On the Capability Separation Between World-Model Policy Learning and Imitated World-Action Models

• URL: http://arxiv.org/abs/2608.22197v1
• Date: Sun, 23 Aug 2026 03:25:31 GMT
• Title: On the Capability Separation Between World-Model Policy Learning and Imitated World-Action Models
• Abstract summary: We compare a direct behavior-cloning policy, an imitation-trained world-action policy, and a policy optimized with an action-conditioned world model.<n>Future prediction changes the learning factorization but not the unrestricted external policy class or ideal imitation target.
• Score: 3.79187263097166
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: World-action models predict a future outcome and then infer an associated action. Although this factorization can improve representation learning and data efficiency, it is unclear whether it provides stronger control capability than direct behavior cloning when both are trained from the same observational demonstrations. We compare a direct behavior-cloning policy, an imitation-trained world-action policy, and a policy optimized with an action-conditioned world model. At the controller-class level, every world-action policy can be flattened into a direct stochastic policy with the same closed-loop trajectory distribution. At the population level, under realizability, exact optimization, common deployment information, and distribution-preserving deployment, direct behavior cloning and world-action imitation both recover the observational behavior policy. Thus, future prediction changes the learning factorization but not the unrestricted external policy class or ideal imitation target. Action-conditioned world-model learning differs by predicting outcomes under specified actions and comparing them through a control objective. We characterize the irreducible action-specific prediction error of future models that do not condition on the candidate action, identify conditions under which a world-action joint can recover an interventional forward model, and show that observational demonstrations do not identify action effects in general. Finally, we construct an environment family in which every observational learner has positive worst-case regret, whereas one informative intervention permits zero regret. The key distinction is therefore between predicting futures associated with observed behavior and predicting consequences of specified actions for policy optimization.

Related papers

• Large Discrete Policy: Advancing Explicit Behavior Modeling with Stochastic Iterative Scoring [77.17706435067372]
  Large Policy (LDiP) is a fully discrete behavior modeling framework that selects actions from a large vocabulary of physically plausible candidates.<n>Rather than perturbing actions, LDiP improves expressivity through iterative scoring.<n>Across end-to-end planning, closed-loop driving, robotic manipulation, and vision-language-action settings, LDiP consistently outperforms strong discrete and continuous baselines.
  arXiv  Detail & Related papers  (2026-09-07T05:12:52Z)
• Do Robotic World Models Really Follow Actions? Diagnosing and Aligning Action-Conditioned Generation for Policy Learning [53.77164209208635]
  We introduce WorldEcho, which probes action following over a broader action distribution using visual integrity and trajectory alignment.<n>We propose WorldSync, which strengthens action following along three axes: distributional coverage, representational grounding, and intervention-effect alignment.<n>Experiments on RoboTwin benchmarks and real-robot tasks show that WorldSync improves WorldEcho metrics and serves as a more reliable simulator for iterative policy improvement.
  arXiv  Detail & Related papers  (2026-08-25T17:59:49Z)
• Predicting Consequences and Reinforcing Navigation Policies with Latent World Models [31.865100130203317]
  We propose a compatibility prediction Latent World Model (LWM) for robot navigation.<n>Our key insight is that spatial proximity correlates with latent feature similarity, enabling action consequences to be evaluated directly in latent space.<n>Our approach significantly outperforms prior world model and imitation learning methods in prediction accuracy, policy learning, and real-world navigation performance.
  arXiv  Detail & Related papers  (2026-08-23T11:58:42Z)
• IADD-TR: Intervention-Aware Dynamics Decoupling with Targeted Regularization for Model-Based Reinforcement Learning [57.16513298507272]
  We propose IADD-TR, a unified framework combining Intervention-Aware Dynamics Decoupling (IADD) and Targeted Regularization (TR)<n>IADD factorizes transitions into an action-intervention stage and an action-free natural evolution stage, using a zero-action anchor to resolve the non-uniqueness of this two-stage factorization for robust generalization.<n>For policy learning, we derive TR from the efficient influence function of a replay-state policy-gradient functional.
  arXiv  Detail & Related papers  (2026-08-11T08:20:02Z)
• On the Identifiability of Controlled World Models [34.31084112465058]
  This paper presents a joint identifiability condition for controlled world models with Gaussian latent states.<n>We prove that when this condition holds, minimizing the LeJEPA-style predictive objective can recover both latent states and controlled dynamics.<n>We also characterize an attainable amplification of counterfactual prediction error that scales inversely with the weakest conditional action-excitation margin.
  arXiv  Detail & Related papers  (2026-07-24T15:49:12Z)
• Feedback World Model Enables Precise Guidance of Diffusion Policy [36.965417653906535]
  We propose feedback world model, a new paradigm that closes the loop between prediction and observation at inference time.<n>We show that our method substantially improves both prediction accuracy and policy performance under distribution shift.<n>In particular, it reduces world model prediction error by up to 76.4% and improves out-of-distribution (OOD) success rate by 30%.
  arXiv  Detail & Related papers  (2026-05-15T07:52:13Z)
• HarmoWAM: Harmonizing Generalizable and Precise Manipulation via Adaptive World Action Models [58.191567345416836]
  World Action Models (WAMs) have emerged as a promising paradigm for robot control by modeling physical dynamics.<n>HarmoWAM fully leverages a world model to unify predictive and reactive control, enabling both general transit and precise manipulation.<n>We construct three training-unseen test environments across six real-world robotic tasks, covering variations in background, position, and object semantics.
  arXiv  Detail & Related papers  (2026-05-11T17:59:56Z)
• PROWL: Prioritized Regret-Driven Optimization for World Model Learning [20.10187986360715]
  We introduce a KL-constrained adversarial curriculum in which a policy is trained to expose high-error trajectories of a diffusion-based world model.<n>We implement our approach in the MineRL framework and evaluate it on held-out out-of-distribution trajectories.
  arXiv  Detail & Related papers  (2026-05-11T14:24:19Z)
• AGWM: Affordance-Grounded World Models for Environments with Compositional Prerequisites [18.635313325437664]
  We propose AGWM (Affordance-Grounded World Model), which learns an abstract affordance structure represented as a DAG of prerequisite dependencies to explicitly track the dynamic executability of actions.<n> Experiments on game-based simulated environments demonstrate the effectiveness of our method by achieving lower multi-step prediction error, better generalization to novel configurations, and improved interpretability.
  arXiv  Detail & Related papers  (2026-05-07T18:46:44Z)
• World Action Verifier: Self-Improving World Models via Forward-Inverse Asymmetry [82.93104394404781]
  General-purpose world models promise scalable policy evaluation, optimization, and planning.<n>We propose World Action Verifier (WAV), a framework that enables world models to identify their own prediction errors and self-improve.
  arXiv  Detail & Related papers  (2026-04-02T12:48:36Z)
• ACT-JEPA: Novel Joint-Embedding Predictive Architecture for Efficient Policy Representation Learning [90.41852663775086]
  ACT-JEPA is a novel architecture that integrates imitation learning and self-supervised learning.<n>We train a policy to predict action sequences and abstract observation sequences.<n>Our experiments show that ACT-JEPA improves the quality of representations by learning temporal environment dynamics.
  arXiv  Detail & Related papers  (2025-01-24T16:41:41Z)
• Policy-Guided Diffusion [30.4597043728046]
  In many real-world settings, agents must learn from an offline dataset gathered by some prior behavior policy. We propose policy-guided diffusion as an alternative to autoregressive offline world models. We show that policy-guided diffusion models a regularized form of the target distribution that balances action likelihood under both the target and behavior policies.
  arXiv  Detail & Related papers  (2024-04-09T14:46:48Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.