About

The Computational Intelligence, Control & Information (CI²) Lab at Texas Tech University studies how intelligent agents can learn to perceive, predict, and control complex dynamical systems. Our work sits at the intersection of deep learning, reinforcement learning, optimal control, and information theory.

We pursue fundamental questions: How should an agent represent its environment to act effectively? What are the information-theoretic limits of perception and control? How can learned models be made reliable enough to drive real physical systems? We approach these with a blend of theory and large-scale experimentation, with applications spanning robotics, autonomous systems, and scientific machine learning.

Research directions

• Deep Learning for Control — learning predictive models and policies to control complex dynamical systems, with an emphasis on sample efficiency, stability, and transfer to real hardware.
• Information Theory of Perception & Control — information-theoretic principles governing how an agent should perceive its environment and choose actions, including intrinsic motivation and empowerment.
• Representation Learning for Dynamical Systems — structured, predictive representations of high-dimensional, partially observed worlds that support planning and control.

See the research page for more detail and the people page for current members.

Join us

The lab is part of the Department of Computer Science in the Edward E. Whitacre Jr. College of Engineering at Texas Tech University. We are actively looking for motivated PhD and undergraduate researchers interested in machine learning for control — reach out to Dr. Stas Tiomkin with a short note about your background and interests.