Wang, K. (2026). Visualising backward information propagation in deep reinforcement learning from a variational data assimilation perspective. Scientific Reports. https://doi.org/10.1038/s41598-026-42086-x