“…First principle methods utilizing quantum chemistry computations augmented with statistical analysis, such as conductor-like screening model for realistic solvents (COSMO-RS) [ 26 , 27 , 28 ], deserve special attention. In general, this method has been found to be successful in modeling various pharmaceutically relevant properties and characteristics, including solubility [ 29 , 30 , 31 , 32 ], partition coefficients [ 33 , 34 , 35 ], acid–base properties [ 36 , 37 , 38 ], and co-crystallization abilities [ 27 , 39 , 40 , 41 , 42 , 43 ]. In recent years, the application of neural networks, including deep learning, to solubility modeling has increased [ 44 , 45 , 46 ].…”