Holley-Stroock perturbation: if a probability measure satisfies a log-Sobolev inequality with constant for bounded cylindrical functions and noise gradients, then the normalized measure with density proportional to , for a bounded Borel with values in , satisfies it with constant .
In the setting of Gaussian Analysis Relative to a Diagonal Gaussian Reference Measure with Noise Weights: Standing Notation, with and the noise gradients of Gaussian Analysis Relative to a Diagonal Gaussian Reference Measure with Noise Weights: Standing Notation §cylindrical, for which is a bounded nonnegative Borel function by The Noise Gradient of a Cylindrical Function and the Noise Tangent Space at a Probability Measure on a Hilbert Space §gradient; for , is the entropy with respect to , and is the function of The Function : Continuity, Young's Inequality and Lower Bounds, with the Elementary Bounds for the Exponential and the Logarithm. is the exponential function. For every and every , the functions and are bounded and Borel, hence integrable with respect to , so that is defined: is bounded and Borel by Bounded C^1 Cylindrical Functions: Linear Structure, the Gradient and the Partial Derivatives, Bounds and Integrability, and Density in the Square-Integrable Functions §bounded-borel, so is Borel by claim 3 of Arithmetic, Absolute Values, and Pointwise Limits of Measurable Real-Valued Functions and its values lie in an interval ; is continuous by The Function : Continuity, Young's Inequality and Lower Bounds, with the Elementary Bounds for the Exponential and the Logarithm §continuous, so is Borel by that clause and bounded on by Extreme Value Theorem on a Closed Real Interval; and bounded Borel functions are integrable with respect to a probability measure by claim 6 of Borel Measurability and Bounded Integration on a Metric Space.
Let and let with be such that
Let be Borel and bounded, and let satisfy for every .
1. (The perturbed measure) The function is Borel and bounded, hence integrable with respect to ; the number satisfies , so ; the function on given by belongs to ; and for every bounded Borel , .
2. (Perturbed log-Sobolev inequality) For every ,
Loading…
No relations recorded yet.