Our next meeting will be on 11 July 2024 at 17:00 Geneva (CH) Time.
We are happy to host Vaibhav Thakkar.
Title: Automatic Differentiation in Roofit for Fast and Accurate Likelihood Fits
Abstract: RooFit is a widely used data analysis library in High-energy physics experiments at CERN, which requires minimizing likelihoods with thousands of free parameters spread over hundreds of likelihood components. To iteratively find the minimum of a likelihood, one generally has to know the gradient of the likelihood with respect to all free parameters. Previously, these were calculated using numerical differentiation, which requires varying one parameter at a time and reevaluating the full likelihood. Although complete reevaluation can be prevented by a caching mechanism, it incurs a significant overhead in bookkeeping. This talk discusses the recent integration of automatic differentiation (AD) in RooFit for these gradient computations using Clad, a compile-time AD library for C++ codebases. We also present the efficiency improvements with the recently released Atlas and CMS Higgs models.
Speaker: Vaibhav Thakkar is currently a
research Intern at CERN, working with
the compiler research group at Princeton University (US) on the
project of compile-time automatic differentiation (Clad) in C++
for
scientific computing applications. Previously, he earned my
bachelor's
in technology in Electrical Engineering from IIT Kanpur in India
and
worked as a software engineer for a couple of years after
graduation.
Feel free to forward this email to everybody who might be
interested. This talk will be recorded.