This focus topic will bring together leaders in the rapidly growing field of data science, artificial intelligence, and machine learning (AI/ML) for materials, processes, and interfaces to drive scientific discovery. AI, ML, and deep learning (DL) are being utilized to understand materials at the atomic scale, discover new scientific laws, and even design the next generation of advanced microelectronics for AI/ML. As researchers from academia to industry search for more effective means of advancing technology, AI/ML is being utilized as a means to reduce the burden on resources that have long relied on traditional experiments and computationally heavy modeling and simulation. This focus topic will bring together the community to disseminate the latest advances in the field, discuss challenges, and share future directions for AI & ML.
Areas of Interest: AI/ML is seeking abstracts in the following areas of interest:
- Driving Scientific Discovery through AI/ML: developing and evaluating new materials/processes/devices with AI/ML to reduce experimental design and computationally expensive modeling; methods for utilizing AI/ML to predict performance (e.g., materials, devices, etc.);
- Experimental Design for the Age of Big Data: design of experiments, testing, and data collection to maximize data generation and improve through-put and fundamental understanding; developing data sets and tools for training models; autonomous experiments and testing; methods for data management; model quality, uncertainty quantification and trust in AI models
- AI/ML for Characterization, including Synthetic Data Generation, from Materials to Systems: applying physics-based models; extraction of physicochemical information; microscopy, spectroscopy, etc.
- AI for AI: catapulting next-generation semiconductors and devices for AI/ML by utilizing AI/ML to drive device design, neuromorphic computation, non-von Neumann architecture, and Beyond Moore
- AI/ML vs Physical Principles: In contrast to AI/ML, the possibility of discovering fundamental approaches to predicting material properties is also being explored (e.g., the discovery of a topological periodic table).
AIML1: AI/ML for Scientific Discovery Oral Session
Invited Speakers:
- Brad Boyce, Sandia National Laboratories, ” BeyondFingerprinting: ML-guided process optimization using high-throughput experiments and simulations”
- Noa Marom, Carnegie Mellon University
- Colin Ophus, Lawrence Berkeley Lab
AIML2: AI/ML for Scientific Discovery Poster Session