhou, Yilun Du, Jiaben Chen, Yandong Li, Dit-Yan Yeung, and Chuang Gan. Robodreamer: Learning composi- tional world models for robot imagination.arXiv preprint arXiv:2404.12377, 2024. 2, 3

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  79. [79]

    Densematcher: Learning 3d semantic correspondence for category-level manipulation from a single demo.arXiv preprint arXiv:2412.05268,

    Junzhe Zhu, Yuanchen Ju, Junyi Zhang, Muhan Wang, Zhecheng Yuan, Kaizhe Hu, and Huazhe Xu. Dense- matcher: Learning 3d semantic correspondence for category- level manipulation from a single demo.arXiv preprint arXiv:2412.05268, 2024. 3

    work page arXiv 2024

  80. [80]

    Grs: Generating robotic simulation tasks from real-world images

    Alex Zook, Fan-Yun Sun, Josef Spjut, Valts Blukis, Stan Birchfield, and Jonathan Tremblay. Grs: Generating robotic simulation tasks from real-world images. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 594–603, 2025. 3 12 Appendix A. Single-Image Scene Reconstruction Details In this section, we describe how IGen reconstruc...

    work page 2025