TY  - RPRT
TI  - Action-Consequence Alignment for Reliable Planning and Self-Improving in Latent World Models
AU  - Jinping Wang1
AU  - Zhiqiang Gao
AU  - Xiantong Zhen
AU  - Ling Shao
PY  - 2026
UR  - https://arxiv.org/abs/2610.04539
ID  - 2610.04539
ER  -