Show simple item record

dc.contributor.authorHong, Yuxi
dc.contributor.authorBergou, Houcine
dc.contributor.authorDoucet, Nicolas
dc.contributor.authorZhang, Hao
dc.contributor.authorCranney, Jesse
dc.contributor.authorLtaief, Hatem
dc.contributor.authorGratadour, Damien
dc.contributor.authorRigaut, Francois
dc.contributor.authorKeyes, David E.
dc.date.accessioned2020-11-22T14:05:42Z
dc.date.available2020-11-22T14:05:42Z
dc.date.issued2020-10-19
dc.identifier.urihttp://hdl.handle.net/10754/666072
dc.description.abstractWe present a new Stochastic Levenberg-Marquardt (SLM) algorithm for efficiently solving large-scale nonlinear least-squares optimization problems. The SLM method incorporates stochasticity into the traditional Levenberg-Marquardt (LM) method. While the traditional LM operates on the full objective function, SLM randomly evaluates part of the objective to compute the corresponding derivatives and function values. Hence, SLM reduces the algorithmic complexity per iteration and speeds up the overall time to solution, while maintaining the numerical robustness of second-order methods. We assess the SLM method on standard datasets from LIBSVM, as well as on a large-scale optimization problem found in ground-based astronomy applications, and in particular adaptive optics systems of the next generation of instruments for the European Very Large and Extremely Large Telescopes. We implement SLM and deploy it on a shared-memory system equipped with multiple GPU hardware accelerators. We demonstrate the performance superiority of the SLM method over not only the traditional LM algorithm but also the state-of-the-art first-order methods. SLM finishes optimization process in less than 1 second on large datasets from the adaptive optics application, where LM and other methods require more than a few minutes. This enables to identify of the system parameters (e.g., atmospheric turbulence and wind speed) and to capture their evolution required during a night of observations with a close to real-time throughput.
dc.description.sponsorshipWe thank Nvidia and Paris Observatory for computing resources.
dc.publisherSubmitted to IEEE
dc.rightsPreprint submitted to 35th IEEE International Parallel & Distributed Processing Symposium. Archived with thanks to IEEE.
dc.titleStochastic Levenberg-Marquardt for Solving Optimization Problems on Hardware Accelerators
dc.typePreprint
dc.contributor.departmentComputer Science
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
dc.contributor.departmentExtreme Computing Research Center
dc.contributor.departmentApplied Mathematics and Computational Science Program
dc.contributor.departmentOffice of the President
dc.conference.dateMay 17-21, 2021
dc.conference.nameSubmitted to 35th IEEE International Parallel & Distributed Processing Symposium
dc.conference.locationPortland Hilton Downtown Portland, Oregon USA
dc.eprint.versionPre-print
dc.contributor.institutionResearch School of Astronomy & Astrophysics, College of Science, Australian National University, Australia
pubs.publication-statusSubmitted
kaust.personHong, Yuxi
kaust.personBergou, Houcine
kaust.personLtaief, Hatem
kaust.personKeyes, David E.
refterms.dateFOA2020-11-22T14:05:42Z


Files in this item

Thumbnail
Name:
main.pdf
Size:
825.8Kb
Format:
PDF
Description:
Preprint

This item appears in the following Collection(s)

Show simple item record