Generative world models used as closed-loop test oracles require a five-level admissibility ladder (L0-L4) because visual fidelity does not predict action-robustness.
hub
World Model for Robot Learning: A Comprehensive Survey
20 Pith papers cite this work. Polarity classification is still indexing.
20
Pith papers citing it
abstract
World models, which are predictive representations of how environments evolve under actions, have become a central component of robot learning. They support policy learning, planning, simulation, evaluation, data generation, and have advanced rapidly with the rise of foundation models and large-scale video generation. However, the literature remains fragmented across architectures, functional roles, and embodied application domains. To address this gap, we present a comprehensive review of world models from a robot-learning perspective. We examine how world models are coupled with robot policies, how they serve as learned simulators for reinforcement learning and evaluation, and how robotic video world models have progressed from imagination-based generation to controllable, structured, and foundation-scale formulations. We further connect these ideas to navigation and autonomous driving, and summarize representative datasets, benchmarks, and evaluation protocols. Overall, this survey systematically reviews the rapidly growing literature on world models for robot learning, clarifies key paradigms and applications, and highlights major challenges and future directions for predictive modeling in embodied agents. To facilitate continued access to newly emerging works, benchmarks, and resources, we will maintain and regularly update the accompanying GitHub repository alongside this survey.
hub tools
citation-role summary
background 1
citation-polarity summary
years
2026 20roles
background 1polarities
background 1representative citing papers
FORGE decouples robotic tool-use into keypoint trajectory prediction from action-free data and action grounding from limited demonstrations, achieving over 2X improvement in functional generalization to unseen tools.
Argues for shifting to diagnosis-driven tension management of offline priors in online RL, supported by a framework on prior roles, experiments showing help-or-hurt reversals, and cross-domain evidence.
Qwen-AgentWorld are language world models that simulate multi-domain agent environments and boost general agent capabilities via decoupled RL simulation and unified foundation model training.
DREAM-Chunk uses test-time sampling and latent-world-model rollouts to select robust action chunks from chunking-based VLA policies, improving performance under stochastic dynamics on simulation and hardware tasks.
ω-EVA is a three-stage latent world model framework that trains action-conditioned dynamics, a language-conditioned flow policy, and a tri-branch refiner to improve embodied action generation in simulation.
FAWAM integrates force signals into perception, prediction, and closed-loop correction, raising success rates 36% over vision baselines in contact-rich manipulation tasks.
Flash-WAM introduces modality-specific consistency parametrizations to distill joint video-action diffusion models to single-step inference, delivering 23x speedup with preserved benchmark performance.
CLAW is an end-to-end self-supervised method that learns semantically meaningful continuous latent actions and predictive world models from action-free videos to support imitation learning and goal-directed planning.
Introduces world-task factorization for robot policies using Bayesian evidence and AICON graph plus learned modulator, outperforming baselines with zero-shot generalization in heterogeneous robotics settings.
Feedback world model closes the prediction-observation loop at inference time to correct errors and improve diffusion policy performance under distribution shift in robotics.
InSight enables autonomous acquisition of manipulation primitives in VLAs via automated segmentation for steerability and a VLM-guided data flywheel that generates and integrates new demonstrations for tasks like pouring and sweeping.
MemoryVAM integrates a Perceiver-based Recap Compressor and Cue Gate into video action models, raising success rates on long-horizon manipulation from 5% to 42.5% on LIBERO-Mem and 75-80% on real-robot counting, spatial recall, and tracking tasks.
GIVE improves pre-trained VLA models for robotic tasks by incorporating gestures via visual skeleton overlays and semantic descriptions, yielding 40% higher object recognition accuracy and 80% higher task success in real-world HRI experiments.
Efficient-WAM delivers 30x lower latency than prior WAMs at 100 ms per chunk while keeping competitive manipulation performance by treating coarse future video as guidance rather than high-fidelity output.
WorldDP combines a high-level object-centric world model for subgoal planning with a low-level diffusion policy for execution, claiming better performance than baselines on multi-stage robotic manipulation benchmarks.
Bridge-WA introduces a lightweight distillation-based world-action model that uses future-change priors to improve robotic task success and robustness without deployment-time dense rollouts.
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.
A survey that clarifies boundaries and organizes World Action Models by generation requirements and predictive substrates, identifying a trend toward generating less of the future.
The authors describe a logistics data flywheel that uses world models for out-of-distribution data synthesis in imitation learning and incorporates operational feedback for continual improvement.
citing papers explorer
-
Validate the Dream Before You Trust Its Verdict: Admissibility for World-Model Simulators
Generative world models used as closed-loop test oracles require a five-level admissibility ladder (L0-L4) because visual fidelity does not predict action-robustness.
-
FORGE: Towards Functional Tool-Use Generalization via Keypoint Trajectory Reasoning
FORGE decouples robotic tool-use into keypoint trajectory prediction from action-free data and action grounding from limited demonstrations, achieving over 2X improvement in functional generalization to unseen tools.
-
Beyond One-Size-Fits-All: Diagnosis-Driven Online Reinforcement Learning with Offline Priors
Argues for shifting to diagnosis-driven tension management of offline priors in online RL, supported by a framework on prior roles, experiments showing help-or-hurt reversals, and cross-domain evidence.
-
Qwen-AgentWorld: Language World Models for General Agents
Qwen-AgentWorld are language world models that simulate multi-domain agent environments and boost general agent capabilities via decoupled RL simulation and unified foundation model training.
-
DREAM-Chunk: Reactive Action Chunking with Latent World Model
DREAM-Chunk uses test-time sampling and latent-world-model rollouts to select robust action chunks from chunking-based VLA policies, improving performance under stochastic dynamics on simulation and hardware tasks.
-
$\omega$-EVA: Envision, Verify, and Act with Latent Interactive World Models
ω-EVA is a three-stage latent world model framework that trains action-conditioned dynamics, a language-conditioned flow policy, and a tri-branch refiner to improve embodied action generation in simulation.
-
FAWAM: Force-Aware World Action Models for Closed-Loop Contact-Rich Manipulation
FAWAM integrates force signals into perception, prediction, and closed-loop correction, raising success rates 36% over vision baselines in contact-rich manipulation tasks.
-
Flash-WAM: Modality-Aware Distillation for World Action Models
Flash-WAM introduces modality-specific consistency parametrizations to distill joint video-action diffusion models to single-step inference, delivering 23x speedup with preserved benchmark performance.
-
CLAW: Learning Continuous Latent Action World Models via Adversarial Latent Regularization
CLAW is an end-to-end self-supervised method that learns semantically meaningful continuous latent actions and predictive world models from action-free videos to support imitation learning and goal-directed planning.
-
World-Task Factorization for Robot Learning
Introduces world-task factorization for robot policies using Bayesian evidence and AICON graph plus learned modulator, outperforming baselines with zero-shot generalization in heterogeneous robotics settings.
-
Feedback World Model Enables Precise Guidance of Diffusion Policy
Feedback world model closes the prediction-observation loop at inference time to correct errors and improve diffusion policy performance under distribution shift in robotics.
-
InSight: Self-Guided Skill Acquisition via Steerable VLAs
InSight enables autonomous acquisition of manipulation primitives in VLAs via automated segmentation for steerability and a VLM-guided data flywheel that generates and integrates new demonstrations for tasks like pouring and sweeping.
-
MemoryVAM: Integrating Memory into Video Action Model for Robot Manipulation
MemoryVAM integrates a Perceiver-based Recap Compressor and Cue Gate into video action models, raising success rates on long-horizon manipulation from 5% to 42.5% on LIBERO-Mem and 75-80% on real-robot counting, spatial recall, and tracking tasks.
-
GIVE: Grounding Human Gestures in Vision-Language-Action Models
GIVE improves pre-trained VLA models for robotic tasks by incorporating gestures via visual skeleton overlays and semantic descriptions, yielding 40% higher object recognition accuracy and 80% higher task success in real-world HRI experiments.
-
Efficient-WAM: A 1B-Parameter World-Action Model with Low-Cost Future Imagination
Efficient-WAM delivers 30x lower latency than prior WAMs at 100 ms per chunk while keeping competitive manipulation performance by treating coarse future video as guidance rather than high-fidelity output.
-
Unifying Object-Centric World Models and Diffusion Policy: A Hierarchical Framework for Multi-Stage Robotic Tasks
WorldDP combines a high-level object-centric world model for subgoal planning with a low-level diffusion policy for execution, claiming better performance than baselines on multi-stage robotic manipulation benchmarks.
-
Bridge-WA: Predicting Where and How the World Changes for Robotic Action
Bridge-WA introduces a lightweight distillation-based world-action model that uses future-change priors to improve robotic task success and robustness without deployment-time dense rollouts.
-
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.
-
World Action Models: A Survey
A survey that clarifies boundaries and organizes World Action Models by generation requirements and predictive substrates, identifying a trend toward generating less of the future.
-
Towards a Data Flywheel for Embodied Intelligence in Logistics
The authors describe a logistics data flywheel that uses world models for out-of-distribution data synthesis in imitation learning and incorporates operational feedback for continual improvement.