The authors investigate the ability of machine learning models to developing new drug-like molecules by learning desired chemical properties versus simply generating molecules that similar to those in the training set.
Read More...Evaluating the feasibility of SMILES-based autoencoders for drug discovery