VLWMs learn variable-length action-conditioned dynamics in latent space with curriculum training, yielding 13% average gains over prior latent world models on long-horizon tasks.
Hierarchical Planning with Latent World Models
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
abstract
World models are a promising path to zero-shot embodied control through planning. However, existing world model planners struggle on long-horizon, multi-stage tasks: prediction errors compound and naive search is exponential in the planning horizon. Hierarchy mitigates both by decomposing tasks into shorter, tractable subproblems; yet prior hierarchical approaches either amortize control into task-specific policies (hierarchical RL) or assume low-dimensional states and known dynamics (classical hierarchical MPC). We present Hierarchical Planning with Latent World Models (HWM), an architecture and planning paradigm for hierarchical model predictive control (MPC) directly on visual world models trained solely via next-latent prediction. HWM learns world models at multiple temporal scales within a shared latent space, so predictions from the long-horizon model serve as subgoals for the short-horizon model via latent matching, without task-specific rewards, skill learning, or hierarchical policies. To keep long-horizon search tractable, HWM learns an action encoder that compresses primitive action chunks into latent macro-actions. On real-world Franka manipulation, HWM solves pick-and-place from a single goal image at 70% success vs. 0% for single-level planning. Across simulated push manipulation and maze navigation, HWM consistently improves performance on long-horizon tasks while requiring up to 3x less planning compute.
citation-role summary
background 1
citation-polarity summary
years
2026 9roles
background 1polarities
background 1representative citing papers
World models succeed when their latent states are built to meet task-specific sufficiency constraints rather than preserving the maximum amount of information.
Graph world-model rollout error splits into a spectral topology factor and a model-norm factor, and Error-Aware training improves long-horizon stability when structure is fixed or evolving.
FF-JEPA introduces a two-model hierarchical structure with an action-free latent planner to decompose long-horizon planning into short subgoals in latent world models.
BISON learns bilevel policies over symbolic world models to generalize long-horizon robotic planning beyond VLA and end-to-end baselines while remaining efficient even at 10,000-object scale.
AdaJEPA performs closed-loop test-time adaptation of latent world models during MPC by executing an action chunk, observing the transition, and taking one gradient step on the model before replanning, yielding higher goal-reaching success.
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.
Embodied AI requires query-conditioned world models that select the simplest physical abstraction sufficient to answer intervention queries.
Survey organizing world models for robotic manipulation into representation families, a functional taxonomy, and infrastructure roles across pretraining, post-training, and inference, while reviewing 34 datasets and evaluation protocols.
citing papers explorer
-
Beyond the Next Step: Variable-Length Latent World Models for Long-Horizon Planning
VLWMs learn variable-length action-conditioned dynamics in latent space with curriculum training, yielding 13% average gains over prior latent world models on long-horizon tasks.
-
Latent State Design for World Models under Sufficiency Constraints
World models succeed when their latent states are built to meet task-specific sufficiency constraints rather than preserving the maximum amount of information.
-
Understanding Rollout Error in Graph World Models
Graph world-model rollout error splits into a spectral topology factor and a model-norm factor, and Error-Aware training improves long-horizon stability when structure is fixed or evolving.
-
FF-JEPA: Long-Horizon Planning in World Models with Latent Planners
FF-JEPA introduces a two-model hierarchical structure with an action-free latent planner to decompose long-horizon planning into short subgoals in latent world models.
-
Learning Bilevel Policies over Symbolic World Models for Long-Horizon Planning
BISON learns bilevel policies over symbolic world models to generalize long-horizon robotic planning beyond VLA and end-to-end baselines while remaining efficient even at 10,000-object scale.
-
AdaJEPA: An Adaptive Latent World Model
AdaJEPA performs closed-loop test-time adaptation of latent world models during MPC by executing an action chunk, observing the transition, and taking one gradient step on the model before replanning, yielding higher goal-reaching success.
-
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.
-
Physically Viable World Models: A Case for Query-Conditioned Embodied AI
Embodied AI requires query-conditioned world models that select the simplest physical abstraction sufficient to answer intervention queries.
-
World Models for Robotic Manipulation: A Survey
Survey organizing world models for robotic manipulation into representation families, a functional taxonomy, and infrastructure roles across pretraining, post-training, and inference, while reviewing 34 datasets and evaluation protocols.