Advancing the physics of learning and neural computation through cutting-edge research and innovative tools.
Understanding how neural networks learn through the lens of statistical physics and dynamical systems theory.
Investigating how capabilities emerge as models scale, using tools from random matrix theory and statistical physics.
Designing neural architectures based on principles from theoretical physics and neuroscience.
Open-source tools for analyzing neural networks as physical systems, including dynamics visualization and emergent property detection.
Infrastructure for bringing together researchers from physics, mathematics, computer science, and neuroscience.
