## Archive for Arianna Rosenbluth

## a film about Stan [not a film review]

Posted in Statistics with tags Adventures of a Mathematician, Arianna Rosenbluth, calculators, Edward Teller, film, Françoise Aron Ulam, Hiroshima, John von Neumann, Mici Teller, Nagasaki, nuclear weapons, Robert Oppenheimer, STAN, Stanislaw Ulam on December 17, 2021 by xi'an## Blackwell-Rosenbluth Awards 2021

Posted in Statistics, University life with tags Arianna Rosenbluth, awards, Bayesian conference, C.R. Rao, David Blackwell, fencing, ISBA, j-ISBA, MCMC, Metropolis-Hastings algorithm, Rao-Blackwell theorem, University of Warwick on November 1, 2021 by xi'an**C**ongratulations to the winners of the newly created award! This j-ISBA award is intended for junior researchers in different areas of Bayesian statistics. And named after David Blackwell and Arianna Rosenbluth. They will present their work at the newly created JB³ seminars on 10 and 12 November, both at 1pm UTC. (The awards are broken into two time zones, corresponding to the Americas and the rest of the World.)

##### UTC+0 to UTC+13

**Marta Catalano**, Warwick University

**Samuel Livingstone**, University College London

**Dootika Vats**, Indian Institute of Technology Kanpur

##### UTC-12 to UTC-1

**Trevor Campbell**, University of British Columbia

**Daniel Kowal**, Rice University

**Yixin Wang**, University of Michigan

## scale matters [maths as well]

Posted in pictures, R, Statistics with tags Arianna Rosenbluth, cross validated, Gaussian mixture, independent proposal, Metropolis-Hastings algorithm, R, simulation, Statistics Forum, Turing's chess on June 2, 2021 by xi'an**A** question from X validated on why an independent Metropolis sampler of a three component Normal mixture based on a single Normal proposal was failing to recover the said mixture…

When looking at the OP’s R code, I did not notice anything amiss at first glance (I was about to drive back from Annecy, hence did not look too closely) and reran the attached code with a larger variance in the proposal, which returned the above picture for the MCMC sample, close enough (?) to the target. Later, from home, I checked the code further and noticed that the Metropolis ratio was only using the ratio of the targets. Dividing by the ratio of the proposals made a significant (?) to the representation of the target.

More interestingly, the OP was fundamentally confused between independent and random-walk Rosenbluth algorithms, from using the wrong ratio to aiming at the wrong scale factor and average acceptance ratio, and furthermore challenged by the very notion of Hessian matrix, which is often suggested as a default scale.