Entropic Trace Estimates for Log Determinants
From MaRDI portal
Publication:6285883
arXiv1704.07223MaRDI QIDQ6285883FDOQ6285883
Authors: Jack Fitzsimons, Diego Granziol, Kurt Cutajar, Michael R. Osborne, M. Filippone, Stephen J. Roberts 
Publication date: 24 April 2017
Abstract: The scalable calculation of matrix determinants has been a bottleneck to the widespread application of many machine learning methods such as determinantal point processes, Gaussian processes, generalised Markov random fields, graph models and many others. In this work, we estimate log determinants under the framework of maximum entropy, given information in the form of moment constraints from stochastic trace estimation. The estimates demonstrate a significant improvement on state-of-the-art alternative methods, as shown on a wide variety of UFL sparse matrices. By taking the example of a general Markov random field, we also demonstrate how this approach can significantly accelerate inference in large-scale learning methods involving the log determinant.
Has companion code repository: https://github.com/11hifish/OptSketchTraceEst
This page was built for publication: Entropic Trace Estimates for Log Determinants
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6285883)
