MLIC Lab’s (Prof. Seungyul Han) paper accepted to ICLR 2026!_Strict Subgoal Execution: Reliable Long-Horizon Planning in Hierarchical Reinforcement Learning