Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
We introduce a novel underdamped Langevin system with momentum preconditioning for sampling from Gibbs distributions on non-convex, non-globally gradient-Lipschitz matrix spaces. This is achieved by employing non-quadratic kinetic energies, where the gradient of the kinetic energy spectrally tames the momentum. The proposed dynamics are proven to preserve the target Gibbs measure and exhibit exponential convergence to equilibrium in weighted total variation distance, with its Euler-Maruyama discretization showing uniform-in-time moment bounds.
Non-convex sampling from matrix spaces is now possible with provable exponential convergence using momentum preconditioning that spectrally tames gradients.
We consider the problem of sampling from Gibbs distributions on matrix spaces whose potential energies are neither convex nor globally gradient-Lipschitz. We introduce a family of non-quadratic kinetic energies that lead to a new underdamped Langevin system with momentum preconditioning, in which the gradient of the kinetic energy acts as a smooth spectral taming of the momentum. We prove that, under these relaxed assumptions on the potential, the resulting dynamics leaves the target Gibbs measure invariant, and we establish exponential convergence to equilibrium in a weighted total variation distance. Finally, we show that the corresponding Euler-Maruyama discretization admits moment bounds that are uniform in time, without any modification of the potential gradient, which ensures the stability of the resulting sampling algorithm.