>

Machine Learning

The Machine Learning Area advances the principles, algorithms, and systems that enable machines to learn from data and experience efficiently. We connect mathematical and computational foundations with applications across science, engineering, health, and society.

Mission Statement

Our mission is to deepen the understanding of learning and develop methods that expand what intelligent systems can achieve. We seek to grow a community with depth across the full spectrum of machine learning, from theoretical foundations to algorithm design and applied research. Fundamental questions about how models represent knowledge, generalize, adapt, and make decisions guide our work. Challenges arising in practice motivate new theories, methods, and measures of progress.

Our faculty bring complementary strengths in mathematical machine learning, optimal transport, geometry and topology, optimization, representation learning, graph learning, foundation models, and decision-making under uncertainty. These strengths connect to research in medical imaging, neuroscience, surgical robotics, physical simulation, societal infrastructure, and beyond. Building on this breadth, we will strengthen the exchange between foundational research and applications, developing general principles while testing their value in complex settings.

Education is central to this mission. Through foundational courses, advanced topics, and project-based learning, we prepare students to understand the principles behind machine learning, develop and evaluate new methods, and collaborate across disciplines. We aim to cultivate researchers and practitioners who can connect mathematical reasoning with computational experimentation and turn advances in learning into scientific and practical contributions.

Research Pillars

Four interconnected research pillars guide this vision. Faculty, students, and projects may contribute to several pillars, connecting theoretical questions with algorithmic advances and challenges arising in practice.

  • Mathematical and Computational Foundations of Learning

    Advance the theory that explains how learning works, when it succeeds, and where its limits lie. We aim to deepen expertise in statistical learning theory, optimization, generalization, and computational complexity, while developing connections to geometry, topology, probability, and dynamical systems. This work provides a basis for understanding existing models and creating new approaches.

  • Learning Algorithms, Representations, and Foundation Models

    Develop models and algorithms that discover structure, represent knowledge, and support reasoning and decision-making. Our interests span generative modeling, graph and multimodal learning, reinforcement learning, and foundation models. We investigate the principles underlying these methods and use those insights to improve their capabilities.

  • Trustworthy, Adaptive, and Efficient Machine Learning

    Build learning methods that remain useful under uncertainty, changing conditions, and limited resources. We pursue robustness, interpretability, privacy, fairness, and rigorous assurance alongside continual learning, model compression, and computational efficiency. These questions connect theoretical guarantees with the requirements of deployed systems.

  • Machine Learning for Scientific Discovery and Societal Benefit

    Develop machine learning through close engagement with consequential problems. Our work spans biomedical imaging, brain and body dynamics, surgical guidance and robotics, physical simulation, transportation, emergency response, and other scientific and civic systems. These collaborations create opportunities both to advance individual fields and to discover broadly useful learning methods.

Leadership

Faculty

Affiliated Faculty

Want to learn more?

Contact us at soheil.kolouri@Vanderbilt.edu