Data-centric materials design in the quantum regime: motif learning and symmetry-guided discovery

Via Zoom

Materials design in the quantum regime call for the integration of multi-tier materials information that go beyond atomic structures. Many quantum behaviors are greatly controlled by local symmetries and local bonding environments. In this talk, motivated by Pauling’s rules, I will show that local bonding environments (motifs) can be incorporated in a graph-based machine learning architecture to make reliable property predictions for solid-state quantum materials including complex metal oxides. I will demonstrate that the unsupervised machine (Atom2Vec and Motif2Vec) can learn the basic properties of atoms and motifs by themselves from the extensive database of known materials. Clustering of atoms and motifs in vector space classifies them into meaningful groups consistent with human knowledge. The proposed atom-motif dual network model demonstrates the feasibility to incorporate beyond-atom materials information in a graph network framework and achieves the state-of-the-art performance in predicting the complex properties of solid-state quantum materials. With these tools developed, I will discuss the potential application of artificial intelligence in the field of quantum materials design, including two-dimensional quantum materials and defect qubit-based quantum information science. I will also discuss the continued development of AI-driven technologies for quantum phenomena, with the consideration of symmetries, orbital interactions, and physical constraints.

Host: Greg Stewart