Academic Celebration

Special Lectures and Homecoming
Keynote Speakers

The lecture will explore the history and recent results on two areas: (i) growth functions of groups, algebras, monoids, and languages; and (ii) complexity functions of infinite sequences.

When electrons are confined to two dimensions and subjected to a strong magnetic field, they enter an astonishing quantum world that defies our everyday intuition. Here, electrons capture quanta of magnetic field and transform into exotic particles known as composite fermions, giving rise to an unexpectedly simple organizing principle for a remarkably rich landscape of quantum phenomena, including the celebrated fractional quantum Hall effects . I will recount the series of surprises that led to this picture, the new forms of quantum matter it has revealed, and the surprises that continue to emerge to this day. 

Recent breakthroughs in generative AI rely on learning score functions—flow fields that generate samples from complex data distributions. Remarkably, these flows behave much like electric fields, revealing a compelling analogy between machine learning and electrodynamics. Guided by Maxwell’s equations, where electric charges act as the source of these fields, we propose parameterizing the score function as a superposition of electric fields generated by various charge geometries. Because these fields are conservative and curl-free by construction, our physically grounded approach inherently satisfies the geometric constraints required to model generative probability distributions. We demonstrate that this physics-inspired model is mathematically elegant, efficient to train, and matches the performance of state-of-the-art neural networks.