Structural-Test Geometry as an Inductive Bias for Partially Observable World Models

概要

Predictive world-model objectives need not organize latent histories according to distinctions required for later control. With privileged simulator access during training, should response-derived structural information be predicted, decoded, or imposed directly on latent geometry--and how do these choices compare with structured belief distillation? We define posterior-marginalized structural tests whose response distributions induce a pseudometric over histories, then train quotient-preserving (QP) world models by aligning shared latent Euclidean distances to that metric. Across a fixed six-task diagnostic suite, the largest differences appear under held-out structural variation: QP improves planning success by 26.1 points $[21.5,30.7]$ over metric decoding and 6.4 points $[3.8,9.0]$ over structured belief distillation. On GridWorld, metric decoding accurately recovers held-out structural distances ($ρ=.91$) while its native latent geometry remains more distorted than QP's (paired IQD difference $.22$ $[.20,.24]$), showing that relational recoverability and direct geometric alignment can diverge.

参考リンク

• Paper ↗

リンク先のライセンス・再現性・データ条件をご確認ください。自動分類や実装の目安は保証ではありません。

実装を始めるプロンプト

参考リンクを確認したうえで、次の内容をコピーして利用できます。

# 論文・記事
Structural-Test Geometry as an Inductive Bias for Partially Observable World Models

# 概要
Predictive world-model objectives need not organize latent histories according to distinctions required for later control. With privileged simulator access during training, should response-derived…

# 原典
https://arxiv.org/abs/2610.38903

上記の手法を実装してください。実行前に依存関係とライセンスを確認してください。