Learning Systems Group

Creating scalable tools to build models from data.

The Learning Systems Group pioneers the next generation of artificial intelligence by designing and deploying sophisticated algorithms across the full spectrum of computational platforms. Our research addresses the fundamental challenge of co-designing algorithmic complexity with hardware capabilities to solve real-world applications. We develop novel machine learning models, deep learning architectures, and data-driven methods optimized for performance, whether running on a remote sensor platform, a national supercomputer, or a quantum computer. This unique approach allows us to push the boundaries of what's possible in a wide array of critical application domains.

We apply our expertise to solve complex, real-world problems with significant societal impact. For national security, we create advanced systems for threat intelligence and data fusion. In toxicology and health sciences, our algorithms accelerate discovery of environmental related health issues and predict adverse health outcomes from chemical exposure. We build state-of-the-art natural language processing (NLP) tools to mine and synthesize knowledge from millions of scientific publications and create knowledge graphs that accelerate discovery. We leverage evolutionary algorithms to navigate vast search spaces and automatically optimize the performance, architecture, and parameters of our artificial intelligence models. Our work in autonomous vehicles focuses on creating energy efficient perception that minimizes energy consumption from computing. By simulating the interactions of individual actors, we use agent-based modeling to unravel the complexity of large-scale systems and construct dynamic digital twins of physical processes and environments. Additionally, we develop new breakthroughs through advanced scientific data analysis, enabling researchers to find meaningful patterns in massive datasets, from law enforcement to renewable energy sources.

Driven by a passion for solving grand challenges that benefit society, our group brings together a vibrant, interdisciplinary team of research staff, postdoctoral fellows, and students from fields such as artificial intelligence, computer science, and computer engineering. We are committed to developing intelligent systems that can learn, reason, and operate effectively using any computational environment. Consequently, the Learning Systems Group has been recognized with seven R&D 100 Awards, and four nominations for the Association of Computing Machinery Gordon Bell Award where two of those nominations were finalist selections.