Jets and jet substructure - Physics Learning Laboratories · Gavin Salam (CERN) Jets and jet...

Jets and jet

substructure

1

Gavin Salam (CERN)with extensive use of material by Matteo Cacciari

and Gregory Soyez

TASIJune 2013

Gavin Salam (CERN) Jets and jet substructure (1) TASI, June 2013 2

JETSCollimated,

energetic bunches of particles


Sterman and Weinberg, Phys. Rev. Lett. 39, 1436 (1977):

Jets date back to the late 1970s


e+e� ! qq̄gevent at LEP

√s = 91.2 GeV, in 1990s

And they’ve been used and studied at every

collider since

event at TASSO√s = 27.4 GeV, in 1979

gluon discovery:



√s = 91.2 GeV, in 1990s


collider since


gluon discovery:

e

p>2.5

p

CTD

CAL

Yoke RMUFMUp>4 GeV

Inner chambersµ

Toroids

Outer chambersµ

BMU pt>2.5 GeV

ZEUS



√s = 91.2 GeV, in 1990s


collider since


gluon discovery:

e

p>2.5

p

CTD

CAL

Yoke RMUFMUp>4 GeV

Inner chambersµ

Toroids

Outer chambersµ

BMU pt>2.5 GeV

ZEUS

CDF


464 records foundFind all papers by ATLAS and CMS


Pull out those that refer to one widely used jet-alg272 records found

60% of papers use jets!

Gavin Salam (CERN) Jets and jet substructure (1) TASI, June 2013

Jet usage at the LHC

7

+N

P co

rrectio

ns

Tree level Monte Carlo (N)NLO

Jets (theory tool)

MC + Tree

Jet X-sctDetector sim.

Jet X-sct

DETECTOR

Detector unfolding

CKKWMLM

BSM searches

Higgs physics

top physics

PDF fits

...

MC validation


Two key aspects to discussing jets

8

How jets come to have the structure they do

Jets as projections[Introduction]

[Background knowledge]

jet 1 jet 2

LO partons

Jet Def n

jet 1 jet 2

Jet Def n

NLO partons

jet 1 jet 2

Jet Def n

parton shower

jet 1 jet 2

Jet Def n

hadron level

π π

Kp φ

Projection to jets should be resilient to QCD e!ects

Gavin Salam (CERN) Jets and jet substructure (1) June 2013 8 / 35



jet 1 jet 2

LO partons

Jet Def n

jet 1 jet 2

Jet Def n

NLO partons

jet 1 jet 2

Jet Def n

parton shower

jet 1 jet 2

Jet Def n

hadron level

π π

Kp φ





jet 1 jet 2

LO partons

Jet Def n

jet 1 jet 2

Jet Def n

NLO partons

jet 1 jet 2

Jet Def n

parton shower

jet 1 jet 2

Jet Def n

hadron level

π π

Kp φ



How we “reconstruct” jets

Why do we see jets? Parton fragmentation[Introduction]


KL

π−

π+

π0

K+

non−perturbative

hadronisation

quark

Gluon emission:!

!sdE

E

d"

"! 1

At low scales:

!s " 1

High-energy partons unavoidably lead tocollimated bunches of hadrons


jet finding interpretation

quarkanti-

quark


Two key aspects to discussing jets

8

How jets come to have the structure they do



jet 1 jet 2

LO partons

Jet Def n

jet 1 jet 2

Jet Def n

NLO partons

jet 1 jet 2

Jet Def n

parton shower

jet 1 jet 2

Jet Def n

hadron level

π π

Kp φ





jet 1 jet 2

LO partons

Jet Def n

jet 1 jet 2

Jet Def n

NLO partons

jet 1 jet 2

Jet Def n

parton shower

jet 1 jet 2

Jet Def n

hadron level

π π

Kp φ





jet 1 jet 2

LO partons

Jet Def n

jet 1 jet 2

Jet Def n

NLO partons

jet 1 jet 2

Jet Def n

parton shower

jet 1 jet 2

Jet Def n

hadron level

π π

Kp φ



How we “reconstruct” jets



KL

π−

π+

π0

K+

non−perturbative

hadronisation

quark

Gluon emission:!

!sdE

E

d"

"! 1

At low scales:

!s " 1



jet finding interpretation

quarkanti-

quark

Most of the content of these lectures

Core understanding


Why do we see jets?

9



KL

π−

π+

π0

K+non−perturbative

hadronisation

quark

Gluon emission:!

!sdE

E

d"

"! 1

At low scales:

!s " 1



Z↵s

dE

E

d✓

✓� 1

Gluon emission

↵s ⇠ 1

Non-perturbativephysics


Why do we see jets?

9



KL

π−

π+

π0

K+non−perturbative

hadronisation

quark

Gluon emission:!

!sdE

E

d"

"! 1

At low scales:

!s " 1



Z↵s

dE

E

d✓

✓� 1

Gluon emission

↵s ⇠ 1

Non-perturbativephysics

soft-collinear structure (no colour)

i

/p+ /kigs✏/ v(p)

drop /k in numerator, use /p/k + /k/p = 2p.k, /pv(p) = 0

get gsp.✏

p.kv(p) take p = (1, 0, 0, 1), k = E(cos ✓, sin ✓, 0, 1), ✏ = (sin ✓,� cos ✓, 0, 0)

square the amplitude, put in phasespace:

d3k

2E(2⇡)3! E dE dcos ✓ d�

16⇡3


Reconstructing jets


Jet finding as a form of projection

11



jet 1 jet 2

LO partons

Jet Def n

jet 1 jet 2

Jet Def n

NLO partons

jet 1 jet 2

Jet Def n

parton shower

jet 1 jet 2

Jet Def n

hadron level

π π

Kp φ



Projection to jets should be resilient to QCD effects


Reconstructing jets is an ambiguous task

12

Seeing v. defining jets[Introduction]


Jets are what we see.Clearly(?) 2 jets here

How many jets do you see?Do you really want to ask yourselfthis question for 109 events?




12






2 clear jets



12






2 clear jets 3 jets?







2 clear jets 3 jets?or 4 jets?



Make a choice: specify a jet definition

14

{pi} {jk}jet definition

particles,4-momenta,

calorimeter towers, ....

jets

• Which particles do you put together into a same jet?• How do you recombine their momenta

(4-momentum sum is the obvious choice, right?)

“Jet [definitions] are legal contracts between theorists and experimentalists’’ -- MJ Tannenbaum

They’re also a way of organising the information in an event1000’s of particles per events, up to 20.000,000 events per second


There is no objectively "correct" clustering algorithm, but […] "clustering is in the eye of the beholder."[1] The most appropriate clustering algorithm for a particular problem often needs to be chosen experimentally, unless there is a mathematical reason to prefer one cluster model over another.


Partitioning / centroid-based clustering

[cone algorithms]

Matteo Cacciari - LPTHE MadGraph School - May 2013 - Beijing

k-means

18

Example of a partitional algorithm1) Choose K centroids at random2) Assign objects to closest centroid, forming K clusters3) Calculate centroid (mean of distances) of each cluster, update centroids4) Check if an object in a cluster is closer to another centroid. Reallocate in case.5) Repeat from step 3 until no object changes cluster anymore.

Step 1(random centroids)

Step 2(allocate objects)

Step 3(move centroids)

Step 5(end of iteration)

[Fig

ures

by

E. G

arre

tt-M

ayer

]

Matteo Cacciari - LPTHE MadGraph School - May 2013 - Beijing

k-means

18

Example of a partitional algorithm1) Choose K centroids at random2) Assign objects to closest centroid, forming K clusters3) Calculate centroid (mean of distances) of each cluster, update centroids4) Check if an object in a cluster is closer to another centroid. Reallocate in case.5) Repeat from step 3 until no object changes cluster anymore.

One of the main shortcomings: result of final convergence can be highly sensitive to choice of initial seeds.

Also, the concept of ‘mean distance’ (to calculate the centroid) must be defined.

Step 1(random centroids)

Step 2(allocate objects)

Step 3(move centroids)

Step 5(end of iteration)

[Fig

ures

by

E. G

arre

tt-M

ayer

]


Iterative Cone, Prog Removal (IC-PR)

60

50

40

20

00 1 2 3 4 y

30

10

pt/GeVOne of the simplest of the cone algs

e.g. CMS iterative cone

! Take hardest particle as seed for

cone axis

! Draw cone around seed

! Sum the momenta use as new

seed direction, iterate until stable

! Convert contents into a “jet” and

remove from event

Notes

! “Hardest particle” is collinear

unsafe more right away...

19



60

50

40

20

00 1 2 3 4 y

30

10

pt/GeV Seed = hardest_particleOne of the simplest of the cone algs



cone axis





remove from event

Notes



20



60

50

40

20

00 1 2 3 4 y

30

10

pt/GeV Draw coneOne of the simplest of the cone algs



cone axis





remove from event

Notes



21



60

50

40

20

00 1 2 3 4 y

30

10

pt/GeV sum of momenta != seedOne of the simplest of the cone algs



cone axis





remove from event

Notes



22



60

50

40

20

00 1 2 3 4 y

30

10

pt/GeV Iterate seedOne of the simplest of the cone algs



cone axis





remove from event

Notes



23



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



24



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



25



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



26



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



27



60

50

40

20

00 1 2 3 4 y

30

10

pt/GeV sum of momenta == seedOne of the simplest of the cone algs



cone axis





remove from event

Notes



28



60

50

40

20

00 1 2 3 4 y

30

10

pt/GeV Cone is stableOne of the simplest of the cone algs



cone axis





remove from event

Notes



29



60

50

40

20

00 1 2 3 4 y

30

10

pt/GeV Convert into jetOne of the simplest of the cone algs



cone axis





remove from event

Notes



30



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



31



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



32



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



33



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



34



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



35



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



36



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



37



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



38



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



39



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



40



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



41



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



42



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



43



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



44



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



45



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



46



60

50

40

20

00 1 2 3 4 y

30

10




cone axis





remove from event

Notes



47

Jets and jet substructure - Physics Learning Laboratories · Gavin Salam (CERN) Jets and jet...

Documents

Transcript of Jets and jet substructure - Physics Learning Laboratories · Gavin Salam (CERN) Jets and jet...