AI and ML for Arts, Science, and Engineering
The use of artificial intelligence and machine learning, closely integrated with domain expertise in the arts, sciences, and engineering, to accelerate research and scholarship, to make data-driven discoveries, and to enable entirely new artistic and scientific tools.
AI and machine learning are having a revolutionary impact on scholarship and research. They are enabling the development of models of unprecedented scale and accuracy, analysis of large and diverse data sets, optimization and acceleration of experiments, and discovery of patterns that lead to new insights and solutions. With the help of IDEaS, Georgia Tech researchers adapt the latest advances in AI to work in the context of their areas of expertise and make discoveries that would not be possible otherwise.
AI/ML in Arts, Sciences, and Engineering brings together data-driven learning, scientific computing, and domain expertise to build models that can (1) emulate or augment physics-based simulations, (2) interpret complex measurements, (3) recommend what experiment or simulation to run next, and (4) support discovery and design in fields from molecular science and astrophysics to art. Landmark advances in the broader field illustrate the potential: for example, modern deep learning systems can dramatically improve protein structure prediction, helping close long-standing gaps between available biological sequences and experimentally determined structures.
Across government, academia, and industry, AI is increasingly regarded as a “force multiplier” for discovery because it can extract insights from large datasets, lower the costs of exploring design options, and accelerate the process from hypothesis to validation. In energy and physical sciences, agencies like the U.S. Department of Energy highlight AI’s ability to speed up work on urgent challenges. Societal benefits include faster development of medicines and materials, improved climate and environmental modeling, more resilient infrastructure and manufacturing, and broader access to advanced research capabilities through shared national resources.
Innovation Through Expertise
IDEaS combines three key strengths: (1) AI methodology (2) domain-driven problem selection where better models are immediately impactful, and (3) the infrastructure to operationalize AI for real research workflows. The following projects supported by IDEaS focus on integrating AI and domain research.
Musicianship Group Robotic
To analyze live musical input, and to create, improvise, and perform with human collaborators. Our robots can listen like humans but improvise like machines, aiming to inspire humans to play and think about music in new ways.
Cybershuttle
Building platforms that make advanced compute/AI usable end-to-end, so researchers can move from data to workflows to A user-facing cyberinfrastructure that provides seamless access to distributed resources for researchers, integrating local, campus, cloud, and NSF-funded national-scale computing centers to boost productivity and enable end-to-end scientific workflows.
Custos
Open-source security middleware for science gateways enabling trustworthy and scalable collaboration around scientific digital artifacts.
VizFold
A research initiative focused on making protein-folding models, including AlphaFold-like system, interpretable and explainable via new visualization and analysis methods, integrating scalable platforms for interactive exploration.
Affiliated Research Centers
Center for Artificial Intelligence in Science and Engineering (ARTISAN)
ARTISAN accelerates science and engineering by integrating cutting-edge AI techniques and supporting researchers in operationalizing workflows with cyberinfrastructure. Cybershuttle, for example, is a user-facing environment that integrates local and campus resources, cloud services, and NSF-funded national computing centers to support end-to end scientific workflows, with a strong user-centered design.
Learn More
Contact for AI & ML 4 Science & Engineering | Suresh Marru; Director, ARTISAN