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:20220812T074334Z
LOCATION:Nairobi Room
DTSTART;TZID=Europe/Stockholm:20220627T163000
DTEND;TZID=Europe/Stockholm:20220627T170000
UID:submissions.pasc-conference.org_PASC22_sess131_msa169@linklings.com
SUMMARY:How Much Chemistry do Machine-Learning Models Learn?
DESCRIPTION:Minisymposium\n\nHow Much Chemistry do Machine-Learning Models
  Learn?\n\nRiniker\n\nFrom simple clustering techniques to sophisticated n
 eural networks, the use of machine learning (ML) has become a valuable too
 l in many fields of chemistry in the past decades. While domain applicabil
 ity is a common concept when assessing ML models in cheminformatics (e.g.,
  predicting biological activity in virtual screening), ML models learning 
 about the potential energy landscape of molecules from quantum-mechanical 
 (QM) data are often assumed to have learned the underlying physics. Here, 
 we attempt to explore – based on our own work and that of others – to whic
 h degree ML models learn chemical concepts when trained on QM data and dis
 cuss the challenges in the field. We are particularly interested in conden
 sed phase systems and show how the ∆-learning scheme can help to simplify 
 the learning task for (QM)ML/MM MD simulations.\n\nDomain: Chemistry and M
 aterials, Life Sciences, Physics
END:VEVENT
END:VCALENDAR