Binned and Template fits in TensorFlow

Tuesday 25 Jun 2019, 12:00 → 14:00 Europe/Zurich

14/4-030 (CERN)

This meeting is about binned and template fits implemented in TensorFlow.

The goal is to discuss and combine the efforts in this topic. The focus is on the technical aspects:
 - using TensorFlow for this task (strength, missing features, ...)
 - defining the problem and summarizing what different experiments need for their analysis
 - ideas for: the optimal structure of such a library/interoperability/data layout/additional needed TF kernels etc.

It's also part of the ongoing effort to implement this functionalities in zfit.

1 Introduction and zfit

Speaker: Jonas Eschle (Universitaet Zuerich (CH))

template_and_binned_TF_zfit.pdf

2 TensorFlow fit in CMS CombinedLimits

Speaker: Josh Bendavid (CERN)

tffit-Jun4-2019.pdf

3 HAMMER and TensorFlow

Speaker: Julian Garcia Pardinas (Universitaet Zuerich (CH))

HAMMERAndTensorFlow_Julian.pdf

4 pyhf

Speaker: Lukas Alexander Heinrich (CERN)

pyhf_zfit.pdf

5 Discussion