Public Research Seminar by Advanced Materials Thrust, Function Hub - AI learning materials synthesis from literature
Supporting the below United Nations Sustainable Development Goals:支持以下聯合國可持續發展目標：支持以下联合国可持续发展目标：
Synthesis prediction is a key accelerator for the rapid design of advanced materials. However, determining synthesis variables such as the choice of precursor materials, operations, and conditions is challenging for inorganic materials because the sequence of reactions during heating is not well understood. This talk will introduce a data-centric approach that leverages decades of historical knowledge to automatically learn the experimental design for synthesizing novel target materials. The marriage of natural language processing and materials science enables large-scale retrieval of structured materials synthesis information from millions of scientific publications. From this comprehensive synthesis knowledge base, chemical insights of materials similarity can be learned and applied for predictive synthesis with a high success rate. This literature-based approach constitutes a significant part of materials informatics and an important step toward the autonomous synthesis of inorganic materials.
Tanjin He earned his Ph.D. in Materials Science and Engineering from UC Berkeley, under the supervision of Prof. Gerbrand Ceder. He continues his research as a postdoctoral scholar in the same group. His research primarily focuses on AI for materials science. He utilizes natural language processing to distill scientific knowledge automatically from various sources, including journal articles and patents. Through big-data-driven analysis and physics-based explainable machine learning, he develops predictive algorithms for the rational design and autonomous synthesis of advanced materials. His research has been published in high-impact journals including Nature, Science Advances, and Nature Communications, and has been cited over 1,000 times, earning him an h-index of 18.