What Drives Success in Physical Planning with Joint-Embedding Predictive World Models?

arXiv 26.1 2026 53.5 method

TLDR

Investigates design choices for joint-embedding predictive world models (JEPA-WMs) in physical planning, outperforming baselines on simulated and real-world tasks.

Reasoning

Strengths include a comprehensive ablation study of key components and validation on both simulated and real-world robotic data. Weaknesses are that the abstract does not detail specific limitations or failure cases, and the core contribution is incremental within the JEPA family.

Read-first score

Read-first score 53.5, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 34.

Recency 6%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Citation impact 18%

78.8

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.78812949

Methodology quality 18%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=baseline,experiment

Reproducibility 18%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=checkpoint,code,github

Topical relevance 29%

48.6

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Citation velocity 12%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FoundationFrontier

Rank sensitivity

Stability: volatile; rank range: 234.

Keyword Scores

world model

model-based reinforcement learning world model

world dynamics prediction

interactive world model

world simulator

generative world model

video world model

Deep Analysis

Innovations

Characterization of a family of world models as Joint-Embedding Predictive World Models (JEPA-WMs)
Comprehensive study of key components (model architecture, training objective, planning algorithm) within the JEPA-WM family
Proposed model that outperforms DINO-WM and V-JEPA-2-AC in navigation and manipulation tasks

Methodology

The paper conducts experiments using both simulated environments and real-world robotic data. It systematically studies the effects of model architecture, training objective, and planning algorithm on planning success. The proposed model is compared against two established baselines, DINO-WM and V-JEPA-2-AC, on navigation and manipulation tasks.

Key Results

The proposed model outperforms both DINO-WM and V-JEPA-2-AC in navigation and manipulation tasks.

Resources

Paper