Title: Predicting the future applications of any stoichiometric inorganic material through learning from past literature

Abstract: Through learning from past literature, artificial intelligence models have been able to predict the future applications of various stoichiometric inorganic materials in a variety of subfields of materials science. This capacity offers exciting opportunities for boosting the research and development (R&D) of new functional materials. Unfortunately, the previous models can only provide the prediction for existing materials in past literature, but cannot predict the applications of new materials. Here, we construct a model that can predict the applications of any stoichiometric inorganic material (regardless of whether it is a new material). Historical validation confirms the high reliability of our model. Key to our model is that it allows the generation of the word embedding of any stoichiometric inorganic material, which cannot be achieved by the previous models. This work constructs a powerful model, which can predict the future applications of any stoichiometric inorganic material using only a laptop, potentially revolutionizing the R&D paradigm for new functional materials