TY  - RPRT
TI  - Diagnosing Faults in Reinforcement Learning Simulators and World Models with Canonical Polynomial Invariants
AU  - Tesfay Zemuy Gebrekidan
AU  - Hadush Hailu Gebrerufael
PY  - 2026
UR  - https://arxiv.org/abs/2609.13194
ID  - 2609.13194
ER  -