2024/10/03 by Chen, August Y., Sridharan, Karthik · 1 citation
#FOS: Computer and information sciences #FOS: Mathematics #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Probability (math.PR) #Statistics Theory (math.ST)
paper · doi:10.48550/arxiv.2410.02979
In this paper, we prove that optimizability of any function F using Gradient Flow from all initializations implies a Poincaré Inequality for Gibbs measures mubeta = e-beta F/Z at low temperature. In particular, under mild regularity assumptions on the convergence rate of Gradient Flow, we establish that mubeta satisfies a Poincaré Inequality with constant O(C'+1/beta) for beta >= Omega(d), where C' is the Poincaré constant of mubeta restricted to a neighborhood of the global minimizers of F. Under an additional mild condition on F, we show that mubeta satisfies a Log-Sobolev Inequality with constant O(beta max(S, 1) max(C', 1)) where S denotes the second moment of mubeta. Here asymptotic notation hides F-dependent parameters. At a high level, this establishes that optimizability via Gradient Flow from every initialization implies a Poincaré and Log-Sobolev Inequality for the low-temperature Gibbs measure, which in turn imply sampling from all initializations. Analogously, we establish that under the same assumptions, if F can be initialized from everywhere except some set S, then mubeta satisfies a Weak Poincaré Inequality with parameters (O(C'+1/beta), O(mubeta(S))) for β= Omega(d). At a high level, this shows while optimizability from 'most' initializations implies a Weak Poincaré Inequality, which in turn implies sampling from suitable warm starts. Our regularity assumptions are mild and as a consequence, we show we can efficiently sample from several new natural and interesting classes of non-log-concave densities, an important setting with relatively few examples. As another corollary, we obtain efficient discrete-time sampling results for log-concave measures satisfying milder regularity conditions than smoothness, similar to Lehec (2023).