Symposium BI01-Democratizing AI in Materials Science—A Pathway to Broaden the Impact of Materials Research

This symposium aims to democratize and streamline materials science by lowering the barriers to adopting data-driven artificial intelligence techniques. Materials research is essential for technological advancements, but can be slow and resource-intensive. The utilization of AI methodologies provides a promising avenue for expediting materials research. Nevertheless, obstacles to embracing these approaches exist. Through collaborative discussions and innovative exploration, our symposium seeks to diminish these obstacles and foster a more accessible and efficient landscape for adopting data-driven artificial intelligence techniques in materials science. By democratizing materials science with AI, we mean to increase the visibility, availability, readiness, and user-friendliness of data, tools, platforms, and innovative concepts that are made available through various means such as web applications, Python packages, GitHub, or other sharing platforms. Our aim is to foster a collaborative and open discussion between data science experts and non-experts. This includes materials scientists and engineers, chemists, physicists, and computer scientists from academia and industry. Through these efforts, we hope to catalyze materials research and facilitate breakthroughs in areas ranging from sustainability to healthcare.

The discussion will revolve around data, tools, platforms, frameworks, and pioneering ideas that can accelerate materials research. It also includes the use of data-driven approaches for educational, explorative, accelerative, disseminative, and knowledge-preservative purposes. We expect our symposium to provide a forum to identify adoption barriers encountered by non-data experts.

Topics will include:

Data extraction, organization, curation, and storage, and materials ontologies
Experimental and computational databases and sharing platforms
High-throughput materials space exploration techniques with computations and experiments
Numerical materials representations (fingerprints or descriptors)
AI-guided experimentation
Knowledge discovery, conservation, and dissemination
Rule mining
Synthesis, prediction, and design strategies
Synergetic materials research strategies that combine experimentation and theory
Accelerating materials discovery with human-assisted AI methods
Machine learning techniques such as active learning, transfer learning, and large language models
Boosting materials research through open-source datasets and software
Broadening impact through outreach, societal interaction, and education

Invited Speakers:

Milad Abolhasani (North Carolina State University, USA)
Maria Chan (Argonne National Laboratory, USA)
Steve Cranford (Cell Press, USA)
Claudia Draxl (Humboldt-Universität zu Berlin, Germany)
Matthew Evans (Université catholique de Louvain, Belgium)
Alysia Garmulewicz (Universidad de Santiago de Chile, Chile)
Neil Gershenfeld (Massachusetts Institute of Technology, USA)
Gabe Gomes (Carnegie Mellon University, USA)
Ivor Lončarić (Institut Ruder Boškovic, Croatia)
Arun Mannodi-Kanakkithodi (Purdue University, USA)
Nicola Marzari (École Polytechnique Fédérale de Lausanne, Switzerland)
Adnan Mehonic (University College London, United Kingdom)
Kristin Persson (University of California, Berkeley, USA)
Lilo Pozzo (University of Washington, USA)
Krishna Rajan (University at Buffalo, The State University of New York, USA)
Kristin Schmidt (IBM T.J. Watson Research Center, USA)
Rama Vasudevan (Oak Ridge National Laboratory, USA)
James Warren (National Institute of Standards and Technology, USA)