Jet issues in Higgs physics

Europe/Zurich
No room. Remote only

No room. Remote only

Bruce Mellado Garcia (University of Wisconsin (US)), Daniele Del Re (Universita e INFN, Roma I (IT)), Frank Tackmann (DESY), Gavin Phillip Salam (CERN)
Description
Discussion on MVA techniques in H+2jets channel.
indico of the meeting:
https://indico.cern.ch/conferenceDisplay.py?confId=231636
 
CMS talk (M. Malberti):
------------------
 
- Content:
    a set of plots which show the effect of the MVA selection on dijet
    e.g. basic jet quantities + ptjj, deltaphi(jj,gg), pt(jjgg)
 
- Questions and comments:
   * what powheg? R: normal powheg, no minlo
   * please redo the plot for ggH. R: done and added into the new
     version of the talk
   * how do you prevent to not have deltaphi(gg,jj) values close to pi?
     In atlas we impose that MVA do not use deltaphi(gg,jj) variables
     above a given threshold. That make efficiency in that region ~flat.
     R: we don't. We will check the efficiency behavior.
 
 
1) ATLAS talk (D. Gillberg):
---------------------
 
- Content
    a set of plots which show the effect of the MVA selection on dijet
    e.g. basic jet quantities + ptjj, deltaphi(jj,gg), pt(jjgg)
 
- Question and comments:
   * why do you evaluate systematics at 95% efficiency point? R:
     arbitrary choice, only to get a cross-check what to expect from method
     shown in later in the other talk
   * would be useful to have efficiency plots for both CMS and ATLAS
   * imposing that MVA does not use deltaphi(gg,jj) above a given
     threshold reduces separation power. Maybe a trade-off 
     between loss in separation and theory uncertainty would 
     be beneficial. R: not for now but interesting for future studies
 
 
2) ATLAS talk (F. U. Bernlochner):
--------------------------
 
- Content
    development of a new technique to evaluate systematics due to
    radiation of extra jets in ggH for dijet channel when using MVAs. 
    Currently ST-like uncertainties used as input (technique itself can
    be used with any input uncertainties)
 
- Question and comments:
    * how do you deal with MCFM for regions with negative cross
      sections? R: Currently, no shape information is used below a certain
      cut but only normalization. (In principle we can extend the method in
      that region but once MCFM-only uncertainties are > 100%
      it becomes difficult to interpret the result.)
    * are there more checks that the method works, even if it uses
      categorization based on a single variable? page 11 looks like a
      trivial closure test. R: it is the test at page 9
    * why is the uncertainty in page 9 different between pure MCFM
      and powheg reweighted? R: spectra in the two MCs are different, we
      were actually happy to see this agreement
    * why not doing it on the MVA directly, i.e. by applying MVA to MCFM 
      ntuple output more than via deltaphi method?
      R: technically more involved, maybe using MVA directly on MCFM would
      be important to understand if it makes a difference if the uncertainty
      is 35% vs. 40% vs. 45% but given the present statistical
      uncertainty of this channel would not make such a difference
    * is it possible to get weights from ATLAS and apply them to CMS
      montecarlo? R: good idea to have consistency.
      Weights could be ready by the beginning of next week.
 
ACTION ITEMS:
------------
   * Experiments should discuss on how to proceed to have a common
     approach and in particular have numbers which can be directly compared.
   * The technique developed by Florian and Dag (ATLAS) and using inputs
     from F. Tackmann et al. paper can be used by CMS physicists
   * There is the Issue that ATLAS is imposing flat efficiency above a
     given threshold in deltaphi and the deltaphi bins are  based on
     that. CMS can check the efficiency above that value.
   * Maybe the starting point is to apply it to the Hgg channel with deltaphi
     as baseline variable
   * Once enough experience other channels can follow (using the
     appropriate baseline variable for them, e.g. pTHjj)
  * Code from ATLAS can be ready by beginning of next week and passed 
     to the LHC working group which will make it public (with
     documentation to follow)
 
There are minutes attached to this event. Show them.