1
00:00:00,000 --> 00:00:01,620
vector boson scattering and vector boson fusion.

2
00:00:03,179 --> 00:00:05,489
So you can start whenever you want Heather.

3
00:00:10,950 --> 00:00:12,360
you have to unmute Heather.

4
00:00:13,710 --> 00:00:16,620
I do have to unmute. that might help,
wouldn't it? okay. uh.

5
00:00:19,980 --> 00:00:23,580
right. so I'm going to talk about vector
boson fusion and vector boson

6
00:00:23,580 --> 00:00:28,050
scattering measurements. and what uh I mean
by this, of course, is that you have

7
00:00:28,050 --> 00:00:32,370
incoming quarks that emit vector bosons and
then those vector bosons interact at a

8
00:00:32,370 --> 00:00:36,360
point, either fused into a single vector boson
or scatter into two new vector

9
00:00:36,360 --> 00:00:41,430
bosons. however, there is an incredibly
important point I need to mention here

10
00:00:41,430 --> 00:00:46,230
before we carry on is that we cannot
directly measure these V B F or V B S

11
00:00:46,230 --> 00:00:50,190
processes. so there's a significant amount
of interference interference with other

12
00:00:50,190 --> 00:00:55,410
diagrams that are of the same order of
alpha electroweak and extracting the V B F

13
00:00:55,440 --> 00:00:59,760
or V B S component is not a gauge invariant
operation. so we cannot measure these

14
00:00:59,760 --> 00:01:00,000
processes

15
00:01:00,000 --> 00:01:05,460
directly. and instead what we do is
measure electroweak production of V J J

16
00:01:05,490 --> 00:01:11,700
of course V is a W boson, a Z boson or a photon,
or V V J J. and I will try to use that

17
00:01:11,700 --> 00:01:14,970
language throughout my talk to be
technically correct, um I might fail.

18
00:01:15,990 --> 00:01:21,120
so aside from our uh electroweak production
V B F electroweak production diagram, I have

19
00:01:21,120 --> 00:01:24,960
another one here, you can see it has the
same order in alpha electroweak on the

20
00:01:25,020 --> 00:01:33,300
bottom left, we also have a very, a very
large amount of strong V J J production. in

21
00:01:33,300 --> 00:01:37,890
general, this has a much higher cross
section with one exception of the same

22
00:01:37,890 --> 00:01:39,480
sign W W

23
00:01:40,650 --> 00:01:48,120
production. so because it's so large, uh we
have a huge task in understanding and

24
00:01:48,150 --> 00:01:51,900
reducing this background that it's
absolutely crucial to do an effective

25
00:01:52,050 --> 00:01:56,370
electroweak V J J or V V J J measurement. um so
that's why most of my talk is actually

26
00:01:56,370 --> 00:01:59,550
going to be discussing these strong or QCD
backgrounds um.

27
00:02:00,000 --> 00:02:03,390
and I will use the word sort of strong and
QCD interchangeably because different

28
00:02:03,390 --> 00:02:06,480
analyses are talked about and call them
different things sort of in their in their

29
00:02:06,750 --> 00:02:09,480
tables and plots, but it means the same
thing.

30
00:02:11,309 --> 00:02:15,539
okay, so I'm going to start by discussing
our brand new electroweak Z J J production

31
00:02:15,539 --> 00:02:21,329
result. the general characteristics of the
events here in the electroweak events are

32
00:02:21,419 --> 00:02:27,089
a leptonically decaying Z boson um and
two jets with a large rapidity separation

33
00:02:27,359 --> 00:02:32,669
and a large invariant mass. additionally, um in
the electroweak production, there's

34
00:02:32,669 --> 00:02:38,909
usually no jets in the gap between these
these two forward fowards jets. um whereas

35
00:02:38,909 --> 00:02:41,789
with for instance, strong production or
main background, there tend to be jets

36
00:02:41,789 --> 00:02:44,999
radiated every which way because this is
what QCD tends to like to do.

37
00:02:46,680 --> 00:02:47,250
so,

38
00:02:48,630 --> 00:02:55,110
here I have a plot that shows the Monte
Carlo data ratio for different processes,

39
00:02:56,040 --> 00:03:00,000
um the strong in blue, electroweak and red
and then the remaining backgrounds

40
00:03:00,390 --> 00:03:06,420
tiny in orange. and you can see this
across the spectrum for M J J. and the

41
00:03:06,510 --> 00:03:10,590
important point takeaway point is a few things
here. first off, the strong background is

42
00:03:10,590 --> 00:03:12,210
the largest across the spectrum.

43
00:03:13,440 --> 00:03:19,470
and the second one is that generators for
both strong and electroweak production

44
00:03:19,470 --> 00:03:25,110
show huge discrepancies between
themselves. um and then, of course, then, none

45
00:03:25,110 --> 00:03:30,990
of them are really modeling the data that
well in the strong Z J J case. it's also

46
00:03:30,990 --> 00:03:34,920
important to note that the strong Z J J
modeling is especially poor in the signal

47
00:03:34,920 --> 00:03:37,590
enriched region when you see this big
divergence between

48
00:03:39,000 --> 00:03:43,050
one of the predictions and the other two.
Uh, and this is right where we want to do our

49
00:03:43,050 --> 00:03:48,240
measurement. so we can't just take the
 Monte Carlo at face value and carry

50
00:03:48,240 --> 00:03:52,890
on. instead, we have to deal with data
driven background modeling, and this will

51
00:03:52,890 --> 00:03:58,710
take care of the, the miss-modeling. so in order to
do this, the analysis is split into four

52
00:03:58,710 --> 00:04:00,000
regions with two uncorrelated

53
00:04:00,000 --> 00:04:04,440
 variables. the first one, the y-axis
here is the number of jets in the gap

54
00:04:04,440 --> 00:04:08,520
between the leading and subleading jets.
and then the second one is the centrality

55
00:04:08,550 --> 00:04:11,580
of the reconstructed Z boson. so
just how, how much it is sort of in

56
00:04:11,580 --> 00:04:17,430
between those two jets. this leaves us
with three background enhanced regions 

57
00:04:17,430 --> 00:04:24,300
C R A, C R B, and C R C, control regions, and then one
signal region that is enhanced, of course

58
00:04:24,300 --> 00:04:32,040
in our electroweak Z J J signal. and we can
use these three C R A B C regions to

59
00:04:32,160 --> 00:04:35,700
constrain and understand our strong
background, um and then use that

60
00:04:35,700 --> 00:04:38,970
understanding to constrain the modeling in
the signal enhanced region.

61
00:04:40,530 --> 00:04:44,700
so of course, I talked about four regions,
we're not just counting events um in each

62
00:04:44,700 --> 00:04:48,480
bin, there's still the observables we
want to perform the measurement in 

63
00:04:48,480 --> 00:04:55,170
that we actually have to consider. so here
I'm showing everything for the dijet

64
00:04:55,170 --> 00:04:59,790
invariant mass, so M J J is done in bin
number, but this is just the whole M J J

65
00:04:59,790 --> 00:05:00,000
spectrum

66
00:05:00,000 --> 00:05:01,170
we do the analysis in.

67
00:05:02,280 --> 00:05:05,970
on the left, these are the raw
distributions before the Monte Carlo is

68
00:05:05,970 --> 00:05:09,630
fit to the data. so you can see
objectively there is some some, some quite bad

69
00:05:09,630 --> 00:05:10,980
agreement there, we need to do something.

70
00:05:12,150 --> 00:05:17,520
but, uh, the control regions A B C have minimal
signal, lots of background, and we can use

71
00:05:17,520 --> 00:05:23,430
these to constrain the background in the
signal region. so in order to do this, we

72
00:05:23,430 --> 00:05:28,560
do a binned maximum likelihood fit. and this is
performed simultaneously in all four

73
00:05:28,560 --> 00:05:35,610
regions. now the, the strong Z J J background is
constrained with bin by bin weights

74
00:05:35,790 --> 00:05:40,620
separately at low and high centrality, but
it's linked between the number of gaps jets

75
00:05:40,620 --> 00:05:45,930
so like C R A and S R have the same bin by
bin weights and C R B and C R C have the same

76
00:05:45,930 --> 00:05:47,100
bin by bin weights.

77
00:05:48,thirty-six0 --> 00:05:52,500
um, then there is a linear function of the
variable being fit. so in this case, M J J

78
00:05:52,500 --> 00:05:57,330
that allows for a correction for any
residual dependence on the number of gap

79
00:05:57,330 --> 00:06:00,000
jets so where that bin by bin linking might
not work

80
00:06:00,000 --> 00:06:06,060
a hundred percent, um there's a linear uh correction that uh
helps fix that. and that's kept the same

81
00:06:06,060 --> 00:06:10,170
between the low and high centrality
regions applied in the signal region and

82
00:06:10,200 --> 00:06:17,790
C R C. so it sort of helps with C R C helps
then fix any any residual C R A to signal

83
00:06:17,790 --> 00:06:23,370
region extrapolation um in the signal region.
So finally, then you have the bin by bin

84
00:06:23,370 --> 00:06:26,670
electroweak signal strength, this is what
actually we want to extract the signal.

85
00:06:27,240 --> 00:06:32,070
um these are kept the same between all of the
regions. so you can see then in the post

86
00:06:32,070 --> 00:06:37,410
fit plot that when you actually perform
the maximum likelihood fit um it nicely

87
00:06:37,410 --> 00:06:40,470
converges to describe the observed data.
this is great. uh.

88
00:06:42,450 --> 00:06:46,380
now, I did talk about of course, that we
had these these three different Monte

89
00:06:46,380 --> 00:06:49,590
Carlo generators that all gave vastly
different predictions. and we didn't

90
00:06:49,590 --> 00:06:54,570
really know which one of them is correct a
priori. um and even after the fact we're not

91
00:06:54,570 --> 00:06:58,170
sure, you know, until we do some proper
unfolding which one better models the data

92
00:06:58,170 --> 00:07:00,000
so, un what we actually

93
00:07:00,000 --> 00:07:05,790
do, uh is do this fit with each of the
generators. so it's done done three times.

94
00:07:06,240 --> 00:07:11,190
and here I have the pre and postfit
plots. also with the next to leading order

95
00:07:11,190 --> 00:07:16,200
madgraph strong Z J J prediction. and, you know,
hopefully, you can take away from this

96
00:07:16,200 --> 00:07:21,630
that the extra or the fit also works with
the, one of the alternative generators.

97
00:07:23,460 --> 00:07:26,490
okay, so of course at the end of the day,
we don't just check these and ignore two

98
00:07:26,490 --> 00:07:32,340
of them, um we end up with three data points
for each bin, for each electroweak signal

99
00:07:32,340 --> 00:07:37,320
strength. so in each bin, the actual
extracted signal is taken as the midpoint

100
00:07:37,350 --> 00:07:41,220
of the envelope span. you can see my
little sketch on the right here um displaying

101
00:07:41,220 --> 00:07:45,330
what I mean. un so it's the midpoint of the
envelope spanned by the three strong data

102
00:07:45,330 --> 00:07:49,020
generators and then the spread is
associated with a systematic uncertainty

103
00:07:49,050 --> 00:07:50,670
or is an associated systematic
uncertainty, sorry.

104
00:07:52,260 --> 00:07:56,790
with these, you know signal strengths,
both the signal and the control regions

105
00:07:56,790 --> 00:08:00,000
are folded. uh so for instance,
electroweak production

106
00:08:00,000 --> 00:08:02,850
is only unfolded in the signal region
because there isn't any in the control

107
00:08:02,850 --> 00:08:07,500
regions. um but the inclusive Z J J
production, we unfold this in the signal

108
00:08:07,500 --> 00:08:11,460
region and in each of the four control
regions. uh and this is so we can basically uh

109
00:08:11,820 --> 00:08:16,260
use the data to the best of our abilities
and show the modeling in more than just

110
00:08:16,500 --> 00:08:21,000
the signal region for future Monte Carlo
studies and generator studies.

111
00:08:22,260 --> 00:08:27,420
so I've also consistently shown plots for
the dijet invariant mass, um but the analysis

112
00:08:27,420 --> 00:08:31,440
is also performed in the dijet rapidity
spread the dilepton transverse momentum

113
00:08:31,440 --> 00:08:36,450
and the delta phi between the higher and
lower rapidity jets in that order.

114
00:08:38,130 --> 00:08:43,380
ok. so beyond the measurement, we can also see
if the measurement shows any hints of new

115
00:08:43,380 --> 00:08:49,470
physics. to do this, um we use the dimension
six effective field theory, uh this is the

116
00:08:49,470 --> 00:08:55,560
SMEFT. and we look at both the interference
terms. that's the first bluey red term, I

117
00:08:55,560 --> 00:08:59,820
guess the second term in the equation at
the top and then the um

118
00:09:00,659 --> 00:09:06,539
quadratic term as well, which is uh the EFT
only quadratic term and that's in the

119
00:09:06,539 --> 00:09:13,469
middle there. now, the analysis tested two
CP even and two CP odd operators. and it,

120
00:09:13,469 --> 00:09:17,969
what's interesting to see in the table on the
right here is for each operator, there's

121
00:09:17,969 --> 00:09:20,429
two rows, the first one is using the
linear um

122
00:09:21,540 --> 00:09:25,680
EFT only and the second one was using the
quadratic. and you can see that it's

123
00:09:25,680 --> 00:09:29,580
really not a large effect. so the
quadratic terms having a minimal effect on

124
00:09:29,580 --> 00:09:32,190
the result of the whole and it's the
linear term that is driving our

125
00:09:32,190 --> 00:09:37,230
sensitivity. this sort of makes sense
because um the linear term is linear in the

126
00:09:37,230 --> 00:09:41,910
coupling whereas with a quadratic term,
you have this uh the coefficient squared. uh.

127
00:09:43,230 --> 00:09:46,830
now, something I should point out is the
reason we actually have sensitivity to

128
00:09:46,830 --> 00:09:53,130
these CP odd operators is by explicitly
checking um a parity odd observable, and

129
00:09:53,130 --> 00:09:56,340
in this case, the observable used is the
one I just mentioned on the previous slide

130
00:09:56,340 --> 00:09:59,670
the azimuthal angular separation between
the two leading jets

131
00:10:00,000 --> 00:10:04,590
calculated by subtracting the lower rapidity
jet from the higher rapidity jet, always

132
00:10:04,590 --> 00:10:08,790
in that order. so I have an example here
of what that variable looks like unfolded

133
00:10:08,790 --> 00:10:14,040
in signal region for the electroweak
signal. and then one other thing you might

134
00:10:14,040 --> 00:10:15,810
have noticed, in the observed

135
00:10:16,830 --> 00:10:22,380
sort of ninety-five percent confidence limit for
C tilde H W B doesn't actually include the

136
00:10:22,380 --> 00:10:27,750
value zero. the value zero, of course, is the
standard model with no EFT. so we'll talk

137
00:10:27,750 --> 00:10:29,670
a little bit about that on the next slide.

138
00:10:30,720 --> 00:10:33,840
so I first want to point out that we're
not talking about an exclusion of the

139
00:10:33,840 --> 00:10:37,410
standard model or discovery, right, this
is just a ninety-five percent confidence interval.

140
00:10:38,490 --> 00:10:42,660
um, but it's interesting to examine a little
further here we can see you know how we

141
00:10:42,660 --> 00:10:48,480
end up with this best fit value this one point two
and not zero. so on the left, we have the

142
00:10:48,480 --> 00:10:53,820
minimization of the negative log
likelihood. and on the right, we have the

143
00:10:53,850 --> 00:10:59,610
associated change in the EFT prediction as
the coupling changes. so you can see both

144
00:10:59,610 --> 00:10:59,910
the

145
00:11:00,000 --> 00:11:03,660
interference and the quadratic terms. and
then most of this contribution is coming

146
00:11:03,660 --> 00:11:07,740
from the interference term. that's the
dashed coral line as compared to the

147
00:11:07,740 --> 00:11:11,820
purple that adds in the quadratic term as
well. and then you can see basically,

148
00:11:12,060 --> 00:11:16,380
given the data points, how we end up with
this best fit value. and as we get farther

149
00:11:16,380 --> 00:11:20,070
and farther away from the Standard Model,
it doesn't really fit well, but then you

150
00:11:20,070 --> 00:11:24,870
get down to this minimum value of one point two. and
that distribution looks kind of a lot like

151
00:11:24,870 --> 00:11:29,910
the data. and then as you get closer and
closer and closer to zero, um, the data no

152
00:11:29,910 --> 00:11:32,070
longer looks quite a lot like the um uh a lot well sorry

153
00:11:33,900 --> 00:11:38,340
the prediction doesn't look a
lot look a lot like the data anymore. so

154
00:11:38,340 --> 00:11:43,830
this is a nice way of seeing how we end up
with a specific value. okay, so that's

155
00:11:43,830 --> 00:11:48,210
enough of a single boson on results. let's
add another that Z boson into the mix and

156
00:11:48,210 --> 00:11:52,710
study electroweak Z Z J J production. so here
we either have two bosons decaying

157
00:11:52,710 --> 00:11:58,080
leptonically, or one leptonic and one
invisible. again, there are two jets.

158
00:11:58,320 --> 00:11:59,970
still the same characteristic topology

159
00:12:00,000 --> 00:12:03,660
explicitly on different sides of the
detector now, with large rapidity

160
00:12:03,660 --> 00:12:05,760
separation and large invariant mass.

161
00:12:06,810 --> 00:12:10,860
the two channels um that I sort of just
mentioned on the previous slide, four

162
00:12:10,860 --> 00:12:15,810
lepton and L L nu nu have a slightly
different selection, but it's largely the

163
00:12:15,810 --> 00:12:21,600
same electroweak type phase space. uh the big
difference in these two channels is really

164
00:12:21,600 --> 00:12:27,600
the backgrounds. so with this L L nu nu
channel, you have a very large uh, high non Z Z J J

165
00:12:27,600 --> 00:12:30,900
background kind of getting in the
way of the analysis. It's not as clean as

166
00:12:30,900 --> 00:12:36,690
having two leptonic Zs. an inclusive
Z Z J J cross section that's inclusive

167
00:12:36,720 --> 00:12:41,910
of the QCD and the electroweak is
extracted in the signal region. um but then

168
00:12:41,910 --> 00:12:45,660
to extract the electroweak Z Z J J
component a boosted decision tree is used

169
00:12:45,660 --> 00:12:51,270
to discriminate between the electroweak and
the QCD components. um now, the BDT isn't just

170
00:12:51,270 --> 00:12:54,750
applied in the signal region we also apply
it in the control region. and this is

171
00:12:54,750 --> 00:12:59,250
better uh to allow us to better constrain the
QCD background right this thing, 

172
00:12:59,250 --> 00:13:00,000
miss-modeling that keeps

173
00:13:00,000 --> 00:13:00,960
popping up everywhere.

174
00:13:02,370 --> 00:13:07,200
and then a separate BDT is trained for the
L L nu nu signal region. so it's also a

175
00:13:07,200 --> 00:13:10,020
multivariate discriminant there on the x-axis,
but it's a different it's a

176
00:13:10,020 --> 00:13:14,220
different one. we don't assume those are
the same by any means. and you can see

177
00:13:14,220 --> 00:13:19,380
that when you fit these simultaneously
you get a combined expected

178
00:13:19,380 --> 00:13:24,390
significance of four point three sigma. of course, we
got a little lucky and we saw a five point five sigma

179
00:13:24,390 --> 00:13:28,380
observed significance and thus we do have
observation of electroweak Z Z J J

180
00:13:28,380 --> 00:13:32,820
production. um the nice neat thing here is
that the electroweak cross section here is

181
00:13:32,820 --> 00:13:35,520
point eight two femtobarns. and this is
actually one of the smallest cross

182
00:13:35,520 --> 00:13:37,470
sections that ATLAS has measured so far.

183
00:13:39,030 --> 00:13:43,440
okay, in my last minute, I'm going to
touch briefly on one last analysis, and

184
00:13:43,440 --> 00:13:47,880
that's electroweak production of a Z
boson and a photon. so again, we have our

185
00:13:47,880 --> 00:13:52,080
leptonic Z, but then we have a one
photon and we make we do our best to ensure

186
00:13:52,080 --> 00:13:55,170
this is not coming from final state
radiation from those leptons because that

187
00:13:55,170 --> 00:13:58,320
wouldn't be sensitive at all to the V B S
diagrams we're trying to really probe

188
00:13:58,320 --> 00:14:00,000
here. um the whole

189
00:14:00,179 --> 00:14:05,639
Z plus photon system is required to be
central sort of between the two jets. like

190
00:14:05,639 --> 00:14:09,629
we have the Z centrality for the
electroweak 
Z J J. and the two jets are

191
00:14:09,629 --> 00:14:12,719
required to have the large rapidity
separation and large invariant mass.

192
00:14:12,959 --> 00:14:17,819
additionally, here, um, there's a no b-jet
requirement in the events, we don't allow

193
00:14:17,819 --> 00:14:18,449
that to happen.

194
00:14:19,770 --> 00:14:23,010
so one thing to point out before going to
the result is that this is a partial

195
00:14:23,010 --> 00:14:26,100
run two result. so it's only using thirty-six inverse
femtobarns of data.

196
00:14:27,660 --> 00:14:32,040
and also that it's really important to
look specifically at all of these

197
00:14:32,040 --> 00:14:thirty-six,990
different electroweak productions, because
we want to ensure we have the full picture

198
00:14:36,990 --> 00:14:41,310
of all of the quartic couplings. so with
Z plus photon final state, we're

199
00:14:41,310 --> 00:14:46,830
uniquely sensitive to the W W Z photon
quartic coupling. whereas for instance, if

200
00:14:46,830 --> 00:14:53,040
you're looking at W Z, you'd have
contributions from um W gamma as your input

201
00:14:53,040 --> 00:14:57,150
vector bosons, or W Z it as your input
vector bosons. so if you saw some anomaly

202
00:14:57,150 --> 00:14:59,970
you wouldn't know necessarily which
diagram it was

203
00:15:00,000 --> 00:15:00,630
coming from.

204
00:15:02,129 --> 00:15:07,409
um now, in addition to this b-jet veto I I
mentioned on the previous slide, the

205
00:15:07,409 --> 00:15:11,819
events then that failed that, that have at
least one b-jet are used to constrain the

206
00:15:11,849 --> 00:15:15,509
ttbar plus photon background and make
sure that we're not susceptible to any

207
00:15:15,509 --> 00:15:20,339
miss-modelings there. so and like with the
Z Z J J um

208
00:15:21,480 --> 00:15:25,500
analysis, a BDT is used to discriminate
between the electroweak and the large QCD

209
00:15:25,500 --> 00:15:29,670
background, you see the BDT score
distribution on the bottom left here, the

210
00:15:29,670 --> 00:15:34,050
full distribution is fit simultaneously
with the b-jet distribution in the ttbar

211
00:15:34,050 --> 00:15:38,970
plus photon control region. and the
extracted observed and expected

212
00:15:38,970 --> 00:15:42,060
significance are both four point one sigma. so
unfortunately, no luckily, lucky

213
00:15:42,060 --> 00:15:46,560
statistical fluctuations here. but we
still have a hundred inverse femtobarns of data to add

214
00:15:47,190 --> 00:15:48,840
and and observe this process.

215
00:15:50,280 --> 00:15:54,990
okay, so there's three additional results
that I haven't talked about due to time

216
00:15:54,990 --> 00:15:57,300
which I'm sure I'm out of. uh, we have.

217
00:15:58,890 --> 00:15:59,910
um yeah, so the, sorry,

218
00:16:00,000 --> 00:16:05,160
same sign W W J J, um this is we have
observed this with six point five sigma.

219
00:16:06,240 --> 00:16:12,510
um, the W Z J J with a five point three sigma observed
significance and a semi-leptonic V V J J uh with

220
00:16:12,510 --> 00:16:18,150
two point seven observed significance. all of these are
thirty-six inverse femtobarn uh, run two uh analyses

221
00:16:18,150 --> 00:16:22,710
and there are details in the backup if you
would like to hear more. um so just in with

222
00:16:22,710 --> 00:16:27,030
that, I'll conclude we have new
differential cross-section measurement of

223
00:16:27,060 --> 00:16:31,440
electroweak Z J J production with strong
limits on new physics through an effective

224
00:16:31,440 --> 00:16:37,380
field theory interpretation. we have
observation of electroweak Z Z J J, and lots of

225
00:16:37,380 --> 00:16:43,320
other partial run two uh dataset analyses. and
many results with a full run two data set

226
00:16:43,320 --> 00:16:46,920
are still to come and there's a link
uh that's not a link at all, there's a list

227
00:16:47,250 --> 00:16:51,330
of the meeting ID and password of for the
poster session zoom chat if you would like

228
00:16:51,330 --> 00:16:53,670
to discuss anything. thanks.

229
00:16:56,100 --> 00:17:00,000
Thank you very much Heather. are there any
questions? yes, there is a question.

230
00:17:00,000 --> 00:17:01,380
from Josh

231
00:17:02,790 --> 00:17:06,210
please go ahead. you're unmuted when you can.

232
00:17:07,500 --> 00:17:10,290
thanks. uh yeah, so I had a question on slide
five.

233
00:17:11,969 --> 00:17:18,599
uh, so, in particular for for the two
predictions here for the for the QCD

234
00:17:18,869 --> 00:17:24,509
Z J J production, which are merged,
merged multi leg N L O. I mean, I

235
00:17:24,509 --> 00:17:28,889
wonder if you could comment on on ongoing
efforts to understand, you know, the

236
00:17:28,889 --> 00:17:33,389
uncertainties on those two predictions.
and whether there's some hope to one day

237
00:17:33,389 --> 00:17:38,009
replace this sort of two or three point
model comparison uncertainty with one

238
00:17:38,009 --> 00:17:42,029
which is driven by let's say, missing
higher orders in QCD match matching

239
00:17:42,029 --> 00:17:45,629
merging uncertainties or shower
resummation uncertainties, which could

240
00:17:45,629 --> 00:17:48,359
perhaps be a bit better defined from a
correlation standpoint.

241
00:17:49,560 --> 00:17:54,630
yeah, for sure. so this is this is part of
the reason there there's been so much effort to

242
00:17:54,960 --> 00:17:59,640
put out not just the electroweak signal
extraction, but also the inclusive

243
00:18:00,000 --> 00:18:05,700
differential cross sections for the
different um different generators um in not just

244
00:18:05,700 --> 00:18:10,080
the signal region, but also the different
control regions in the hopes that this

245
00:18:10,080 --> 00:18:14,070
information will be fed back to the
community. and we can use this to look at

246
00:18:14,070 --> 00:18:19,170
new predictions and then sort of run three
iterations might be um using whole new

247
00:18:20,310 --> 00:18:23,370
Monte Carlo predictions that fit the data
much better in the end.

248
00:18:26,190 --> 00:18:28,260
hope that answers your question. thanks.

249
00:18:31,050 --> 00:18:36,900
so we can take another question from
Kenneth. and then we will have to move

250
00:18:36,900 --> 00:18:39,810
on. Kenneth you're unmuted as you can speak.

251
00:18:41,790 --> 00:18:48,000
on slide seven, maybe you said I just
didn't understand but the the bin by bin

252
00:18:48,000 --> 00:18:55,140
weights that you apply to the V J J. these
are derived independently of the fit and

253
00:18:55,140 --> 00:18:59,940
basically applied before you do the
maximum likelyhood fit, is that right?

254
00:19:00,570 --> 00:19:01,830
no, they're part of the fit.

255
00:19:05,520 --> 00:19:12,240
I mean, so is there is there a parameter
that's free and the fit for each bin? uh no.

256
00:19:12,240 --> 00:19:16,560
so not not for each bin. so uh the fit is
done with three times the number of bins

257
00:19:16,560 --> 00:19:23,850
plus two parameters. uh and there's obviously
four times the number of bins um of actual

258
00:19:23,880 --> 00:19:25,080
data to fit.

259
00:19:26,220 --> 00:19:29,220
so is there I mean, are these weights
allow for

260
00:19:30,630 --> 00:19:35,010
essentially just the normalization to
change? or do they somehow allow uh

261
00:19:thirty-six,600 --> 00:19:40,440
a slope could be to be introduced to the
shape of the M J J?

262
00:19:41,760 --> 00:19:45,180
yeah, so they definitely do because
they're done bin by bin right, the weight

263
00:19:45,180 --> 00:19:47,820
in each bin that comes out of the fit can
be different,

264
00:19:48,990 --> 00:19:53,970
are they constrained in some way. sorry, so
they're linked between the the number of

265
00:19:54,510 --> 00:19:59,250
gap debts of of greater than or equal to one
and zero. so it's the same sort of in each

266
00:19:59,250 --> 00:19:59,820
column.

267
00:20:00,000 --> 00:20:05,790
so C R A and S R, the strong Z J J weights
are identical, the same in that low

268
00:20:05,790 --> 00:20:12,270
centrality region the same in the low in
the high centrality region. um but they are

269
00:20:12,360 --> 00:20:18,420
allowing for a bin by bin difference in
each of the columns basically of that

270
00:20:18,420 --> 00:20:19,050
little figure

271
00:20:20,610 --> 00:20:28,740
ok thanks they'll be a more wide description uh
in the paper. these are preliminary

272
00:20:28,740 --> 00:20:32,190
results but in the paper when it ends up
on arxiv, I encourage you to read it.

273
00:20:34,770 --> 00:20:39,150
there is one more question from Stefano
but we are a bit uh over time

274
00:20:39,150 --> 00:20:44,760
now. so maybe I mean, Heather you will be
available at the end on your own zoom room

275
00:20:44,850 --> 00:20:48,240
so probably Stefano you can ask a
question there. sorry.