COSMO-NET: Fast and Accurate Machine Learning Surrogates for COSMO-based Molecular Descriptors

Accurate predictive thermodynamic models are essential tools for the computational design of new molecules. Models based on COnductor-like Screening Models (COSMO), such as COSMO-SAC and COSMO-RS, are well-suited for this purpose but require computationally expensive quantum mechanical (QM) calculations to generate key properties-the surface charge density distribution (p(σ)), the surface area (A) and the cavity volume (V)-for each molecule. Here we develop graph-based neural network surrogate models, based on a directed message passing neural network (DMPNN) and a graph convolutional network (GCN), to predict these molecular properties rapidly and accurately. We train the models on a dataset of over 16,000 compounds generated through an automated QM calculation pipeline. The DMPNN model outperforms GCN for probability of surface charge density prediction, while the GCN shows higher accuracy for surface area and cavity volume. We build on these strengths to propose a hybrid model, COSMO-NET. When applied to predict octanol-water partition coefficients, COSMO-NET achieves a mean absolute error (MAE) of 0.31 compared to 0.34 and 0.36 for DMPNN and GCN, respectively. These results demonstrate that machine learning surrogates can replace costly QM calculations while maintaining accuracy, supporting the discovery of new molecules and the evaluation of their performance.