BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Integrating machine learning and quantum chemistry with fully diff
 erentiable quantum chemistry - Muhammad Firmansyah Kasim\, University of O
 xford
DTSTART:20210531T153000Z
DTEND:20210531T160000Z
UID:TALK160783@talks.cam.ac.uk
CONTACT:Bingqing Cheng
DESCRIPTION:Automatic differentiation programming offers paradigm shifts t
 o enable researchers focusing only on forward calculation without having t
 o derive the gradients. By taking down the barrier in deriving gradients\,
  automatic differentiation has fueled the growth of deep learning and coul
 d have the same potential in other scientific fields\, such as quantum che
 mistry. As computational quantum chemistry involves abundant gradient calc
 ulations\, writing the code with automatic differentiation allows explorat
 ion of new ideas that were prohibited by the difficulty of calculating gra
 dients. Here we introduce an open-source differentiable quantum chemistry\
 , DQC\, and explore some applications made easy by having the automatic di
 fferentiation capability. One of the applications is learning the exchange
 -correlation (xc) functional in DFT. Using only eight experimental data po
 ints on diatomic molecules\, our trained xc networks enable improved predi
 ction accuracy of atomization energies across a collection of 104 molecule
 s containing new bonds and atoms that are not present in the training data
 set. Besides learning xc functional\, DQC also has other applications such
  as learning a new basis set for a family of molecules\, checking converge
 nce stability\, and predicting molecular properties with alchemical pertur
 bation among others. To conclude the talk\, I will briefly discuss the cha
 llenges in implementing software using modern automatic differentiation li
 braries\, specifically PyTorch. The code is available at: https://github.c
 om/mfkasim1/dqc/\n\n
LOCATION:virtual ZOOM meeting ID: 263 591 6003\, Passcode: 000042\, https:
 //us02web.zoom.us/j/2635916003?pwd=ZlBEQnRENGwxNmJGMENGMWxjak5nUT09
END:VEVENT
END:VCALENDAR