Kuo-Ying Wang. "Visualising backward information propagation in deep reinforcement learning from a variational data assimilation perspective." Scientific Reports, 2026. doi:10.1038/s41598-026-42086-x