Researchers from ISMN-CNR have introduced GrapheNet, a deep learning framework based on an Inception-Resnet architecture using image-like encoding of structural features for the prediction of the properties of nanographenes.
Scheme of the GrapheNet framework. Image from Scientific Reports
By exploiting the planarity of quasi-bidimensional systems and through encoding structures into images, and leveraging the flexibility and power of deep learning in image processing, Graphenet is said to achieve significant accuracy in predicting the physicochemical properties of nanographenes.
This approach is able to efficiently encode structures composed of hundreds of atoms, scaling efficiently with the size of the model and enabling the prediction of the properties of large systems, which contrasts with the limitations of current atomistic-level representations for deep learning applications.
The approach proposed based on image encoding exhibits a significant numerical accuracy and outperforms the computational efficiency of current representations of materials at the atomistic level, with significant advantages especially in the representation of nanostructures and large planar systems.
To develop the GrapheNet approach, the team exploited the topological correlation between the quasi-2D morphology of nanographenes and the standard encoding of images. The GrapheNet framework was tested on datasets of graphene oxide (GO) and defected nanographene (DG) samples.
The GrapheNet framework provides accurate predictions of key electronic properties of nanographenes, outperforming the computational efficiency of current representation methods.
