Ritabrata 'Rito' Dutta
@ritothebayesian.bsky.social
Associate professor of statistics in the University of Warwick. Excited about everything Bayesian, related to Prediction, and smells of Computation!
Reposted by Ritabrata 'Rito' Dutta
The authors write: "If your pseudo-marginal MCMC (e.g. ABC) is stuck trying to reach to an approximate posterior, use adjusted SGLD (hugely efficient) with scoring rule posterior (guaranteed to contract to correct parameters)!"
November 29, 2024 at 12:00 PM
The authors write: "If your pseudo-marginal MCMC (e.g. ABC) is stuck trying to reach to an approximate posterior, use adjusted SGLD (hugely efficient) with scoring rule posterior (guaranteed to contract to correct parameters)!"