Recognition: unknown
Fleming-Viot processes : two explicit examples
read the original abstract
The purpose of this paper is to extend the investigation of the Fleming-Viot process in discrete space started in a previous work to two specific examples. The first one corresponds to a random walk on the complete graph. Due to its geometry, we establish several explicit and optimal formulas for the Fleming-Viot process (invariant distribution, correlations, spectral gap). The second example corresponds to a Markov chain in a two state space. In this case, the study of the Fleming-Viot particle system is reduced to the study of birth and death process with quadratic rates.
This paper has not been read by Pith yet.
Forward citations
Cited by 2 Pith papers
-
VASR: Variance-Aware Systematic Resampling for Reward-Guided Diffusion
FVD applies Fleming-Viot population dynamics to diffusion model sampling at inference time to reduce diversity collapse while improving reward alignment and FID scores.
-
VASR: Variance-Aware Systematic Resampling for Reward-Guided Diffusion
VASR separates continuation and residual variance in reward-guided diffusion SMC, using optimal mass allocation and systematic resampling to achieve up to 26% better FID scores and faster runtimes than prior SMC and M...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.