INITIAL DATA
LATENT
REPRESENTATION
AUTO ENCODER
QSVM
Simulated
G → WW Events
Two jets with highest p
T
Each with 100 particles
for each event
Δη, Δϕ, p
T
for each particle
Purpose: Dimensionality Reduction
Transforms data into new space, called latent representation,
with different dimensionality
(here from
300
into 4, 8 or 16).
Trained using SM event data, the model will transform
BSM event data in a noticeably different way.
SM
BSM
Standard Model
Beyond the Standard Model
QCD multijet production
narrow Randall-Sundrum gravitons
decaying to two W-bosons
Quantum Support Vector Machine
Classic SVM
Kernel Trick
1
overview
classic svm
optimization
prediction
kernel
mapping
mapping
ارتقاء امنیت وب با وف بومی