Machine Learning for Optimizing the Optical Propertiesof Quantum Dots

The application of machine learning (ML) to the synthesis of quantum dots (QDs) represents a significant innovation in material optimization. Due to the quantum confinement effect, the optical properties of QDs can be tuned by controlling the size of the material during synthesis, but this traditionally requires a long time of experimental synthesis practice. ML is an efficient approach that analyzes databases of synthesis variables to predict final properties, thereby reducing the experimental effort required. Supervised and unsupervised models play complementary roles, in this context, helping to predict properties or identify patterns in unlabeled data. In this chapter, the main steps that include the construction of structured databases, feature engineering, and model training for accurate predictions are presented, highlighting the potential of ML in developing QDs with specific optical properties.