BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20211207T055407Z
LOCATION:Online
DTSTART;TZID=America/Chicago:20211115T110000
DTEND;TZID=America/Chicago:20211115T113000
UID:submissions.supercomputing.org_SC21_sess423_ws_mlhpce113@linklings.com
SUMMARY:HYPPO: A Surrogate-Based Multi-Level Parallelism Tool for Hyperpar
 ameter Optimization
DESCRIPTION:Workshop\n\nHYPPO: A Surrogate-Based Multi-Level Parallelism T
 ool for Hyperparameter Optimization\n\nDumont\n\nWe present a new software
 , HYPPO, that enables the automatic tuning of hyperparameters of various d
 eep learning models. Unlike other hyperparameter optimization methods, HYP
 PO uses adaptive surrogate models and directly accounts for uncertainty in
  model predictions to find accurate and reliable models that make robust p
 redictions. Using asynchronous nested parallelism, we are able to signific
 antly alleviate the computational burden of training complex architectures
  and quantifying the uncertainty. HYPPO is implemented in Python and can b
 e used with both TensorFlow and PyTorch libraries. We demonstrate various 
 software features on time-series prediction and image classification probl
 ems as well as a scientific application in computed tomography image recon
 struction. Finally, we show that we can reduce by an order of magnitude th
 e number of evaluations necessary to find the most optimal region in the h
 yperparameter space and reduce by two orders of magnitude the throughput f
 or such HPO process to complete.\n\nTag: Online Only, Machine Learning and
  Artificial Intelligence\n\nRegistration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR