Our status and plans in CMS

1
Our status and plans in CMS
Group members in CMS Prof. W. deBoer group
leader V. Zhukov M. Niegel

This talk is aimed to increase collaboration
between groups
Main research subject CMS SUSY BSM group
Institute CMS duties (M.Niegel) CMS grid jobs
monitoring, CMSSW support
History 01.2005 join CMS SUSY BSM group.
Responsibility mSUGRA trileptons signature in
large m0 region, inspired by indirect DM search
and Tevatron search 05.2005 setup analysis
framework in ORCA and FAMOS, implemented FAMOS
PU 06.2005 CMSIN2005/39 Trileptons from SUSY
production. First results. 10.2005 Implemented
NN and Genetic algo for the bkg.
suppression. Generated gt107 events in full and
fast simulation using Grid and ekpplus. 02.2006
study influence of fake leptons on trilepton
signal. 05.2006 CMS PTDR2 contribution and CMS
NOTE2006/113. Finished trileptons
What is the next?
2
1. Fake leptons
Why fake leptons?
MSUGRA trileptons
1. Important for SUSY search, especially in low
MET regions. 2. For all leptonic channels
(SM,Higgs) Fakes -increase bkg. -introduce
uncertainties Tevatron experience more fakes
than expected...
With fakes 50 increase bkg
Without fakes
Our experience standard lepton isolation cuts
tracker SPTlt1.5 GeV in cone Rlt0.3 , for muons
Ecallt5 GeV Rlt0.3
Rate per event (FAMOS)
CMS IN2005/28 Wjets ORCA Fe10 4 Fm 105
Large uncertainties in Fake rates, up to 50
Fake rate depends on the bkg. channel, event MC
generator (PS,ME, I/FSR) and detector
simulation (FAMOS,ORCA)
No clear understanding in CMS how to estimate
fake rates in simulation and with the real
data....
3
Fake leptons classification
in simulation -trace back all leptons and
calculate fake rates per each type of object
(jet, hadrons, leptons) -improve selection
algorithms (use NN), spot important detector
parameters.
with real data calibrate with standard candles (
Zjets,Wjets...) , but still PU, ISRF effects.
We need an algorithm for continuous monitoring!
Plans This is a 'long' term activity related to
different CMS working groups. 1. FakeFinder
package, first for FAMOS, later for CMSSW
10/2006 2. Optimization of leptons Id (in
collaboration with ...)
12/2006 3. Calibration with the real data (
algorithms)
5/2007
4
Fakes Finder
For example (preliminary)
FakeElectron
Fake
FakesFinder
Fake getfake(RecMuon ) Fake
getfake(ElectronCandidate ) ....
FakeMuon
bool fake double fakeness ...
FakeJet
FakesConfig
SimParticle SimHit CaloClusters RecMuon ...
void read() double dRmatch .....
Tracer
Containers of related pointers for easy navigation
HepParticle getMC() SimHit getSim() ...
How to match the MC, Sim and Rec objects? -use
existing reconstruction history -independently by
the isolation cone .....
As a result optimization of the leptons
identification
5
2. SUSY topology selector
SUSY (mSUGRA) free parameters (mo, m1/2, A0, tanb
, sgnm ) define mass spectrum and coupling of
new particles (q,g,c) and therefore signal
signature
METJetsLeptons
With different mSUGRA parameters
6
Inclusive and exclusive SUSY search
1. Inclusive search Check any deviations from
SM predictions already at low statistics (small
s, low Lint ) First days Physics! Counting like
experiments. Uncertainties are very important.
Can we constrain SUSY parameters already in
inclusive search?
2. Exclusive search With enough statistics
reconstruct kinematic variables , endpoints in
invariant mass distribution. Constrains mass
spectrum and SUSY parameters. Works in limited
regions , at low m0. Requires large statistics.
ATLAS
Reconstructed mass spectrum from kinematic end
points at 100 fb-1 for low mass region
Region accessible for kinematic end points
7
mSUGRA cross sections
Main production channels
Fractions of different production channels
Total mSUGRA cross section
s tot pb
tan?50 A00
Different channels -gt different topologies can be
used to separate regions
8
MSUGRA observables Jets
Reconstructed (FAMOS) averaged observables
ltNjgt ETgt30GeV
ltNBjgt B jets
mgmq long cascades
c02-gt H -gtbbar
Highest jet ltETgt
Second jet ltETgt
200
250
Hardest jet in q decays
Third jet ltETgt
Cosf between first and second
120
Cosf between first and MET
Cosf between second and MET
9
more observables MET
S ET from CaloTowers
MET from CaloTowers
2 body
Low MET region
MinvminvH1minvH2/2 invariant mass per hemisphere
Heff METS ET jets S ETleptons
10
And more Leptons
ltN gt electrons
ltN gt muons
ltN gt SameSignSameFlavor
ltN gt OppositeSignSameFlavor
Highest muon ltPTgt m1/2
Cosf OSSF and MET
back2back
Nss/Nss-
ltMinvgt OSSF
Expect asymmetry
Zpeak
11
Event selection in inclusive search
Inclusive analysis in CMS
Bkg hypothesis Hb
Signal hypothesis Hs
signal
Selector
ns
S
sb
nb
backgrounds
a
/- uncert.
S
nb
nbns
Selector is optimised for some particular test
point to get maximum significance S. For small
statistic regime at S5 (discovery reach) ns5,
nb1 shape of Hb distribution is very important!
(and not well known)
a -probability to accept wrongly the Hs
Significance Sns/sb , where sb
sqrt(nbdnb2) nb -Poisson term dnb - systematic
uncertainties

• Only one parameter (S) to characterize a
  hypothesis
• Systematic uncertainties of the background are
  very important
• Can not say much about the model parameters

12
Region Dependant event selection
-The mSUGRA topologies depends on parameters ,
can define n- region dependant selectors Seli
-There is only one solution , i.e. Selk the
Sel n-m should give no signal at the defined
threshold (5s) . Probability to observe a signal
in k -region Ppk Si?k (1-pi)
Multichannel inclusive analysis
signal
Selector i optimized for i-topology
Si
ns
ns
S
Si
Selector i optimized for i-topology
ns
nb
backgrounds
Preselector
nb
/- uncert.
Preselector suppresses most of the SM
background Selectors - tuned to different
mSUGRA topologies
Problem The bkg suppression efficiency should
be mSUGRA model independent, i.e. selection
efficiency Ei(Dbkg)const.
identify mSUGRA region, increases the model
significance. Can do a BLIND analysis, i.e.
exclude the signal topology and test the signal
and bkg hypothesis in neighbor regions. - very
model dependant . How to control bkg?
13
Region separation.
Use Neural Network (NN) for region separation.
Train NN for region i against other regions
j-i
Sig2/bkg13
Preselection efficiency
Preselect METgt50, Njgt2 (gt30GeV)
NN parameters MET, Heff, ETj1, ETj2, ETj3, Ptl1,
Nj,Njb cosf(j1j2), cosf(j1met),cosf(j2met)
3 test points (m0,m1/2,tanb50) Sel1(500,450) Sel
2 (1100,300) Sel3(2500,500)
Sel Efficiency
ExampleTeacher out
Selection efficiencies of 3 NNs trained for
Sig1/bkg23
Sig2/bkg13
Sig3/bkg12
Can constrain mSUGRA regions!
14
Constrain MSUGRA parameters
Can we separate channels and identify the
fractions(f.ex. gg, qq, etc)? Train NNs for
individual channels against others for one test
point (m0500 m1/2200)
qq
gg
cc
MC and Rec fractions
NN selection efficiencies ( m0500 m1/2200)
The selection efficiency E is a convolution of
NN and preselection efficiencies Eie (presl)ki
e (NN)k . Find the fractions F i from the
observed Di DiEi X Fi and compare with the
expectation from Mc(normalized)
Another point(m02500 m1/2600) with the same
NN( m0500 m1/2200)
Larger errors, but still can constrain the ratios.
The s(gg)s(qq)s(cc) ratios constrain the
mSUGRA region
15
Bkg suppression
Preselect data samples and use NN to train ttbar
against different mSUGRA regions, produce many
NNs and check efficiency for the bkg and signal.
Preselection METgt50, Njgt2(ETgt30), 2m(PTgt10)
NNs (preselection) efficiency for bkg(ttbar).
Best is const, i.e. region independent
NNs preselection efficiency for two mSUGRA points
m0500 m1/2200
m02500 m1/2600
bkgttbar
Preselection efficiency dominates, similar
pattern. Are there improvement in discovery?
Up to 10 spread, is it acceptable?
Still work to be done...
16
Combining results
ATLAS experience kinematic edges cross
sections Use of Markov chain to Likelihood
maximization in multiparameters space
Can do 1. Combine outputs of different selectors
in one likelihood L(p), where p are the model
parameters. 2. Find the most probable set of
parameters p for the observed data using the
Markov chain. p(md)p(dm) p(m) p(dm) PLi
Powerful method of hypothesis testing in
multiparameters space. Looks promising..
17
Summary
General CMS interest activity Fakes Work is
started. Will produce the large data samples
(Wjets, Zjets, QCD) during summer. Implement and
debug the FakesFinder
Hot topic- how to identify SUSY regions? (SUSY
BSM) Topology selector based on NN in inclusive
search is working in principle. Need optimization
of the bkg suppression for different
topologies Select of observables less prone to
uncertainties. Study of the bkg uncertainties
(PYTHIA-ALPGEN)-started.
Combining results and hypothesis testing with the
Markov chain. Interesting statistical aspect,
can be used in many studies. Just started...
This is just enough...
Anybody interested can joint our efforts... Fun
and important results are guaranteed.
18
Selection efficiency for the PTDR analysis
Use selection cuts defined in PTDR analysis. Most
of them are tuned to mo60, m12250 tanb10 bulk
region
tan?50 A00
19
Discovery reaches 10 fb-1
Taking the Nbkg from analysis, the results are
very close... PTDR is tuned to tanb10, here is 50