Subadditivity of the log-Sobolev constant on convolutions
By: Thomas A. Courtade, Edric Wang
Potential Business Impact:
Makes math rules work for more complex problems.
We present a general subadditivity inequality for log-Sobolev constants of convolution measures. As a corollary, we show that the log-Sobolev constant is monotone along the sequence of standardized convolutions in the central limit theorem.
Similar Papers
Talagrand's convolution conjecture up to loglog via perturbed reverse heat
Probability
Makes guessing better for computers.
Failure of uniform laws of large numbers for subdifferentials and beyond
Optimization and Control
Math rules break for bumpy shapes.
Contractive kinetic Langevin samplers beyond global Lipschitz continuity
Probability
Makes computer models learn faster and more accurately.