Adaptive multiple importance sampling resubmited
Our AMIS paper (with Jean-Marie Cornuet, Jean-Michel Marin and Antonietta Mira) has been once again revised and resubmitted to Scandinavian Journal of Statistics and arXived as well. Since this is the second round of revisions, the difference is not major with the earlier version… The reason why we needed this revision is that a new referee made requests about the theoretical validation of the AMIS algorithm, sending us back to the earliest version of the paper where we had described our endeavours and lack of success to completely circumvent the convergence problem. We agree it would have been nice to end up with a convergence result for large T, small N asymptotics, but we talked to many (brilliant) colleagues over the years without finding any useful hint. The paper thus concludes with the relevance of this perspective, nothing more… (We have now reached the point where we think that the stylised case Nt=1 is not really useful.) A revision a minima thus, hopefully sufficient to conclude the rounds of revisions for a paper on an algorithm that is already implemented in the literature!
May 23, 2011 at 12:13 am
[…] for publication Scandinavian Journal of Statistics, which is a very good news. As mentioned in the earlier post, two weeks ago, I am very glad the paper is finally accepted as (a) the method is already in use in […]