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:20191206T142453Z
LOCATION:Substanz 1\, 2
DTSTART;TZID=Europe/Stockholm:20180626T083000
DTEND;TZID=Europe/Stockholm:20180626T100000
UID:isc_hpc_ISC High Performance 2018_sess113_post115@linklings.com
SUMMARY:(RP06) A Deep Learning Tool for fast Simulation
DESCRIPTION:Research Poster\n\n(RP06) A Deep Learning Tool for fast Simula
 tion\n\nVallecorsa\n\nMachine Learning techniques have been used in differ
 ent applications by the HEP community: in this talk, we discuss the case o
 f detector simulation. The need for simulated events, expected in the futu
 re for LHC experiments and their High Luminosity upgrades, is increasing d
 ramatically and requires new fast simulation solutions. We will describe a
 n R&D activity, aimed at providing a configurable tool capable of training
  a neural network to reproduce the detector response and replace standard 
 Monte Carlo simulation. This represents a generic approach in the sense th
 at such a network could be designed and trained to simulate any kind of de
 tector and, eventually, the whole data processing chain in order to get, d
 irectly in one step, the final reconstructed quantities, in just a small f
 raction of time. We will present the first application of three-dimensiona
 l convolutional Generative Adversarial Networks to the simulation of high 
 granularity electromagnetic calorimeters. We will describe detailed valida
 tion studies comparing our results to standard Monte Carlo simulation, sho
 wing, in particular, the very good agreement we obtain for high level phys
 ics quantities and detailed calorimeter response. We will show the computi
 ng resources needed to train such networks and the implementation of a dis
 tributed adversarial training strategy (based on data parallelism). Finall
 y we will discuss how we plan to  generalize our model in order to simulat
 e a whole class of calorimeters, opening the way to a generic machine lear
 ning based fast simulation approach.\n\nPasses: Conference Pass, AI/Machin
 e Learning/Deep Learning, MPI, Parallel Applications, Scientific Software 
 Development\n\nTag: Conference Pass, AI/Machine Learning/Deep Learning, MP
 I, Parallel Applications, Scientific Software Development
URL:https://2018.isc-program.com/presentation/?id=post115&sess=sess113
END:VEVENT
END:VCALENDAR