BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning Feynman Diagrams using Graph Neural Networks - Harrison M
 itchell\, dept of Physics Cambridge
DTSTART:20221124T170000Z
DTEND:20221124T180000Z
UID:TALK192824@talks.cam.ac.uk
CONTACT:Pietro Lio
DESCRIPTION:In the wake of machine learning growing in popularity in the f
 ield of particle physics\, this work finds a new application of geometric 
 deep learning on Feynman\n diagrams to make accurate and fast matrix eleme
 nt predictions with the potential to be used in analysis of quantum field 
 theory. This research uses the graph attention layer which makes matrix el
 ement predictions to 1 significant figure above 90% of the time. Peak perf
 ormance was achieved in making predictions to 3 significant figures over 1
 0% of the time with less than 200 epochs of training. Finally\, a procedur
 e is suggested\, to use the network to make advancement in quantum field t
 heory by constructing Feynman diagrams with effective particles that repre
 sent non-perturbative calculations.\n\n\nzoom:\nhttps://zoom.us/j/99166955
 895?pwd=SzI0M3pMVEkvNmw3Q0dqNDVRalZvdz09
LOCATION:lecture theatre 2 dept of computer science and zoom
END:VEVENT
END:VCALENDAR