Dual-Mind World Models: A General Framework for Learning in Dynamic Wireless Networks

arXiv 25.10 2025 58.4 method, application

TLDR

Proposes dual-mind world model combining pattern and logic learning for long-term link scheduling in dynamic wireless networks.

Reasoning

Strengths include a novel cognitive psychology-inspired framework that addresses data inefficiency and short-sightedness, enabling reasoning without observations. Weaknesses are the abstract's truncation, unclear real-world validation, and focus on a specific V2X scenario.

Read-first score

Read-first score 58.4, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 44.

Recency 8%

86.7

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

Methodology quality 25%

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

Topical relevance 42%

62.9

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

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

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 272.

Keyword Scores

world model

model-based reinforcement learning world model

world dynamics prediction

generative world model

world simulator

interactive world model

video world model

Deep Analysis

Innovations

Dual-mind world model inspired by cognitive psychology with pattern-driven System 1 and logic-driven System 2 components
End-to-end differentiable imagined trajectories with logical consistency for long-term link scheduling
Joint reasoning of network states and planning link scheduling through imagination rollouts
Capability to make efficient decisions during intervals without observations

Methodology

The proposed dual-mind world model framework integrates a pattern-driven System 1 and a logic-driven System 2 to learn dynamics and logic of the wireless network. Link scheduling is learned via end-to-end differentiable imagined trajectories with logical consistency over an extended horizon, avoiding reliance on direct environment interactions. The model is evaluated on a realistic simulator based on Sionna with real-world physical channel, ray-tracing, and scene objects with material properties, comparing against state-of-the-art RL baselines and a System 1-only world model.

Key Results

The proposed world model achieves a significant improvement in data efficiency and demonstrates strong generalization and adaptation to unseen environments compared to state-of-the-art RL baselines and the System 1-only world model.

Resources

Paper