ABSTRACT Organic compounds play a crucial role in various industrial applications, and their properties, such as surface tension, significantly impact their performance. Designing new organic compounds with desired properties is a challenging task, and machine learning (ML) has emerged as a promising approach to accelerate this process. In this study, we developed an ML model to predict the surface tension of organic compounds using 2048‐bit fingerprints. Among various algorithms, XGBoost demonstrated the best performance with an R 2 score of 0.87. Leveraging this model, we designed 910 new organic compounds with surface tension values as high as 35. The data was visualized using SALI scatter plots, providing insights into the chemical space of the designed compounds. Furthermore, we calculated the synthetic accessibility scores for these compounds and identified 20 promising candidates for experimental synthesis. This work showcases the potential of ML in designing new organic compounds with desired properties, paving the way for accelerated materials discovery and development.
Kyhoiesh et al. (2026) studied this question.