Skip to main content

Advertisement

Table 2 The correctness of learned polices using RL and IRL methods in the test data set

From: Inverse reinforcement learning for intelligent mechanical ventilation and sedative dosing in intensive care units

Policy Overall Action Ventilation Sedative
π IBL 53.9% 99.7% 54.2%
π BL 53.5% 99.6% 53.9%
\(\phantom {\dot {i}\!}\pi _{{BL}_{1}}\) 23.5% 45.7% 51.0%
\(\phantom {\dot {i}\!}\pi _{{BL}_{2}}\) 14.1% 35.5% 39.1%
\(\phantom {\dot {i}\!}\pi _{{BL}_{3}}\) 17.2% 34.9% 54.1%