TY  - RPRT
TI  - Behavior-Invariant Task Representation Learning with Transformer-based World Models for Offline Meta-Reinforcement Learning
AU  - Fuyuan Qian
AU  - Menglong Zhang
AU  - Song Wang
AU  - Quanying Liu
PY  - 2026
UR  - https://arxiv.org/abs/2606.00780
ID  - 2606.00780
ER  -