JOPAT jointly models pixels, point tracks, and actions in a diffusion transformer and reports gains over pixel-only baselines on long-horizon robot tasks with occlusion and off-screen motion.
hub Canonical reference
Masked visual pre-training for motor control
Canonical reference. 100% of citing Pith papers cite this work as background.
30
Pith papers citing it
Background
100% of classified citations
abstract
This paper shows that self-supervised visual pre-training from real-world images is effective for learning motor control tasks from pixels. We first train the visual representations by masked modeling of natural images. We then freeze the visual encoder and train neural network controllers on top with reinforcement learning. We do not perform any task-specific fine-tuning of the encoder; the same visual representations are used for all motor control tasks. To the best of our knowledge, this is the first self-supervised model to exploit real-world images at scale for motor control. To accelerate progress in learning from pixels, we contribute a benchmark suite of hand-designed tasks varying in movements, scenes, and robots. Without relying on labels, state-estimation, or expert demonstrations, we consistently outperform supervised encoders by up to 80% absolute success rate, sometimes even matching the oracle state performance. We also find that in-the-wild images, e.g., from YouTube or Egocentric videos, lead to better visual representations for various manipulation tasks than ImageNet images.
hub tools
citation-role summary
background 5
citation-polarity summary
roles
background 5polarities
background 5representative citing papers
π₀.₇ is a steerable generalist robotic model that uses rich multimodal prompts including language, subgoal images, and performance metadata to achieve out-of-the-box generalization across tasks and robot bodies.
UMI enables zero-shot deployment of robot manipulation policies trained solely on portable human demonstrations captured with custom handheld grippers, supporting dynamic bimanual tasks across novel environments and objects.
A collaborative dataset spanning 22 robots and 527 skills enables RT-X models that transfer capabilities across different robot embodiments.
VIP learns a visual embedding from human videos whose distance defines dense, smooth rewards for arbitrary goal-image robot tasks without task-specific fine-tuning.
CAIP learns action-aligned visual representations via contrastive pre-training on human hand keypoints from egocentric video, outperforming DINOv2, SigLIP, MVP, and R3M with >30% gains on real dexterous manipulation tasks.
ACE-Ego-0 is a VLA pretraining framework that turns egocentric human videos into robot-format pseudo-actions via a video-to-action pipeline and trains jointly with robot data under a reliability-aware objective.
T-Rex introduces a large tactile dataset and MoT architecture that achieves over 30% higher success rates than baselines on 12 tasks requiring force control and deformable object handling.
DynaFLIP pre-trains dynamics-aware image encoders by aligning image, language, and 3D flow modalities through simplex-volume minimization plus regularizers on video triplets, yielding reusable backbones that improve manipulation policies by up to 22.5% in out-of-distribution settings.
GAP pre-trains the spatial adapter on a lightweight simulated proxy task with free object masks to generate repeatable geometric keypoints, yielding higher success rates than baselines in low-data robotic manipulation on RoboMimic and ManiSkill.
PALM improves long-horizon robotic manipulation success by distilling affordance representations for object interaction and predicting within-subtask progress in a VLA model.
Uni-Hand forecasts 2D/3D hand waypoints, head motion, and contact states in egocentric views using vision-language fusion and dual-branch diffusion, with new benchmarks for downstream robotics and action tasks.
Training models to generate videos of robot actions produces policies that generalize better to new objects and tasks while using far less demonstration data than standard behavior cloning.
π_{0.5} is a VLA model that achieves long-horizon dexterous manipulation in entirely new homes through co-training on heterogeneous tasks and multi-source data including web and semantic predictions.
DINO-WM builds world models on pre-trained DINOv2 features to enable zero-shot planning from offline data without rewards or demonstrations.
GR-2 pre-trains on web-scale videos then fine-tunes on robot data to reach 97.7% average success across over 100 manipulation tasks with strong generalization to new scenes and objects.
Gen2Act enables generalizable robot manipulation for unseen objects and novel motions by using zero-shot human video generation from web data to condition a policy trained on an order of magnitude less robot interaction data.
A GPT-style model pre-trained on large video datasets achieves 94.9% success on CALVIN multi-task manipulation and 85.4% zero-shot generalization, outperforming prior baselines.
Lift3D-VLA integrates 3D point cloud encoding and temporal action modeling into Vision-Language-Action models, achieving higher success rates on simulated and real-world robotic manipulation tasks.
DexAC-WM improves FID, FVD, and PCK in high-DoF action-conditioned video prediction via structured action modeling and semantic grounding on EgoDex and EgoVerse.
FAME combines a factor-aware MoE with frozen pretrained encoders via staged adapter training and joint fine-tuning, reporting 34% gains on Meta-World and 35% in real-world pick-and-place under environmental changes.
An end-to-end policy learns robust humanoid locomotion directly from noisy depth images via high-fidelity sensor simulation, vision-aware distillation from privileged maps, and terrain-specific multi-critic reward shaping.
MSDP pre-trains a transformer encoder with masked multisensory autoencoding, then uses an asymmetric actor-critic bridge (cross-attention for critic, pooling for actor) to accelerate and robustify contact-rich RL across simulation and real robots.
GR-3 is a VLA model that generalizes to novel objects, environments, and abstract instructions, outperforms the π0 baseline, and integrates with the new ByteMini bi-manual mobile robot.
citing papers explorer
-
Point Tracking Improves World Action Models
JOPAT jointly models pixels, point tracks, and actions in a diffusion transformer and reports gains over pixel-only baselines on long-horizon robot tasks with occlusion and off-screen motion.
-
${\pi}_{0.7}$: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities
π₀.₇ is a steerable generalist robotic model that uses rich multimodal prompts including language, subgoal images, and performance metadata to achieve out-of-the-box generalization across tasks and robot bodies.
-
Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots
UMI enables zero-shot deployment of robot manipulation policies trained solely on portable human demonstrations captured with custom handheld grippers, supporting dynamic bimanual tasks across novel environments and objects.
-
Open X-Embodiment: Robotic Learning Datasets and RT-X Models
A collaborative dataset spanning 22 robots and 527 skills enables RT-X models that transfer capabilities across different robot embodiments.
-
VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training
VIP learns a visual embedding from human videos whose distance defines dense, smooth rewards for arbitrary goal-image robot tasks without task-specific fine-tuning.
-
Contrastive Action-Image Pre-training for Visuomotor Control
CAIP learns action-aligned visual representations via contrastive pre-training on human hand keypoints from egocentric video, outperforming DINOv2, SigLIP, MVP, and R3M with >30% gains on real dexterous manipulation tasks.
-
ACE-Ego-0: Unifying Egocentric Human and Robotic Data for VLA Pretraining
ACE-Ego-0 is a VLA pretraining framework that turns egocentric human videos into robot-format pseudo-actions via a video-to-action pipeline and trains jointly with robot data under a reliability-aware objective.
-
T-Rex: Tactile-Reactive Dexterous Manipulation
T-Rex introduces a large tactile dataset and MoT architecture that achieves over 30% higher success rates than baselines on 12 tasks requiring force control and deformable object handling.
-
DynaFLIP: Rethinking Robotics Perception via Tri-Modal-Dynamics Guided Representation
DynaFLIP pre-trains dynamics-aware image encoders by aligning image, language, and 3D flow modalities through simplex-volume minimization plus regularizers on video triplets, yielding reusable backbones that improve manipulation policies by up to 22.5% in out-of-distribution settings.
-
GAP: Geometric Anchor Pre-training for Data-Efficient Visuomotor Learning of Manipulation Tasks
GAP pre-trains the spatial adapter on a lightweight simulated proxy task with free object masks to generate repeatable geometric keypoints, yielding higher success rates than baselines in low-data robotic manipulation on RoboMimic and ManiSkill.
-
PALM: Progress-Aware Policy Learning via Affordance Reasoning for Long-Horizon Robotic Manipulation
PALM improves long-horizon robotic manipulation success by distilling affordance representations for object interaction and predicting within-subtask progress in a VLA model.
-
Uni-Hand: Universal Hand Motion Forecasting in Egocentric Views
Uni-Hand forecasts 2D/3D hand waypoints, head motion, and contact states in egocentric views using vision-language fusion and dual-branch diffusion, with new benchmarks for downstream robotics and action tasks.
-
Video Generators are Robot Policies
Training models to generate videos of robot actions produces policies that generalize better to new objects and tasks while using far less demonstration data than standard behavior cloning.
-
$\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization
π_{0.5} is a VLA model that achieves long-horizon dexterous manipulation in entirely new homes through co-training on heterogeneous tasks and multi-source data including web and semantic predictions.
-
DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning
DINO-WM builds world models on pre-trained DINOv2 features to enable zero-shot planning from offline data without rewards or demonstrations.
-
GR-2: A Generative Video-Language-Action Model with Web-Scale Knowledge for Robot Manipulation
GR-2 pre-trains on web-scale videos then fine-tunes on robot data to reach 97.7% average success across over 100 manipulation tasks with strong generalization to new scenes and objects.
-
Gen2Act: Human Video Generation in Novel Scenarios enables Generalizable Robot Manipulation
Gen2Act enables generalizable robot manipulation for unseen objects and novel motions by using zero-shot human video generation from web data to condition a policy trained on an order of magnitude less robot interaction data.
-
Unleashing Large-Scale Video Generative Pre-training for Visual Robot Manipulation
A GPT-style model pre-trained on large video datasets achieves 94.9% success on CALVIN multi-task manipulation and 85.4% zero-shot generalization, outperforming prior baselines.
-
Lift3D-VLA: Lifting VLA Models to 3D Geometry and Dynamics-Aware Manipulation
Lift3D-VLA integrates 3D point cloud encoding and temporal action modeling into Vision-Language-Action models, achieving higher success rates on simulated and real-world robotic manipulation tasks.
-
Not All Actions Are Equal: Rethinking Conditioning for Dexterous World Model
DexAC-WM improves FID, FVD, and PCK in high-DoF action-conditioned video prediction via structured action modeling and semantic grounding on EgoDex and EgoVerse.
-
Factor-Aware Mixture-of-Experts with Pretrained Encoder for Combinatorial Generalization
FAME combines a factor-aware MoE with frozen pretrained encoders via staged adapter training and joint fine-tuning, reporting 34% gains on Meta-World and 35% in real-world pick-and-place under environmental changes.
-
Now You See That: Learning End-to-End Humanoid Locomotion from Raw Pixels
An end-to-end policy learns robust humanoid locomotion directly from noisy depth images via high-fidelity sensor simulation, vision-aware distillation from privileged maps, and terrain-specific multi-critic reward shaping.
-
Self-Supervised Multisensory Pretraining for Contact-Rich Robot Reinforcement Learning
MSDP pre-trains a transformer encoder with masked multisensory autoencoding, then uses an asymmetric actor-critic bridge (cross-attention for critic, pooling for actor) to accelerate and robustify contact-rich RL across simulation and real robots.
-
GR-3 Technical Report
GR-3 is a VLA model that generalizes to novel objects, environments, and abstract instructions, outperforms the π0 baseline, and integrates with the new ByteMini bi-manual mobile robot.
-
Self-Predictive Representations for Combinatorial Generalization in Behavioral Cloning
BYOL-γ uses self-predictive representations to approximate successor representations, improving zero-shot combinatorial generalization in goal-conditioned behavioral cloning.
-
Intention-Conditioned Flow Occupancy Models
InFOM applies flow matching to model intention-conditioned occupancy measures for RL pre-training, reporting 1.8x median return gains and 36% higher success rates on benchmarks.
-
Robot Self-Improvement via Human-Video Dynamics Models
Human-video dynamics models enable cross-embodiment robot self-improvement via training-free Dynamics-Guided Action Correction, raising success rates from 40% to 81% on seven real-world tasks.
-
Multimodal Fusion for Sim2real Transfer in Visual Reinforcement Learning
A multimodal RGB-depth fusion backbone with vision transformer, masked-token contrastive learning, and curriculum domain randomization outperforms baselines in simulation and enables zero-shot real-world robot manipulation.
-
From Human Videos to Robot Manipulation: A Survey on Scalable Vision-Language-Action Learning with Human-Centric Data
The paper surveys four classes of techniques that derive action-related supervision from human videos for VLA robot models and identifies three open challenges in episode structuring, embodiment grounding, and evaluation.
- Action-Effect Memory Pretraining for Robot Manipulation