cs.RONov 11, 2026

The Posterior-Prior Epistemic Gap: When World Model Imagination Fails to Carry Task-Relevant Signals

Authors: Jaerock Kwon, Elahe Delavari, Donghyun Kim, Haewoon Nam

Organizations: Department of Electrical and Computer Engineering, University of Michigan–Dearborn · Department of Electrical Engineering, Hanyang University, South Korea

Abstract

Model-based planners that roll out imagined futures assume the transition model faithfully propagates task-relevant signals---a property rarely measured directly. In a single-seed case study of a Recurrent State Space Model world model with active-inference planning in the CARLA driving simulator, we identify a posterior-prior gap---posterior (observation-filtered) states retain obstacle information that imagined rollout states progressively lose: a supervised obstacle probe achieves AUC 0.85 on posterior states, but the same probe applied to imagined prior states shows monotonic decay (AUC 0.85→\rightarrow0.71 over 15 steps; logit separation drops 57%). Because the probe is identical, the decay isolates the transition model---not the representation---as the information bottleneck. Obstacle avoidance drops to 0% without privileged runtime coordinates, and the planner's own ensemble-based epistemic term does not flag the loss---ensemble disagreement yields AUC 0.52 as an obstacle classifier (chance level), making the failure an unflagged imagination failure---an unknown unknown at planning time. As a secondary contribution, we show that task-conditioned weighting of Effective Free Energy components over the same world model yields lane keeping (81.2±1.881.2 \pm 1.8% route completion) and obstacle avoidance (28/28 avoided, using privileged coordinates) in the CARLA simulator, with runtime weight modulation enabling post-evasion centering.

Figures & tables

Explore similar work

CardsList
  1. EpicWorldModel: Exploration-driven Planning with Latent World Models

    Oct 5, 2026Bowen Feng, Julian Ost, May Mei +2Partially Observable RLWorld Model-Based Planning