Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
June 11, 2026Open Access

Interacting Particle Systems for Sampling from Non-Gaussian Targets

View Full Paper
Ask AI
Bookmark
Share

Authors

RTRitvik Sai Teegavarapu

Discussion

Loading...

Member takes

Overview

Randomized trial evaluates a new particle algorithm for non-Gaussian target distributions, suggesting improved performance.

Key Points

  • This research aims to enhance sampling methods from non-Gaussian target distributions in applied mathematics.
  • Developed a regularized variant of the Ensemble Kalman Sampler (EKS) mean-field PDE.
  • Replaced potential and entropy terms with mollified counterparts parameterized by a smoothing scale.
  • Proved existence of weak solutions via a covariance-modulated JKO variational scheme.
  • Numerical experiments show that the new particle algorithm matches or outperforms EKS for various benchmark targets.
  • Characterized the steady states of the covariance-modulated blob flow and their associated bias.
  • Established convergence of mollified solutions to the EKS mean-field PDE as the smoothing scale decreases.

Cite This Study

Ritvik Sai Teegavarapu (2026) studied this question.

synapsesocial.com/papers/6a2a51d780c8f91e7f39df3chttps://doi.org/10.7907/jkfa-cc97
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Sequential Kalman Monte Carlo for gradient-free inference in Bayesian inverse problems2024
  2. 2Importance sampling for rare event tracking within the ensemble Kalman filtering framework2024
  3. 3Efficient, Multimodal, and Derivative-Free Bayesian Inference With Fisher-Rao Gradient Flows2024 · 1 citations
  4. 4Weak Generative Sampler for Stationary Distributions of McKean-Vlasov System2025
  5. 5A Method of Moments Estimator for Interacting Particle Systems and their Mean Field Limit2024 · 7 citations