Symposium MT06-Generative AI Meets Materials Modeling—Emerging Opportunities and Challenges

With materials modeling and simulations playing a more prominent role in all aspects of materials research, from nanoscale to macroscale, new simulations are being performed every day, generating tremendous amount of data in a variety of formats. At the same time, generative artificial intelligence (AI), such as large language models (LLMs), has incubated revolutionary changes in every corner of our life, where materials research is not an exception. When generative AI meets materials modeling, we are about to see the dawn of achieving inverse materials design, thus now it is a critical moment for us to clarify those emerging trends as well as aware the accompanying new challenges, which motivates us to organize the proposed symposium. One part of the symposium will focus on the development of novel algorithms and models for constructing meaningful descriptors to interpret multi-modal simulation data at different scales – which should be the foundation for us to develop new generation generative models, to reveal structure-property relationships. It is of interest to demonstrate how the efforts, including physics-based, data-centric, or hybrid approaches, contribute to the understanding and engineering of material systems with complex compositions, structures, morphologies and geometries. The symposium will also provide a platform to showcase new concepts and protocols for us to better learn from simulation data, with a special focus on the recent progresses of large language models and their applications in materials research. Relevant topics include but are not limited to data augmentation, feature enhancement, generative models, autonomous workflow design, active learning, transfer learning, multi-fidelity learning, contrastive learning in materials modeling. Finally, the symposium will be interested in innovative methods for capturing thermodynamics and kinetics during simulated materials processes and phenomena, and how machine learning can reform these computations. Abstracts are welcomed on reporting latest progresses of free energy calculations, enhanced sampling, kinetic pathway recognition, prediction of critical conditions, etc.

artificial intelligence autonomous Computing kinetics machine learning modeling simulation