| Abstract Scope |
Traditional glass discovery relies on trial and error approaches, leading to a design-to-deploy period of 20-30 years. Here, we will discuss the application of artificial intelligence (AI) and machine learning (ML) in accelerating glass discovery. Specifically, three aspects where AI and ML can be used: (i) ML models for glass property predictions, (ii) natural language processing (NLP) for extracting information from the glass literature, and (iii) reinforcement learning for glass structure optimization. The first focuses on developing interpretable ML models for predicting properties of inorganic and chalcogenide glasses. The second focuses on MatSciBERT, the first materials-aware language model. We will also discuss how MatSciBERT can be used to extract information regarding composition and properties from the glass literature. Third, we will discuss how to optimize glass structures using reinforcement learning. Altogether, the talk will cover various aspects of AI and ML for accelerating glass discovery. |