World Models as Reference Trajectories for Rapid Motor Adaptation

arXiv 25.5 2025 48.6 method, application

TLDR

A dual control framework using world model predictions as reference trajectories for rapid motor adaptation in changing dynamics.

Reasoning

The paper introduces a novel dual architecture that separates long-term RL from rapid latent control, achieving faster adaptation with low computational cost. However, the abstract lacks explicit mention of real-world experiments or benchmarks, and the evaluation appears limited to simulated continuous control tasks.

Read-first score

Read-first score 48.6, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 36.

Recency 8%

86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Topical relevance 42%

51.4

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

Methodology quality 25%

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

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 446.

Keyword Scores

world model

world dynamics prediction

model-based reinforcement learning world model

world simulator

generative world model

interactive world model

video world model

Deep Analysis

Innovations

Using world model predictions as implicit reference trajectories for rapid adaptation
Dual control framework separating long-term reward maximization via reinforcement learning and robust motor execution via rapid latent control
Combining flexible policy learning with rapid error correction capabilities

Methodology

Reflexive World Models (RWM) is a dual control framework that uses world model predictions as implicit reference trajectories. It separates the control problem into long-term reward maximization through reinforcement learning and robust motor execution through rapid latent control. The approach is evaluated on high-dimensional continuous control tasks under varying dynamics, compared to model-based RL baselines.

Key Results

RWM achieves significantly faster adaptation with low online computational cost compared to model-based RL baselines, while maintaining near-optimal performance.

Resources

Paper