1970-. American computer scientist and mathematician, professor at Yale.
Developed spectral graph theory as a practical computational tool, connecting eigenvalues of the Laplacian to graph partitioning, clustering, and linear system solving.
Co-invented nearly-linear time solvers for Laplacian systems ($Lx = b$ in $O(m \log^c n)$ time), making spectral methods scalable to massive graphs.
His spectral sparsification technique approximates dense graphs with sparse ones that preserve spectral properties, enabling efficient computation on large networks.
Proved the Kadison-Singer conjecture (with Marcus and Srivastava), resolving a 50-year-old problem in functional analysis with direct implications for graph algorithms.
His lecture notes on spectral graph theory are the standard reference for the mathematical machinery behind the cyber tri-kernel and Laplacian-based computation.