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:20210916T132450Z
LOCATION:Henry Dunant
DTSTART;TZID=Europe/Stockholm:20210706T143000
DTEND;TZID=Europe/Stockholm:20210706T150000
UID:submissions.pasc-conference.org_PASC21_sess136_msa267@linklings.com
SUMMARY:Distributed Deep Reinforcement Learning of Molecular Mechanisms
DESCRIPTION:Minisymposium\n\nDistributed Deep Reinforcement Learning of Mo
 lecular Mechanisms\n\nCovino, Jung, Wadhawan, Bolhuis, Hummer\n\nWe presen
 t a deep reinforcement learning artificial intelligence (AI) that learns t
 he mechanism of self-organization in molecular systems from computer simul
 ations. The AI initiates molecular dynamics simulations of molecular reorg
 anizations and progressively learns how to predict their outcome. We integ
 rate path theory, transition path sampling (TPS), deep learning, and symbo
 lic AI. TPS is a Markov chain Monte Carlo method to sample the rare trajec
 tories connecting metastable states. Using reinforcement learning, we iter
 atively train a deep neural network on the outcomes of TPS simulation atte
 mpts. In this way, we increase the rare-event sampling efficiency while gr
 adually revealing the underlying mechanism. The AI learns the rare events&
 rsquo; committor function, encoded in the trained neural network. By using
  symbolic regression, we distill simplified quantitative models that revea
 l mechanistic insight in a human-understandable form. Our innovative AI en
 ables the sampling and mechanistic interpretation of rare events by autono
 mously driving many parallel simulations with minimal human intervention.\
 n\nDomain: Chemistry and Materials, Physics, Life Sciences, Engineering
END:VEVENT
END:VCALENDAR