人工智能在材料发现中的应用OA

This paper reviews the application of artificial intelligence(AI)in materials discovery,focusing on the latest advancements in three core areas:material characterization,generation,and property prediction.First,the principles of descriptor design for material characterization were introduced,emphasizing the transfor-mation of complex information,such as micro-structure and chemical composition,into machine-readable quan-titative features.Then,an in-depth analysis of the technical pathways employed by generative models in the inverse design of materials was provided,detailing the specific applications and advantages of the generative adversarial networks(GANs),graph neural networks(GNNs),variational autoencoders(VAEs)diffusion models,and large language models(LLMs).Furthermore,material property prediction technologies based on GNN and Transformer architectures were reviewed,summarizing the performance of these methods in terms of prediction accuracy,computational efficiency,and model interpretability.Finally,the core challenges and potentials of AI-driven materials discovery were discussed,offering insights into future research directions.