Choose timezone
Your profile timezone:
Powered by Indico
v3.3.2-pre
--------------------------------------------------------------------------------
A Summary of The 39th General Meeting of the ILC Physics Subgroup
Nov. 1, 2014 (Sat) at Room 425, Building 3, KEK
--------------------------------------------------------------------------------
0) Attendants:
Present at KEK:
A.Ishikawa (Tohoku)
J.Strube (Tohoku)
A.Dubey (Tohoku)
C.Calancha (KEK)
K.Fujii (KEK)
J.Fujimoto (KEK)
J.Tian (KEK)
T.Ogawa (Sokendai)
T.Tanabe (Tokyo)
K.Hidaka (Gakugei)
H.Ono (NDU)
T.Suehara (Kyushu)
Remote connection:
K.Kotera (Shinshu)
K.Tsuchimoto (Shinshu)
S.Kawada (Hiroshima)
Y.Sudo (Kyushu)
T.Tomita (Kyushu)
0) Opening comments (K.Fujii)
1) ZH: Recoil mass using Z->ll (Jan Strube on behalf of Shun Watanuki)
Status:
- Mass template method
-> x = 125.018 +/- 0.021 (GeV) 250fb-1 @ 250GeV
This should be considered as the ultimate stat. power.
-> trying to define the template function from 1st principle.
- CP mixture from production angle
-> Now trying to understand the data samples
SM: there seems to be some
BSM: physsim: generating samples for different sets of (a,b,bt)
Next:
- estimate correct eta error
- find optimal (a,b,bt) for eta analysis
- higgs decay plane
- start writing masters thesis.
2) ZH: Recoil mass using Z->qq (Tatsuhiko Tomita)
Status:
ZH->qqH @250GeV
ISR finder: efficiency=92%, purity=71%
Tau finder: efficiency78%, purity=62%
Cuts: mass box for ZZ, WW BGs.
Z mass
sphericity to kill 2f BG
Thrust to cut 2f BG
Result
-> delta sigma / sigma = 4.7% (Left pol), 3.3% (right pol)
Model independence?
-> checking efficiencies for different categories of event shapes:
#leptons, #taus, #jets
efficiencies are within 1% except for 0lep 0tau category.
Next Step:
- Use likelihood method for signal selection
- Study more the 0lep-0tau category
If you look at bb, cc, gg differently the efficiencies are different but
if you put them together the efficiency becomes pretty close to the average
efficiency. Need to understand why this is so.
- Improve tau separation by optimizing tau finder.
- Estimate systematic errors.
3) H->bb,cc,gg (Hiroaki Ono)
Status:
250fb-1@250GeV, 330fb-1@350GeV
Small difference from extrapolation from 120GeV.
Maybe this is due to low BG statistics for qqH.
-> still under investigation.
New X-likeliness (X1+X2)/2 seems to give a better result.
e.g. h->bb: no change
h->cc: 12.6% -> 11.3% @ 250GeV
: 8.8% -> 7.9%
h->gg: 8.6% -> 7.7%
6.0% -> 5.4%
kt clustering tried for qqh @ 250GeV
-> no significant difference as expected
(since gam gam -> hadron BG is small at 250GeV)
Next Step:
- TMVA
- Evaluate different polarization case.
- Update 500GeV analysis.
4) Higgs to tau tau (Shin-ichi Kawada)
Status:
Current numbers:
250fb-1@ 250GeV
3.4% qqh : done
46.0% vvh : extrapolation from mh=120GeV
16.1% eeh : same
14.7% mmh : same
500fb-1 @ 500GeV
qqh : 4.7% buggy tau finder
vvh : 6.8% zh ww-fusion separation
eeh : 31.2% MC stat. was not enough
mmh : 17.6% same
Today’s topic: vvh @ 250GeV
\x{2014}> 43.1% (cut based)
TMVA (BDT and BDTG) ongoing
Next:
- TMVA for vvh @ 250GeV
- new collinear
- llh @ 250GeV and then @500GeV.
5) ee -> HW at ILC (Jan Strube on behalf of Yuko Shinzaki)
Status:
LHC limit: m_H+ > 150GeV
Setup:
m_H+=150GeV
F_HWZ=1, F_HWA=0
Signal 1:
WH -> jj tau nu
Recoil mass
57.-sigmas -> 1.75%
Signal 2:
WH -> jj WZ*
Trying different methods to select the correct W for
recoil mass calculation.
C: (m_W,m_W(*)) (p_W, p_W(*))
Next:
- For H+->tau nu, derive mH vs FHWZ limit
- H+->WZ
jet paring
C: Use single leptonic or even both leptonic W decay modes.
6) Heavy Higgs search at ILC (Abhinav Dubey)
Status:
e+e- -> HA -> bbbb
mA=mH=400GeV, tanb=10, 2.4fb @ 1TeV
H->bb (77%) A->bb (65%)
Event selection:
jet pairs selected by chi2 requiring two pairs to have the same mass.
Preselection:
cut variables include thrust, major thrus, dphiJ1J2, dphiJ3J4, Et_vis
BDT analysis:
b-tag values, Y45, Y56, Y67, Y78
\x{2014}> efficiency=30%, significance=17%
Q: In the top right fig. of p12 why you have points below 100 and above 400 GeV?
A: Will check.
Next:
- Tag leptons
- Understand beam BG
- Limit on mA-tanb
- Extraction of model parameters/
7) enW (Koya Tsuchimoto)
Motivations:
Detector benchmark (JER)
Direct mW measurement
Anomalous WWA, WWZ couplings
Status:
Ecm=250GeV, 250fb-1
Strategy:
Identify the isolated electron and calculate the invariant mass of the rest
of the final state particles.
Next Step:
- ISR elimination.
- Better jet clustering algorithm.
C: Take a look at Graham Wilson’s work.
C: Prepare MC sample w/o ISR, beam BG overlay and try it to separate
the effect of ISR and beam BG from the detector effect.
8) Top Yukawa coupling study at 500 GeV (Yuji Sudo)
Status:
new: overlay low Pt beam background
- Setup:
mh=125 GeV, 500fb-1 at Ecm=500GeV,
beam polarization=(-0.8,+0.3)
- Results:
As of LCWS2014
S/sqrt(S+B)=2.35 for 8-jet
=2.04 for l+6-jet
Systematic error degrade significance < 5%.
Now trying to improve lepton ID and to include the tth to the 2Lbb mode.
S/sqrt(S+B)=2.38 for 8-jet
=2.11 for l+6-jet
= 0.8 for 2lbb [NEW]
Nest Step:
- Lepton ID using cluster shape (cf. Kurata).
- Increase tbW statistics.
- MVA and MC generation needed for it.
- Optimize 2lbb selection.
Q: How many tau nu events are included in lv6J sample?
A: home work.
9) TPC radius (Tomohisa Ogawa)
Motivation:
Assess the effect of changing the TPC outer radius on recoil mass measurements.
Results:
Signal only study:
At Ecm=250GeV
Recoil mass resolution degrades, when reducing R_out from 1.8m to 1.4m, by
~33% (detector resolution only)
~10% (including beam energy spread)
At Ecm=350GeV
Recoil mass resolution degrades, when reducing R_out from 1.8m to 1.4m, by
~41% (detector resolution only)
~25% (including beam energy spread)
Inclusion of BG
Line shape model: GPET or Kernel Fn.
As expected the Kernel function models the line shape better and mitigates
the systematic mass shift.
At Ecm=250GeV
Reducing R_out from 1.8m to 1.4m makes resolution worse by
~10% (detector resolution)
<~ 5% (sigma_zh)
~12% (mh)
At Ecm=350GeV
Reducing R_out from 1.8m to 1.4m makes resolution worse by
~25% (detector resolution only)
<~ 5% (sigma_zh)
~30% (mh)
10) H -> gamma gamma, mu+ mu- (Tino Calancha)
Current Status:
h->mm @ 1TeV
S/sqrt(S+B) -> 44% : 500fb-1
h->gam gam
18.9% @ 500GeV (500fb-1)
7.4% @ 1TeV (1000fb-1)
both for vvh (W-fusion)
Next:
- concentrate on 500GeV
- 500GeV with qqh, vvh and prepare detailed documentation.
- h->mm at 500GeV.
Q: Z gamma?
C: e+e- -> h gamma?
11) LCFIPlus (Taikan Suehara)
Status:
NIM paper will be submitted very soon.
Jan and Masakazu joined the collaboration.
Plan:
- automatic creation of joint prob./ bc-q prob. (Taikan)
- pair BG (Jan)
- jet clustering - overlay tagger
FastJet inclusion? (Tomohiko)
- vertex charge optimization (volunteer?)
- better lepton ID/PID (Masakazu)
- pi0 tagging using w/wo conversion (Masakazu)
- more optimization on (volunteer?)
- fewer tracks (1+1 etc)
- vertices closer to IP (current cutoff is at 300um)
- other ideas?
12) ME method (Junping Tian)
Status:
Verification (ZHH)
phi_F-M(Z) plane \x{2014}> weighted by 1/|ME|^2 -> flat
eeH analysis
ln(L0/L1) is the best discriminant
\x{2014}> better than MLP, BDT, Likelihood
Recoil mass analysis
M = M(mmh) x D(propagator of h)
Tested w/o ISR and beamstrahlung.
\x{2014}> ISR deteriorates the separation of signal from BG.
Next Step:
- ISR treatment.
- Use ME method.
- Recoil mass analysis with MEM.
- Then to the self-coupling analysis.
13) Physics case for the ILC: Post LHC8 (Tomohiko Tanabe)
Tomohiko showed the slides prepared primarily for the ILC review by the MEXT
Physics WG.