Latent Dynamics · 2026-09-30
Lucid Dreaming for World Models: Learning to Doubt Imagination and Decide by Trust
Ziqi Wen, Ting Xu, Lianyu Wang, Xian Lin, Yanda Meng, Huazhu Fu, Meng Wang, Ching-Yu Cheng
Auto-summarized: this summary was generated by a language model from the paper’s text and has not been reviewed by an editor. Check the paper before relying on it. Benchmark numbers appear only after a human has verified them.
TL;DR
LucidWM is a world model that learns “doubt” about imagined transitions from experience using Subjective Logic, then turns doubt into “trust” that compounds along imagination. Trust reweights λ-returns for learning and guides action choice. It needs no extra parameters or forward passes for uncertainty estimation, and the authors report earlier alarms for drift and fewer steps to reach a goal in a navigation study.
Why it matters
World-model planning can be misled by overconfident predictions on unfamiliar state-action pairs, especially when uncertainty signals ignore whether a transition is supported by experience and when trust depends on the full imagined rollout. The authors’ doubt/trust mechanism targets this by separating predictive shape from evidential support and discounting imagined rewards when trust collapses, affecting both learning and decision-making.
Method
- Learn doubt from experience by treating the categorical transition head as a Subjective Logic opinion from its logits; doubt is derived from the evidential support (Dirichlet uncertainty mass) rather than output spread alone.
- During imagination, propagate trust through rollouts by multiplying transition-level trust and using it to reweight the λ-return, reducing reliance on steps preceded by low-trust transitions.
- Decide by scoring candidate imagined futures with the trust-weighted objective; reward beyond a trust collapse can be vetoed (set to zero contribution).
Limitation
Uncertainty estimation requires no additional parameters or forward passes, but the paper text does not state other limitations on settings, architectures, or compute beyond the described experimental protocol.
Abstract (from arXiv)
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.