Artificial Intelligence and Machine Learning

Overview

Artificial Intelligence at Caltech researches the development of intelligent systems capable of emulating human-like intelligence and performing complex tasks. We study the fundamental aspects of AI, spanning machine learning, mathematics, and statistics, in order to delve into applications in perception, robotics, reinforcement learning, and decision making. We are witnessing a new era in AI with the emergence of foundation models, which hold great potential for transforming and improving the quality of human life. At Caltech, our mission is to propel scientific discoveries forward through collaborative efforts across disciplines, including robotics, biology, chemistry, geophysics, ecology, economics, and more. These endeavors are spearheaded by Caltech's AI faculty: Anima Anandkumar, Katie Bouman, Georgia Gkioxari, Eric Mazumdar, Pietro Perona, Yang Song, Adam Wierman and Yisong Yue.

Anima Anandkumar works on foundations of machine learning and deep learning methods, and its applications to scientific domains. Anima's work on neural operators accelerate scientific simulations and design in applications such as weather modeling, climate mitigation and medical devices. She also works on learning for control, tensor methods, optimization, and generalization of AI foundation models.

Katie Bouman focuses on developing computational cameras and solving inverse problems for a variety of applications, including medical imaging, recovering material properties, seismic tomography and imaging dark matter. Katie's work on integrating algorithms and sensors to image the invisible has led to breakthroughs in imaging techniques, such as the imaging of black holes.

Georgia Gkioxari works on machine perception, namely teaching machines to see. Georgia's interests lie in the design of visual perception systems that bridge the gap between 2D imagery and our 4D world, including visual recognition, 3D reconstruction, video understanding and novel learning techniques for learning with less supervision.

Eric Mazumdar works at the intersection of machine learning and economics by exploring learning techniques in the presence of strategic agents, multi-agent reinforcement learning, and online learning. Eric studies the consequences of algorithmic decision-making in the context of bias, fairness and polarization.

Pietro Perona studies the computational foundations of vision with an emphasis in behavior analysis and learning in laboratory animals, AI for conservation and ethical AI. Pietro collaborates with Caltech psychologists, social scientists and neuroscientists to understand the nature of behavior, search and decision in humans and animals.

Yang Song focuses on developing scalable methods for modeling, analyzing and generating complex, high-dimensional data. Yang's interests span multiple areas, including generative modeling, representation learning, probabilistic inference, AI safety, and AI for science.

Adam Wierman works on online learning and online optimization, with applications to the sustainability of networked systems.

Yisong Yue works broadly in two areas. The first is learning neurosymbolic models, which aim to be symbolically interpretable and amenable to formal verification. The second is autonomous decision making. His research agenda spans both fundamental and applied pursuits, from novel learning-theoretic frameworks all the way to deployment in autonomous driving on public roads.