Chiral nanostructure prediction model for chiral characteristic optical materials in biological art

Chiral features are a common physical structural feature in nature and are often applied in biological art. Traditional chiral nanostructure prediction methods have low efficiency in dealing with complex optical materials and are difficult to meet high-precision requirements. To address this issue and improve the accuracy of predicting chiral nanostructures, this study utilises machine learning algorithms to construct a deep learning-based convolutional neural network model for predicting the circular dichroism of chiral nanostructures. Moreover, by fusing multiple deep learning models, the accuracy of the prediction model has been improved. Results showed the Stacking ensemble model achieved 94.15% accuracy, 11.99 to 18.48% higher than other models, and closely matched actual circular dichroism spectra. This study enhances the prediction accuracy of chiral nanostructures, providing crucial support for the design and application of nanomaterials in biological art.