MahdiehNejati / RL-Set-Valued-Policy

Implementation for "Clinician-in-the-Loop Decision Making: Reinforcement Learning with Near-Optimal Set-Valued Policies" (Tang, Modi, Sjoding, Wiens). ICML 2020

Clinician-in-the-Loop Decision Making: Reinforcement Learning with Near-Optimal Set-Valued Policies

Source code submission for ICML 2020

Notes

• The implementation of the main algorithms can be found under synthetic/lib
• Run notebooks in synthetic to recreate all synthetic experiments (Figures 1-5)
• Follow instructions in mimic-sepsis to recreate the real data experiment (Figure 6-7)