Foundations of AI and ML
Foundations of AI includes topics such as AI theory, AI systems, combinations of reasoning and learning, and verification of AI systems that impact all of AI. This area is important because it underpins all of AI. We have theorists working on ML theory, systems researchers focused on scaling up the training and inference pipelines of AI systems, and formal verification and automated reasoning researchers focused on verifying AI as well as combining learning and reasoning systems.
Foundations of AI includes topics such as AI theory, AI systems, combinations of reasoning and learning, and verification of AI systems that impact all of AI. This area is important because it underpins all of AI. We have theorists working on ML theory, systems researchers focused on scaling up the training and inference pipelines of AI systems, and formal verification and automated reasoning researchers focused on verifying AI as well as combining learning and reasoning systems.
Given the increasing impact of AI on all aspects of society, it is important to study its theoretical and systems underpinnings, as well as ensure that these systems behave reliably and as intended. The Foundations of AI research area studies topics that help us build more scalable and reliable AI systems. The benefits for society are many: when we better understand these AI systems both from a theory and systems perspective, it enables us to improve their efficiency and scalability, while research at the intersection of formal methods and AI helps us make them more robust.
IDEaS supports research at the intersection of AI and formal methods, combinations of reasoning and learning, as well as AI systems by providing hardware credits, directing funding towards faculty working in these areas, and organizing relevant workshops and seminars. For example, IDEaS facilitates access to the Nexus supercomputing system, critical in advancing their research. Many of our faculty members are also leading researchers in the Foundations of AI research area.
Data & Infrastructure
We aim to create and deploy AL/ML tools and workflow practices tailored explicitly for a range of scientific research disciplines. We develop trustworthy and interpretable AI models, and the key data sets that underlie them. We seek to support the Georgia Tech research community by working with faculty to develop infrastructure to streamline research workflows, reduce redundancies and ensure reproducibility, in scientific AI applications.
Machine Learning
The foundation of ML is the development of models with the ability to learn to make concise predictions based on prior data input and experience. At IDEaS we specialize in integrating machine learning models with advanced fields like quantum chemistry and physical sciences to process large datasets for scalable and domain-specific applications.
Ethical Impacts
We aim to develop and deploy domain-knowledge informed AI tools that make accurate predictions and provide meaningful insights to developing scientific theories and knowledge. Our research is aimed at making complex AI/ML models explainable and transparent for scientific discovery by employing innovative methods to improve model interpretability in scientific research.
Learn More
Foundations of Artificial Intelligence Seminar Series
Contact for Foundations of AI/ML | Vijay Ganesh; Associate Director; IDEaS | vganesh@gatech.edu