| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Computational Discovery and Design of Materials
|
| Presentation Title |
Representation-based Generative Models for Materials |
| Author(s) |
Victor Fung |
| On-Site Speaker (Planned) |
Victor Fung |
| Abstract Scope |
Data-driven machine learning methods can greatly accelerate materials discovery and design over conventional human-guided approaches. Specifically, generative models could be used for inverse design by generating new materials samples with desired properties. However, when applying generative models for atomic structures, suitable structural fingerprints or representations will be needed which are analogous to the graph-based or SMILES representations used in molecular generation. Ideally these representations should be invariant to translations, rotations, and permutations, while remaining invertible back to their Cartesian coordinates. The challenges associated with simultaneously meeting both invariance and invertibility requirements have prompted us to propose an alternative approach to this problem by developing methods for accurately reconstructing the structure using optimization-based techniques which can be applied towards non-invertible representations. Our recent findings show this approach can reliably reconstruct atomic structures with high accuracy, and when paired with a generative model, may be used to efficiently produce new atomic structures. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Modeling and Simulation, Computational Materials Science & Engineering |