tag:blogger.com,1999:blog-10448119.post2678874300321724245..comments2024-03-23T06:42:53.608-04:00Comments on Hogg's Research: exoplanets, exoplanets, exoplanetsHogghttp://www.blogger.com/profile/18398397408280534592noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-10448119.post-62049951375586717502013-09-24T12:29:35.422-04:002013-09-24T12:29:35.422-04:00All I can say is (a) we are exploiting sparsity as...All I can say is (a) we are exploiting sparsity as much as we can, and (b) we are making approximations that let us do many more hypothesis tests than matrix determinant / inversion operations, and (c) we have a wavelet idea that might make things even faster (employing what some call the "matrix inversion lemma"). Right now I think we are still technically n^3 but with a very tiny prefactor.Hogghttps://www.blogger.com/profile/18398397408280534592noreply@blogger.comtag:blogger.com,1999:blog-10448119.post-2201290105014639792013-09-23T22:37:24.172-04:002013-09-23T22:37:24.172-04:00"People were shocked that some of our models ..."People were shocked that some of our models have hundreds of thousands of parameters."<br /><br />:-)<br /><br />"using Gaussian Processes under the hood. They told us it couldn't be done (too slow) but we are doing it."<br /><br />Do you know of any methods faster than O(N^3) for evaluating log det C and C^{-1}y?Brendon J. Brewerhttp://stat.auckland.ac.nz/~brewer/noreply@blogger.com