BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Energy landscapes: from molecules to machine learning - David Wale
 s (Dept. Chemistry\, University of Cambridge)
DTSTART:20200330T153000Z
DTEND:20200330T160000Z
UID:TALK140701@talks.cam.ac.uk
CONTACT:Bingqing Cheng
DESCRIPTION:Energy landscapes: from molecules to machine learning\n\nThe p
 otential energy landscape provides a conceptual and computational framewor
 k for \ninvestigating structure\, dynamics and thermodynamics in atomic an
 d molecular science.\nThis talk will highlight how new approaches for glob
 al optimisation\, enhanced sampling of systems\nexhibiting broken ergodici
 ty\, and rare event dynamics can provide new insight into the\nsolution la
 ndscape for neural networks. The key aim is to explain how\nthe energy lan
 dscape perspective can unify our understanding of apparently disparate\nsy
 stems. A range of applications will be presented including recent results 
 for machine learning landscapes.\n\nSelected Publications:\n Perspective: 
 New Insights From Loss Function Landscapes of Neural Networks. Machine Lea
 rning: Science and Technology\, in press\, 2020\n Machine learning landsca
 pes and predictions for patient outcomes. R Soc Open Sci 4\, 170175\, 2017
 .\n Perspective: Energy Landscapes for Machine Learning\, PCCP\, 19\, 1258
 5-12603\, 2017.\n Feature Article: Exploring Biomolecular Energy Landscape
 s\, Chem. Commun.\, 53\, 6974\, 2017\n Machine learning prediction for cla
 ssification of outcomes in local minimisation. Chemical Physics Letters 66
 7\, 158\, 2017\n Exploring Energy Landscapes. Ann. Rev. Phys. Chem.\, 69\,
  401-425\, 2017\n Energy Landscapes: Some New Horizons\, Curr. Op. Struct.
  Biol.\, 20\, 3\, 2010.\n Energy Landscapes\, Cambridge University Press\,
  Cambridge\, 2003
LOCATION:Mott Seminar (531) room\, top floor of the Mott Building\, in the
  Cavendish Laboratory\, West Cambridge.
END:VEVENT
END:VCALENDAR