Mind the Gap: Promises and Pitfalls of Hierarchical Planning in LeWorldModel

Published July 15, 2026 · Category: Robotics

Overview

arXiv:2607.12547v1 Announce Type: new Abstract: We investigate whether temporal hierarchy can improve LeWorldModel on long-horizon goal-conditioned control. We introduce Hi-LeWM, an extension that freezes the pretrained low-level LeWM and adds high-level planning over latent subgoals. We evaluate Hi-LeWM on PushT and Cube across increasing goal offsets. Hierarchy does not automatically improve performance: at short horizons, the best configuration uses a one-step high-level horizon, while longer horizons reveal a mismatch between the learned high-level action space and the inference-time search distribution. Experiments with true future latent subgoals show that the frozen low-level controller can execute well-aligned intermediate targets, indicating that high-level subgoal generation is the main bottleneck. Unconstrained search can select latent macro-actions that appear favorable under the learned model but produce poor control targets. Constraining search around macro-actions encoded from training trajectories, with appropriate subgoal execution timing, recovers useful hierarchical regimes, improving over flat LeWM by +11.3 percentage points at medium-range horizons and +14.7 percentage points at the longest PushT horizon. Overall, temporal abstraction can benefit compact frozen LeWM, but only when high-level search remains compatible with the low-level controller

Source

Originally published at arxiv.org.

Related Articles

• Autonomous Tracking and Terminal Guidance of Moving Targets for Fixed-Wing UAVs
• Expert Knowledge-driven Reinforcement Learning for Autonomous Racing via Trajectory Guidance and Dynamics Constraints
• Analysis of Mutual and Referential Human and Robot Gazes in a Collaborative Word Association Game