A Graph Theory – Based Molecular Generative Model for Practical Use in Chemical Industries

Abstract

Material informatics has been attracting great attentions across broad industrial domains. Especially molecular generative model is important. However, most of existing solutions do not address industrial practicality, because they need a heavily loaded pre-training by large amount of data and the models are uninterpretable. In this paper, we will present our algorithm-based molecular generative model tool, in which pre-training is unrequired thereby models are fully interpretable and fine-controllable in atomistic level.