BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20241120T082410Z
LOCATION:HG F 30 Audi Max
DTSTART;TZID=Europe/Stockholm:20240604T095800
DTEND;TZID=Europe/Stockholm:20240604T095900
UID:submissions.pasc-conference.org_PASC24_sess158_pos141@linklings.com
SUMMARY:P46 - Scaling Laws for Machine-Learned Reconstruction
DESCRIPTION:Poster\n\nEric Wulff (CERN), Joosep Pata (National Institute o
 f Chemical Physics and Biophysics), and Maria Girone (CERN)\n\nMachine Lea
 rning (ML) methods have been successfully applied to various High Energy P
 hysics (HEP) problems, such as particle identification, event reconstructi
 on, jet tagging, and anomaly detection. However, the relationship between 
 the model size, i.e., the number of model parameters, and the physics perf
 ormance for different HEP tasks is not well understood. In this work, we e
 mpirically determine the scaling laws for different commonly used ML model
  architectures such as Graph Neural Networks (GNNs) and Transformers on a 
 challenging ML problem from HEP with the goal of finding how much physics 
 performance can be gained by increasing the model size as opposed to inves
 tigating more complex model architectures. We also take memory usage and c
 omputational complexity, which is not directly related to model size, into
  account. High Performance Computing resources are used to train and optim
 ize the models on large-scale HEP datasets for supervised learning. We eva
 luate the model performance in terms of accuracy, efficiency, and inferenc
 e speed. We also observe that the optimal model size varies depending on t
 he complexity and structure of the input data. Our work demonstrates the p
 otential and challenges of applying ML methods to HEP problems, and contri
 butes to the advancement of both fields.\n\nSession Chair: Iva Kavcic (Met
  Office)
END:VEVENT
END:VCALENDAR