BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Direct Learning of Systematics-Aware Summary Statistics
DTSTART;VALUE=DATE-TIME:20180803T171100Z
DTEND;VALUE=DATE-TIME:20180803T171200Z
DTSTAMP;VALUE=DATE-TIME:20191113T130744Z
UID:indico-contribution-3051475@indico.cern.ch
DESCRIPTION:Speakers: Pablo De Castro Manzano (Universita e INFN\, Padova
(IT))\nComplex machine learning tools\, such as deep neural networks and g
radient boosting algorithms\, are increasingly being used to construct pow
erful discriminative features for High Energy Physics analyses. These meth
ods are typically trained with simulated or auxiliary data samples by opti
mising some classification or regression surrogate objective. The learned
feature representations are then used to build a sample-based statistical
model to perform inference (e.g. interval estimation or hypothesis testing
) over a set of parameters of interest. However\, the effectiveness of the
mentioned approach can be reduced by the presence of known uncertainties
that cause differences between training and experimental data\, included i
n the statistical model via nuisance parameters. This work presents an end
-to-end algorithm\, which leverages on existing deep learning technologies
but directly aims to produce inference-optimal sample-summary statistics.
By including the statistical model and a differentiable approximation of
the effect of nuisance parameters in the computational graph\, loss functi
ons derived form the observed Fisher information are directly optimised by
stochastic gradient descent. This new technique leads to summary statisti
cs that are aware of the known uncertainties and maximise the information
that can be inferred about the parameters of interest object of a experime
ntal measurement.\n\nhttps://indico.cern.ch/event/648004/contributions/305
1475/
LOCATION:Arts building
URL:https://indico.cern.ch/event/648004/contributions/3051475/
END:VEVENT
END:VCALENDAR