Lucid Dreaming for World Models: Learning to Doubt Imagination and Decide by Trust

2026-09-29

AI Summary

This study addresses the unreliability of imagined planning in world models, where overconfidence in unseen state-action pairs misleads decision-making. To mitigate this, we propose LucidWM, a framework that introduces subjective logic to assign degrees of doubt to categorical latent transitions, enabling uncertainty estimation without additional parameters. By accumulating trust over multi-step trajectories to reweight returns, the method integrates doubt into the imagination process, thereby guiding reinforcement learning policy optimization. The effectiveness of this approach is validated against four baseline models and seventeen uncertainty readouts. In navigation tasks, LucidWM reduces the number of steps required for goal attainment from 362 to 190, significantly enhancing the robustness of trust-based decision-making.

Abstract

World models enable agents to learn and plan in imagination, but predictions beyond their experience can become unreliable and mislead decisions. Existing uncertainty estimates derived from predictions can remain overconfident on unfamiliar state-action pairs. We propose the Lucid World Model (LucidWM), which learns doubt from experience and propagates trust through imagination. By integrating Subjective Logic into categorical latent transitions, LucidWM distinguishes predicted outcomes from their evidential support and assigns each transition a degree of doubt. The complement of this doubt defines transition-level trust, which accumulates multiplicatively along imagined trajectories to reweight returns for policy learning and guide action selection. Uncertainty estimation requires no additional parameters or forward passes. Evaluated on four base world models against seventeen uncertainty readouts, LucidWM detects environmental changes and signals uncertainty during action-corrupted rollouts. In a controlled navigation case study, acting on trust reduces the number of steps required to reach the goal from 362 to 190. Fifteen demonstration videos show how LucidWM doubts its dreams and acts on that doubt. Videos are available at https://lucidwm.github.io.

Problem

• World Models
• Uncertainty Estimation
• Overconfidence
• Imagination-based Planning

Innovation

• World Models
• Subjective Logic
• Uncertainty Estimation
• Trust Propagation
• Reinforcement Learning