Symposium DS01—Integrating Machine Learning and Simulations for Materials Modeling, Design and Manufacturing

This symposium aims to promote an integrated vision of material design—informed by data and channeled by physics-based simulations. Although numerical simulations have revolutionized materials design, they face several challenges, including high computing cost, limited accuracy, and limited potential for inverse design. Machine learning models also suffer from some limitations, e.g., need for large, consistent, and accurate datasets, questionable extrapolations, potential violations of physics and chemistry laws, and limited interpretability. In that regard, data-driven machine learning models and knowledge-driven simulations have the potential to inform, advance, and complement each other—and to address each other’s deficiencies. This symposium builds on the idea that the lack of meaningful integration between data- and knowledge-driven modeling is a missed opportunity in materials science. This symposium will explore new modeling approaches that seamlessly combine and integrate machine learning and simulations—wherein simulation informs machine learning, machine learning advances simulations, or closed-loop integrations thereof.

• Multi-fidelity models, data-fusion, and transfer learning approaches
• Machine learning to inform simulations (e.g., machine-learned interatomic forcefields)
• Physics-informed machine learning and symbolic learning
• "Self-driving" simulations, reinforcement learning, and robotic synthesis
• Graph neural networks for materials modeling
• Automatic differentiation, inverse problems, and deep generative models
• Machine learning for "finding needles in haystacks" in simulation output data
• Rare events sampling and automated identification of collective variables
• Machine learning for structural and topology optimization
• Machine-learned surrogate simulators
• Natural language processing for materials modeling
• Use of hardware dedicated to deep learning (e.g., TPUs) to accelerate simulations

• Christine Aikens (Kansas State University, USA)
• Raymundo Arroyave (Texas A&M University, USA)
• Alán Aspuru-Guzik (University of Toronto, Canada)
• Muratahan Aykol (Toyota Research Institute, USA)
• Amanda Barnard (The Australian National University, Australia)
• Peter Battaglia (DeepMind, United Kingdom)
• Miguel Bessa (Delft University of Technology, Netherlands)
• Souvik Chakraborty (Indian Institute of Technology Delhi, India)
• Maria Chan (Argonne National Laboratory, USA)
• Jacqueline Cole (University of Cambridge, United Kingdom)
• Ekin Dogus Cubuk (Google, USA)
• Payel Das (IBM T.J. Watson Research Center, USA)
• Marjolein Dijkstra (Utrecht University, Netherlands)
• George Em Karniadakis (Brown University, USA)
• Ian Foster (The University of Chicago, USA)
• Rodrigo Freitas (Massachusetts Institute of Technology, USA)
• Rafael Gomez-Bombarelli (Massachusetts Institute of Technology, USA)
• Bjork Hammer (Arhus University, Denmark)
• N M Anoop Krishnan (Indian Institute of Technology Delhi, India)
• Emine Kucukbenli (Harvard University, USA)
• Artrith Nongnuch (Columbia University, USA)
• Rampi Ramprasad (Georgia Institute of Technology, USA)
• Subramanian Sankaranarayanan (University of Illinois at Chicago, USA)
• Yizhou Sun (University of California, Los Angeles, USA)
• Rama Vasudevan (Oak Ridge National Laboratory, USA)
• Xiaonan Wang (National University of Singapore, Singapore)
• Jie Xu (Argonne National Laboratory, USA)
• Lusann Yang (Google, USA)
• Tarek Zohdi (University of California, Berkeley, USA)