ATLAS+CMS: Boosted topologies (Run1 results, Run2 potential)
John Stupak III
on behalf of the ATLAS and CMS collaborations
Purdue University Calumet, Department of Chemistry and Physics,
2200 169th Street, Hammond, IN
As no sign of beyond the Standard Model physics has yet been observed at the LHC, experimental searches continue to probe ever higher mass scales. The decay of new, heavy resonances
can produce highly Lorentz boosted particles, such as top quarks and Higgs or vector bosons.
The decay products of these boosted particles are very highly collimated, and novel techniques
are needed to identify such decays. In this paper, recent analyses from ATLAS and CMS in
the boosted regime are reviewed, with an emphasis on jet substructure techniques and their
application.
1
Introduction
Many models of beyond the Standard Model (BSM) physics predict the existence of states which
decay to top quarks and Higgs or vector bosons. If the mass difference between the mother and
daughter particles is sufficient, the daughter particles are produced with significant Lorentz
boost and their decay products will be narrowly collimated. In the event the daughter decays
hadronically, the resulting shower produces a “fat” jet with a radius of R ≈ 2m/pT , where m
and pT are the mass and transverse momentum of the daughter, respectively.
In order to search for such a decay, one must contend with overwhelming background from
QCD interactions. Fortunately, a number of techniques have been proposed in the literature to
identify the hadronic decays of top quarks and Higgs and vector bosons. The most powerful
handle at our disposal to discriminate between jets from heavy particle decays and those from
light quarks and gluons (“QCD jets”) is the jet mass. Unfortunately, the jet mass is highly
sensitive to pileup and underlying event activity. However, a number of grooming algorithms
have been proposed (pruning [1], trimming [2], filtering [3], soft drop [4], etc.) to remove soft
and wide angle radiation from the jet clustering history, which significantly pushes the mass
distribution for light jets towards zero while having only a minimal effect on jets from heavy
particles decays. The effect of grooming on the jet mass distribution [5] and the stability with
respect to pileup [6] are shown in Figure 1.
In addition to the mass, a number of jet shape observables are useful for discriminating
between QCD jets and those from heavy flavor decays. The kT -splitting scale [7] and mass
drop [3] observables exploit the symmetric nature of heavy particle decays. N-subjettiness (τN )
variables [8] characterize the consistency of a jet with being composed of N or more subjets.
Energy correlation functions [9], quark/gluon likelihood [10], jet charge [11], pull angle [12], and
Q-jets volatility [13] are useful as well, among others.
Both ATLAS [14] and CMS [15] have developed b-tagging algorithms optimized for the
dense environment inside highly-collimated jets. ATLAS performs b-tagging on standard antikT R=0.4 jets and then performs a geometric matching to fat jets to tag them. CMS performs
ATLAS Simulation
jet
0.12
0.1
C/A LCW jets, 600 ≤ pT < 800 GeV
Ungroomed Z'→ tt
Ungroomed Dijets
Trimmed Z'→ tt
Trimmed Dijets
0.08
0.06
< mtrimming > (GeV)
Arbitrary units
0.14
120
110
100
90
80
CMS Simulation Preliminary
13 TeV
QCD, Anti-kT (R=0.8)
<nPU>= 40
p > 300 GeV
r sub=0.2,pT
=0.05
r sub=0.1,pT
=0.03
r sub=0.2,pT
=0.03
|η|< 2.5
r sub=0.3,pT
=0.03
PF with trimming
ungroomed
frac
frac
T
frac
frac
70
60
50
0.04
40
0.02
30
0
0
50
100
150
200 250 300
Jet mass [GeV]
20
15
20
25
30
35
40
45
50
nPV
Figure 1 – The jet mass distribution for the leading-pT jet in simulated Z 0 →tt and QCD dijet events before and
after trimming (left), and the stability of the average jet mass versus pileup for ungroomed and trimmed jets
(right).
b-tagging directly on fat jets or subjets, depending on the kinematic regime.
Finally, all of the above techniques can be combined into a dedicated “tagger.” This typically
involves the application of a groomed mass window requirement along with requirements on the
jet substructure. In the case of top and Higgs tagging, b-tagging can also be exploited. The
BDRS [3], HEP Top Tagger [16], and CMS Top Tagger [17] are all examples. These top taggers
also exploit the presence of a real W boson in the top decay chain, and require a pair of subjets
to be compatible with the W mass hypothesis.
2
2.1
Analyses in the Boosted Regime
Searches for Fermion+Fermion Resonances
Both ATLAS and CMS have extensive programs of searches for Z 0 →tt and W 0 → tb resonances.
For very high mass resonances with small cross sections, the all-hadronic decay modes with
large branching ratios are extremely important. It is also in this regime where jet substructure
techniques are most powerful.
CMS has performed a combination of Z 0 →tt searches in zero, one, and two lepton final
states [18]. The all-hadronic channel exhibits a dijet topology. Both high-pT jets were required
to be top-tagged. Separate optimizations were performed in the low and high mass regimes; in
the low (high) mass channel, the HEP Top Tagger (CMS Top Tagger) was used, which is based
on R=1.5 (R = 0.8) jets. Subjet b-tagging was applied, as well as a requirement on the ratio
τ32 = τ3 /τ2 in the high mass channel.
The dominant background in the all-hadronic final state is QCD dijet production. This
background was modeled using a sophisticated data-driven technique. No excess with respect
to the background expectation was observed. Model independent 95% confidence level (CL)
cross section upper limits are shown in Figure 2, based on a combination of all channels. These
were interpreted in the context of a variety of models to obtain mass exclusions. For a narrow
leptophobic topcolor [19] Z 0 resonance with ΓZ 0 /mZ 0 = 1.2%, masses below 2.4 TeV are excluded.
ATLAS performed a similar search in the lepton+jets final state [20]. Separate optimizations
were performed for the cases where the top quark decay products are merged into a single fat
jet, or resolved separately. Events were required to contain a high-pT lepton, at least 1 bjet and 1 top-jet, and large missing transverse energy (/
ET ) and transverse mass. Top √
tagging
was performed on trimmed R=1.0 jets, requiring m > 100 GeV and kT -splitting scale d12 >
40 GeV.
107
106
105
104
103
102
10
1
10-1
20
Data
SM tt
SM W+jets
Other SM
g 2.0 TeV, 15.3%
ATLAS
s=8 TeV, 20.3 fb-1
boosted l+jets
2
10
KK
Expected (95% CL)
Observed (95% CL)
Z' 1.2% width
±1σ Expected
±2σ Expected
CMS
Preliminary
10
1
10-1
0.5
1
1.5
2
2.5
3
3.5
10-2
1
0
19.7 fb-1 (8 TeV)
103
Upper limit on σZ' × B(Z' → tt) [pb]
Data/BG
Events/0.08 TeV
The tt invariant mass distribution, shown in Figure 2, was used to test for the presence of
signal. No significant excess was observed. Like the CMS analysis, 95% CL cross section upper
limits were established and interpreted in the context of a variety of models. For a narrow
leptophobic topcolor Z 0 resonance with ΓZ 0 /mZ 0 = 1.2%, masses below 1.8 TeV are excluded.
0
0.5
1
1.5
2
2.5
3
3.5
[TeV]
mreco
tt
10-3
0.5
1
1.5
2
2.5
3
MZ' [TeV]
Figure 2 – The ATLAS reconstructed tt invariant mass distribution in the boosted channel (left), and the CMS
observed 95% CL cross section upper limit (right).
Similar searches have been performed by ATLAS [21] and CMS [22] for W 0 →tb resonances.
2.2
Searches for Fermion+Boson Resonances
Many BSM theories predict the existence of fermion+boson resonances; from vector-like quarks
in little Higgs models [23; 24], models with extra dimensions [25; 26], and composite Higgs
models [25; 26; 27], to excited fermions in composite models [28; 29; 30]. CMS recently performed
searches for vector-like top [31] and bottom [32] quark partners, decaying via T 0 → tH and
B 0 → bH to all hadronic final states, as well as excited leptons [33] decaying via `? → `γ/`Z.
The search for `? → `Z considered both leptonic and hadronic Z decays; in the latter case, jet
substructure techniques were used to reject the overwhelming Z+jets background.
The all hadronic T 0 and B 0 searches were very challenging, requiring sophisticated jet substructure techniques to reject the QCD background. The T 0 analysis was particularly groundbreaking, as it represented the first use of a Higgs tagger combining both substructure information as well as subjet b-tagging, as well as the first vector-like quark search in an all-hadronic
final state.
The search was optimized for T 0 pair-production, where at least one T 0 decays via tH to the
all hadronic bbbjj final state. Events were selected with at least one top-jet and one Higgs-jet.
The top-jets were tagged with the HEP Top Tagger, also requiring a subjet b-tag. R=1.5 jets
were Higgs-tagged by requiring a double subjet b-tag and trimmed mass m > 60 GeV. Events
were categorized based on the number of Higgs-tagged jets in the event, and a joint likelihood
was constructed based on the scalar sum of the pT of all reconstructed jets and the mass of the
Higgs-tagged jet. No significant deviation from the background prediction was observed in this
likelihood distribution. 95% CL cross section upper limits were derived, and interpreted in the
triangular branching ratio space of a vector-like top quark partner, shown in Figure 3.
2.3
Searches for Diboson Resonances
A wide variety of diboson resonance searches have been conducted recently by ATLAS [34;
35] and CMS [36; 37; 38; 39] using jet substructure techniques. With the discovery of the
19.7 fb-1 (8 TeV)
T(bW)
σ(pp→ TT) (pb)
1
900
19.7 fb-1 (8 TeV)
850
CMS
0.8
800
750
0.6
-1
10
0
700
0
650
0.4
600
Obs. 95% CL
Exp. 95% CL
10
550
0.2
Theory TT
-2
500
Central 95% CL exp.
Central 68% CL exp.
1
600
800
1000
mT (GeV/c2)
1
0
T(tZ)
450
Observed T quark mass limit (GeV/c2)
CMS
T(tH)
Figure 3 – The 95% CL cross section upper limits for T 0 pair production (left) and T 0 mass lower limits (right).
Higgs boson, searches for W H and ZH resonances have become viable and are being pursued
vigorously. Initially these searches were focused on the dominant H → bb decay mode, but in
order to maximize search sensitivity recent searches have begun to investigate sub-dominant
Higgs decay modes as well.
The first search for a V H resonance in a fully-hadronic final state [36] included channels
optimized to select events consistent with H → bb and H → W W → 4q decays of the Higgs
boson. This required development of a novel H → 4q tagger. Pruned R=0.8 jets were used to
tag V → qq, H → bb, and H → 4q decays. In addition to mass window requirements, b-tagging
and N-subjettiness information was utilized as well. B-tagging was applied either to the fat jet or
the subjets, depending on the geometric separation of the subjets. The V → qq tagger required
the N-subjettiness ratio τ21 = τ2 /τ1 to be small, while the H → 4q tagger instead required the
ratio τ42 = τ4 /τ2 to be small, owing to the 4-pronged nature of the decay. The t42 distribution
for the H → 4q signal and other processes is shown in Figure 4. No significant deviation with
respect to the background prediction was observed, and resonance masses below 1.7 TeV were
excluded in the Heavy Vector Triplet model [40], as shown in Figure 4.
19.7 fb-1 (8 TeV)
Data
25
20
Unmatched t → Wb → qq'b
CMS
10
Preliminary
σtotal × BR(V'→ HV) (pb)
Events / 0.01
3
30 ×10
Matched t→ Wb → qq'b
q/g MADGRAPH+PYTHIA
H → WW* → 4q
W/Z → qq'
15
110 < mj < 135 GeV
10
5
0
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
N -subjettiness ratio τ42
19.7 fb-1 (8TeV)
CMS
Preliminary
Observed
Expected (68%)
Expected (95%)
V'->HV, HVT(B)
1
10-1
10-2
10-3
1
1.5
2
2.5
Resonance mass (TeV)
Figure 4 – The τ42 distribution for various processes (left) and the 95% CL cross section upper limit for a resonance
decaying to V H (right).
Another recent analysis [37], optimized to search for a ZH resonance, developed a novel
H → τ τ tagger. In this analysis, pruned R=0.8 jets were used to tag Z-jets, along with a
mass window requirement and a requirement on τ21 . The H → τ τ tagger also used pruned
R=0.8 jets as a starting point. Jets with a large mass drop µ1,2 = max(m1 , m2 )/m12 were
used as inputs to the hadron-plus-strips algorithm [41] with modified isolation requirements. A
likelihood-based fit was performed to reconstruct the H candidate from the visible daughters,
and a mass window requirement was subsequently applied. Again, no significant deviation from
the backrgound prediction was observed.
2.4
Searches for Supersymmetry
1
φ [radians]
1
6
5
0
0
~~
~
g-g production, g→qq∼
χ ,∼
χ →qqq
ATLAS Simulation
1
1
6
5
4
3
3
T
T
4
p [GeV]
0
0
~~
~
g-g production, g→qq∼
χ ,∼
χ →qqq
ATLAS Simulation
p [GeV]
φ [radians]
Jet substructure techniques have recently found application in high mass stop searches [42;
43]. An ATLAS R-parity violating SUSY search [44] made use of a novel application of jet
substructure, so-called “accidental substructure.” The analysis was optimized to search for
gluon pair production, with cascades containing R-parity violating UDD couplings, ultimately
producing 10 or more final state partons.
Figure 5 shows a typical signal event clustered with anti-kT R=0.4 and R=1.0 jets. When
clustered with R=0.4 jets, 17 unique jets are reconstructed, whereas only 5 jets are reconstructed
with the larger R parameter. Unlike the jet substructure applications described above, here the
goal is not to capture all the decay products from a heavy parent in a single jet; but rather
to capture radiation from partons with different parents that “accidentally” fall in the same
fat jet, giving rise to large mass. The observable which then discriminates between signal and
background is the scalar sum of the masses (after trimming) of the four leading jets mΣ
J . This
has the advantage over more traditional analyses which rely on the scalar sum of the jet pT in
that it also exploits the rich angular structure of signal events.
2
3
10
2
102
102
1
1
anti-kt (R=0.4) jets
0
−4
3
10
−3
−2
−1
anti-kt (R=1.0) jets
kt (R=0.3) subjets
10
0
1
2
3
4
y
1
0
−4
−3
−2
−1
10
0
1
2
3
4
y
1
Figure 5 – A signal event clustered with anti-kT R=0.4 (left) and anti-kT R=1.0 (right) jets.
In addition to containing at least four R=1.0 jets, selected events were required to have a
small separation in η between the two leading jets. Backgrounds were modeled with a datadriven approach. The observed mΣ
J distribution is shown in Figure 6, which agrees well with the
background prediction. The resulting 95% CL mass limits in the mχe0 vs. meg plane are shown
1
in Figure 6.
2.5
Standard Model Measurements
The jet substructure techniques outlined above are now sufficiently-well understood for use in
precision measurements. As such, they were recently used in a V +jets cross section measurement [45], as well as a tt differential cross section measurement [46]. These measurements were
able to extend earlier leptonic measurements to a previously inaccessible kinematic regime.
525 - 725 GeV
960±25
> 725 GeV
71±7.0
200
+32
27
1070 280.±8.4±57±49
79
77±0.60±3.2
35.±2.9±18±6.0 35±0.40±9.9±9.0
Table III. Table showing the predicted in the SM and observed number of events in SR100
scenarios. The background uncertainties are displayed as statistical + systematic; the
statistical + systematic + theoretical.
103
102
3-Jet Template, Data
4-Jet Sample, Data, SR
Uncertainties
∼0) = 800, 175 GeV
m(~
g, χ
1
ATLAS
s = 8 TeV, 20.3 fb-1
3
4jSR, p > 100 GeV
T
Ratio to Template
10
2.0
1.5
1.0
0.5
0.0
1
104
mχ∼0 [GeV]
Events / 50 GeV
13
~~
~
∼
permit
studies
full MJ⌃ spectrum as well M
as⌃ comg-g production,
g→ qq∼
χ ,of
χ → the
qqq
J results in SR100 a
parisons of
data
and
SM
predictions without significant
that can be easily rein
Expected limit (±1 σ )
ATLAS
signal contamination.
ses. In the case of the b
Observed limit (±1 σ
)
0
0
1
1
exp
SUSY
theory
fit (where the number a
is performed that take
s = 8 TeV,
20.3 fb-1
each MJ⌃ range. This
1000 C.
Validation
and tisystematic
uncertainties
es
n
i
ta
tivity to small deviatio
r
ce
Un
n
The correlation of th
tio
a
i
d
Many tests dare
performed
using the 3jCR as both
the
Ra
en
MJ⌃ spectrum
are acc
D
d
bi
UD
or
training sample
order
to matrix. Th
to and the kinematic sample in correlation
∼χ f
e
qq
Du
~g→ the
ed
the different bins as fu
determine
robustness of the method. The selection
at
alu
ev
which
is the largest co
500
requires Unthat
there be exactly three large-R jets
in the
event, as described in Tab. I. The dependence of tainties.
the tem-All other unc
spectrum
plate on the jet in question (leading, subleading,
etc) is as fully corr
All limits at 95% CL
0
1
Total Jet Mass [GeV]
Events / 50 GeV
tested, as well as the dependence of the template on the
Figure 9 shows bot
jet kinematics. It is determined that it is optimal
to
CL limits in the (mg̃ ,
0
0.2
define
separate
templates
for
each
of
the
three
jet
cate400
600
800
1000
1200
1400
region that provides t
m~g [GeV]
gories (leading, subleading, and third
jet) and toforbin
themass combina
each
5
templates according to the jet pT and ⌘. In the
4-jet
the expected
exclusion
(a) MJ⌃ in SR100 using p3T > 100 GeV
regions, the fourth jet uses the template derivedresents
for the
the experiment
Σ
Figure 6 – The predicted and observed mJ distribution forFigure
selected
events
(left)
and exclusion
the
expected
observed
9. Expected
observed
limits
the (m
g̃ ,
solid line
third
jet inand
the
3jCR:
tests
in
theinand
4jCR
and
4jVR
in-shows the ob
95% CL limits in the me
plane (right).
m ˜0 ) plane
forvery
the ten-quark
model givenbetween
by the total-jetlines
indicating
the ±1
0 vs. me
dicate
good
agreement
this
template
and
1
χ
g
1
104
3-Jet Template, Data
mass analysis. Limits are obtained by using the signal re- ⌃ certainties on the sign
the
observed
spectrum.
As
a
first
test,
the
M
template
J
gion with the best expected sensitivity at each point. The
4-Jet Sample, Data, SR
by renormalization an
from
3-jetlimits
kinematic
dashedconstructed
black lines show
the the
expected
at 95% sample
CL, with is compared
103
Uncertainties
certainties. All mass
⌃
to (yellow)
the actual
in 3-jet due
events,
and
very the 1 SUS
the light
bandsM
indicating
the 1 excursions
to
J distribution
m(~
g, ∼
χ ) = 800, 175 GeV
assuming
theo
Previous 2V +jets cross section
measurements inexperimental
leptonic
decay
modes
only
probed
the
region
and
background-only
theory
uncertainties.
Obgood agreement is observed.
10
At low m ˜01 , the region
ATLAS
served limits are indicated by medium dark (maroon) curves,
There
are
two
intrinsic
sources
of
systematic
uncerof phase space with vector boson
p
<
300
GeV.
In
the
recent
ATLAS
analysis
[45]
based
T20.3 fb
s = 8 TeV,
is excluded. Exclud
where the solid contour represents the nominal limit, and the
4jSR, p > 250 GeV
tainty
associated
the
template
procedure:
the
dottedby
linesselecting
are
obtainedevents
bywith
varying
the
signal
cross-section
inguncerm ˜01 , up to a max
on hadronic10decays, the cross section
was measured
with
R=0.6
jets with
due to finite
statistics scale
in the
3jCR
sample
by the tainty
renormalization
and factorization
and
PDF training
un< 870 GeV at m
m
g̃
⇠
pT > 320 GeV
was applied.
A further
enhancement
certainties.
1 and |η| < 1.9. A mass window requirement
(the variance),
and the
uncertainty
due to the m
smooth˜01 > 650 GeV are ex
of the signal sensitivity was obtained with the use of a ing
likelihood
discriminant
constructed
from
procedure in the template derivation (the bias). The
formersection
is estimated
by generating
2.0
jet shape variables
in the jet rest frame. The V +jets cross
was obtained
fromanaensemble
binned, of MJ⌃
1.5
templates and taking the ±1 deviations (defined as the
Mass [GeV]
maximum-likelihood
fit toTotal
theJetobserved
jet mass distribution.
A value of 8.5 ± 0.8(stat) ±
1.0
±34% quantile) with respect to the median of those vari0.5
1.5(syst) pb was obtained, in reasonable ×agreement
with as
the
theoretical
prediction
of unations
theNLO
uncertainty,
bin by bin.
The systematic
103
0.0
0.2
0.5
0.8
1
1.2
certainty due to the smoothing procedure is determined
5.1 ± 0.5 pb. 0
using the fact that a Gaussian kernel smoothing is apTotal Jet Mass [GeV]
plied to the template. The full difference between the
(b) MJ⌃ in SR250 using p3T > 250 GeV
nominal template and a template constructed using a
3 Conclusion and
Outlook
leading-order correction for the bias, derived analytically
Figure 8. Total-jet-mass in the 4jSR (a) using p3T > 100 GeV
in Ref. [69], is taken as the systematic uncertainty. The
Novel
jet substructure
provedForextremely
during
Run I ofdue
theto LHC.
The sensi(SR100)
and (b) using p3Ttechniques
> 250 GeV (SR250).
the SR100 useful
systematic
uncertainty
finite control
region statisselection,
the searches
reweightedwas
template
(built in
the 3jCR, andthough
tics is
chosen
to be
larger
(by setting
the size of the kernel
tivity
of many
increased
significantly
their
use,
and
precision
measurements
reweighted jet-by-jet in the 4jCR) is shown in the hatched
smoothing)
than that
dueCMS
to the
smoothing
procedure
wereblue
extended
extreme
kinematic
regimes.
During
Run II, ATLAS
and
will
probe yet
histogram.toThe
total systematic
uncertainty
due to
the
since the former is more accurately estimated.
smoothing
and At these scales, heavy particles will be
higher
mass procedure,
scales infinite
the statistics
search in
forcontrol
new regions,
physics.
proFigure 7 presents the total-jet-mass MJ⌃ in
the 4jVR
the difference between template prediction and the data ob3
duced
with
significant
boost,
making
jet
substructure
techniques
essential
in
many
analyses
(if
using pT > 100 GeV. A small level of disagreement
is
served in the 4jCR is shown in green.
observedPileup
when comparing
the will
observed
to the prethey are not already). New challenges will also be presented.
mitigation
be amass
serious
dicted mass in the 4jCR: a reweighting derived in the
challenge with higher instantaneous luminosity and 254jCR
ns bunch
spacing. It will also be chal(as a function of the jet mass) is then applied to
The
jet
multiplicity
and
|
⌘|
are
used
to
define
the
lenging to keep all-hadronic trigger rates at acceptable
without
losing
signal of the
thelevels
individual
jet masses
priorsignificant
to the construction
control regions. The 3jCR, with exactly three jets, is
⌃
MJwith
for each
event.
After
the
reweighting
the agreement
efficiency.
Fortunately,
the
experimental
collaborations,
input
form
the
theory
community,
used to train the background templates previously disis substantially
improved
at high total-jet-mass.
are cussed.
alreadyInwell
on theircontrol
way to
these challenges,
and jet
substructure
techniques The
the remaining
andaddressing
validation regions,
reweighted template agrees very well with the observed
requiring
4 jets, tool
the | during
⌘| selection
will each
remain
a powerful
Runsuppresses
II of thethe
LHC.
signal contribution and is used to define the 4jCR and
4jVR. In the 4-jet regions, the | ⌘| selection value for
5 It is observed that the difference between the leading and subthe control regions is chosen to be larger than an inverReferences
leading jet templates is minimal, but that the third jet exhibits
sion of the signal region selection, resulting in the selecqualitatively different masses as function of the jet pT .
tions presented in Tab. I. These control region definitions
×103
0.5
0.8
1
1.2
Total Jet Mass [GeV]
0
1
-1
3
Ratio to Template
T
[1] S. D. Ellis, C. K. Vermilion, and J. R. Walsh, “Recombination Algorithms and Jet
Substructure: Pruning as a Tool for Heavy Particle Searches,” Phys.Rev. D81 (2010)
arXiv:0912.0033.
[2] D. Krohn, J. Thaler, and L.-T. Wang, “Jet Trimming,” JHEP 1002 (2010)
arXiv:0912.1342.
[3] J. M. Butterworth, A. R. Davison, M. Rubin, and G. P. Salam, “Jet substructure as a
new Higgs search channel at the LHC,” Phys.Rev.Lett. 100 (2008) arXiv:0802.2470.
[4] A. J. Larkoski, S. Marzani, G. Soyez, and J. Thaler, “Soft Drop,” JHEP 1405 (2014)
arXiv:1402.2657.
[5] ATLAS Collaboration, “Performance of jet substructure techniques for large-R jets in
√
proton-proton collisions at s = 7 TeV using the ATLAS detector,” JHEP 1309 (2013)
arXiv:1306.4945.
[6] CMS Collaboration, “Pileup Removal Algorithms,” Tech. Rep. CMS-PAS-JME-14-001,
CERN, Geneva, 2014.
[7] S. D. Ellis and D. E. Soper, “Successive combination jet algorithm for hadron collisions,”
Phys.Rev. D48 (1993) arXiv:hep-ph/9305266.
[8] J. Thaler and K. Van Tilburg, “Identifying Boosted Objects with N-subjettiness,” JHEP
1103 (2011) arXiv:1011.2268.
[9] A. J. Larkoski, G. P. Salam, and J. Thaler, “Energy Correlation Functions for Jet
Substructure,” JHEP 1306 (2013) arXiv:1305.0007.
[10] CMS Collaboration, “Performance of quark/gluon discrimination in 8 TeV pp data,”
Tech. Rep. CMS-PAS-JME-13-002, CERN, Geneva, 2013.
[11] D. Krohn, M. D. Schwartz, T. Lin, and W. J. Waalewijn, “Jet Charge at the LHC,”
Phys.Rev.Lett. 110 (2013) arXiv:1209.2421.
[12] J. Gallicchio and M. D. Schwartz, “Seeing in Color: Jet Superstructure,” Phys.Rev.Lett.
105 (2010) arXiv:1001.5027.
[13] S. D. Ellis, A. Hornig, T. S. Roy, D. Krohn, and M. D. Schwartz, “Qjets: A
Non-Deterministic Approach to Tree-Based Jet Substructure,” Phys.Rev.Lett. 108 (2012)
arXiv:1201.1914.
[14] ATLAS Collaboration, “The ATLAS Experiment at the CERN Large Hadron Collider,”
JINST 3 (2008).
[15] CMS Collaboration, “The CMS experiment at the CERN LHC,” Journal of
Instrumentation 3 (2008).
[16] T. Plehn, M. Spannowsky, and M. Takeuchi, “How to Improve Top Tagging,” Phys.Rev.
D85 (2012) arXiv:1111.5034.
[17] CMS Collaboration, “Boosted Top Jet Tagging at CMS,” Tech. Rep.
CMS-PAS-JME-13-007, CERN, Geneva, 2014.
[18] CMS Collaboration, “Search for resonant ttbar production in proton-proton collisions at
sqrt(s)=8 TeV,” arXiv:1506.03062.
[19] R. M. Harris and S. Jain, “Cross Sections for Leptophobic Topcolor Z’ Decaying to
Top-Antitop,” Eur.Phys.J. C72 (2012) arXiv:1112.4928.
[20] ATLAS Collaboration, “A search for tt̄ resonances using lepton-plus-jets events in
√
proton-proton collisions at s = 8 TeV with the ATLAS detector,” arXiv:1505.07018.
√
[21] ATLAS Collaboration, “Search for W 0 → tb → qqbb decays in pp collisions at s = 8 TeV
with the ATLAS detector,” Eur.Phys.J. C75 (2015) arXiv:1408.0886.
[22] CMS Collaboration, “Search for W prime to tbbar in the all-hadronic final state,” Tech.
Rep. CMS-PAS-B2G-12-009, CERN, Geneva, 2014.
[23] N. Arkani-Hamed, A. G. Cohen, and H. Georgi, “Electroweak symmetry breaking from
dimensional deconstruction,” Phys.Lett. B513 (2001) arXiv:hep-ph/0105239.
[24] M. Schmaltz and D. Tucker-Smith, “Little Higgs review,” Ann.Rev.Nucl.Part.Sci. 55
(2005) arXiv:hep-ph/0502182.
[25] I. Antoniadis, K. Benakli, and M. Quiros, “Finite Higgs mass without supersymmetry,”
New J.Phys. 3 (2001) arXiv:hep-th/0108005.
[26] Y. Hosotani, S. Noda, and K. Takenaga, “Dynamical gauge-Higgs unification in the
electroweak theory,” Phys.Lett. B607 (2005) arXiv:hep-ph/0410193.
[27] K. Agashe, R. Contino, and A. Pomarol, “The Minimal composite Higgs model,”
Nucl.Phys. B719 (2005) arXiv:hep-ph/0412089.
[28] J. Pati, A. Salam, and J. Strathdee, “Are quarks composite?,” Physics Letters B 59
(1975).
[29] O. W. Greenberg and C. A. Nelson, “Composite models of leptons,” Phys. Rev. D 10 Oct
(1974).
[30] U. Baur, M. Spira, and P. M. Zerwas, “Excited-quark and -lepton production at hadron
colliders,” Phys. Rev. D 42 Aug (1990).
[31] CMS Collaboration, “Search for vector-like T quarks decaying to top quarks and Higgs
bosons in the all-hadronic channel using jet substructure,” arXiv:1503.01952.
[32] CMS Collaboration, “Search for pair-produced vector-like quarks of charge -1/3 decaying
to bH using boosted Higgs jet-tagging in pp collisions at sqrt(s) = 8 TeV,” Tech. Rep.
CMS-PAS-B2G-14-001, CERN, Geneva, 2014.
√
[33] CMS Collaboration, “Search for excited leptons in proton proton collisions at s=8
TeV,” Tech. Rep. CMS-PAS-EXO-14-015, CERN, Geneva, 2015.
[34] ATLAS Collaboration, “Search for production of W W/W Z resonances decaying to a
√
lepton, neutrino and jets in pp collisions at s = 8 TeV with the ATLAS detector,”
Eur.Phys.J. C75 (2015) arXiv:1503.04677.
[35] ATLAS Collaboration, “Search for resonant diboson production in the ``q q̄ final state in
√
pp collisions at s = 8 TeV with the ATLAS detector,” Eur.Phys.J. C75 (2015)
arXiv:1409.6190.
[36] CMS Collaboration, “Search for a massive resonance decaying into a Higgs boson and a
W or Z boson in hadronic final states in proton-proton collisions at sqrt(s) = 8 TeV,”
Tech. Rep. CMS-PAS-EXO-14-009, CERN, Geneva, 2015.
[37] CMS Collaboration, “Search for narrow high-mass resonances in proton-proton collisions
√
at s = 8 TeV decaying to Z and Higgs bosons,” arXiv:1502.04994.
[38] CMS Collaboration, “Search for massive resonances in dijet systems containing jets
√
tagged as W or Z boson decays in pp collisions at s = 8 TeV,” JHEP 1408 (2014)
arXiv:1405.1994.
[39] CMS Collaboration, “Search for massive resonances decaying into pairs of boosted bosons
√
in semi-leptonic final states at s = 8 TeV,” JHEP 1408 (2014) arXiv:1405.3447.
[40] D. Pappadopulo, A. Thamm, R. Torre, and A. Wulzer, “Heavy Vector Triplets: Bridging
Theory and Data,” JHEP 1409 (2014) arXiv:1402.4431.
[41] CMS Collaboration, “Performance of τ -lepton reconstruction and identification in CMS,”
Journal of Instrumentation 7 (2012).
[42] ATLAS Collaboration, “Search for direct pair production of the top squark in all-hadronic
√
final states in proton-proton collisions at s = 8 TeV with the ATLAS detector,” JHEP
1409 (2014) arXiv:1406.1122.
[43] ATLAS Collaboration, “Search for top squark pair production in final states with one
√
isolated lepton, jets, and missing transverse momentum in s =8 TeV pp collisions with
the ATLAS detector,” JHEP 1411 (2014) arXiv:1407.0583.
[44] ATLAS Collaboration, “Search for massive supersymmetric particles decaying to many
√
jets using the ATLAS detector in pp collisions at s = 8 TeV,” Phys.Rev. D91 (2015)
arXiv:1502.05686.
[45] ATLAS Collaboration, “Measurement of the cross-section of high transverse momentum
vector bosons reconstructed as single jets and studies of jet substructure in pp collisions at
√
s = 7 TeV with the ATLAS detector,” New J.Phys. 16 (2014) arXiv:1407.0800.
[46] ATLAS Collaboration, “Measurement of the differential cross-section of highly boosted
√
top quarks as a function of their transverse momentum using the ATLAS detector in s
= 8 TeV proton-proton collisions,” Tech. Rep. ATLAS-CONF-2014-057, CERN, Geneva,
Sep 2014.