AI for Materials Science
- [University of California at Berkeley]
- Overview
Artificial intelligence and automation are accelerating the development of science and chemistry by helping scientists choose which experiments to conduct and focus on promising new materials.
- Why it matters: These fields are under pressure to produce new materials faster and cheaper to support and advance technologies that could transform industries and economies.
- The big picture: Batteries, drugs, and semiconductors require new materials and molecules to underpin green grids, precision medicine, and next-generation computing and communications.
Ongoing initiatives are underway in the United States, China, the European Union, and Japan to stimulate materials development by creating libraries of compounds that can be tested and potentially developed into new materials. Twenty years ago, the United States led the world in publishing papers in this field, but by this measure, China now leads the field of materials science research.
- Data Analytics and Machine Learning in Materials Science
Artificial intelligence (AI) seems to be finding its way into every field these days, and the field of materials science is no exception.
Data analytics and machine learning researchers are exploring and advancing applications “big data” in materials science. They develop and use sophisticated data mining, data analytics, and machine learning approaches to extract the maximum possible information from high-throughput simulations and experiments that produce volumes of data too large for manual analysis by humans.
They develop tools to identify patterns in images and other data to identify the most promising materials and processes from enormous numbers of possibilities at an unprecedented scale.
- [Bagan, Myanmar]
- Applications of AI in Materials Science
Artificial intelligence (AI) technology promises to accelerate material discovery through high-throughput computation and high-throughput experimentation. The application of artificial intelligence (AI) tools such as machine learning, deep learning, and various optimization techniques is critical to making this happen.
Some of the key application areas for applying AI techniques to materials include: development of well-curated and diverse datasets, selection of effective material representations, inverse material design, integration of autonomous experiments and theory, and selection of appropriate algorithms/workflows. The idea of including physics-based models in AI frameworks is also appealing.
Finally, uncertainty quantification in AI-based material property prediction and issues related to building the infrastructure for disseminating AI knowledge are critical to making AI-based materials research successful.
- AI Research Topics in Materials Science
Topics addressed in this AI for Materails Science include (but not be limited to):
- Dataset and tools for employing AI for materials
- Integrating experiments with AI techniques
- Graph neural network
- Comparison of AI techniques for materials
- Challenges applying AI to materials
- Uncertainty quantification and building trust in AI predictions
- Generative modeling
- Using AI to develop classical force-fields
- Natural language processing
[More to come ...]
