Extended Data Table 2 Comparison of generated set composition of each agent trained during SrIOP and REINVENT for TPSA optimization

From: Testing the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learning