4 papers
Internal Guidance as a Sampling Method for Tilted Diffusion Models
Curtis McDonald, Peter Bartlett
Reward guidance methods sample an exponential tilt of a base density using a diffusion model trained with only samples from the base distribution. Existing methods use what we call…
Technical Note on Relating Scores of Tilted Distributions
Curtis McDonald
Recent results have shown that for a linear tilt to a reference measure, the scores that would be produced under convolution with a normal variable can be expressed in terms of con…
Rapid Bayesian Computation and Estimation for Neural Networks via Log-Concave Coupling
Curtis McDonald, Andrew R. Barron
This paper presents the study of a Bayesian estimation procedure for single-hidden-layer neural networks using controlled neuron weight vectors. We study the structure o…
Log-Concave Coupling for Sampling Neural Net Posteriors
Curtis McDonald, Andrew R Barron
In this work, we present a sampling algorithm for single hidden layer neural networks. This algorithm is built upon a recursive series of Bayesian posteriors using a method we call…