1 00:00:00,000 --> 00:00:05,010 To the rare processes first innopolis and then we'll go to 2 00:00:07,470 --> 00:00:12,840 Yeah. So Hello, everybody, thank you for being here. And it's pleasure for me to 3 00:00:13,050 --> 00:00:19,380 give you an overview on the latest results from Atlas regarding roto pork production 4 00:00:19,410 --> 00:00:25,920 and yeah, just a bit of context. So, the huge amount of data that we collected 5 00:00:25,920 --> 00:00:31,260 during the run to lead us to test the center model tighter precision, but also 6 00:00:31,260 --> 00:00:38,250 to look so for two very, very tiny signal in basically exploring the right hand, the 7 00:00:38,250 --> 00:00:44,310 right bottom plot, corner of this plot, where top Quark production process, cross 8 00:00:44,310 --> 00:00:52,950 sections are here, and in this talk, I will cover three main analysis. I will fly 9 00:00:52,950 --> 00:00:58,740 over because Carrillo rarely has mentioned a lot of details around the top two gamma 10 00:00:58,740 --> 00:01:05,220 coupling and then CNC and then looking for standard model processes such as these 11 00:01:05,220 --> 00:01:10,560 that qm for tops starting from the first one, so the D three on my CNC coupling 12 00:01:10,560 --> 00:01:15,330 that's what that's because I've lost doesn't have really any latest results 13 00:01:15,750 --> 00:01:20,370 addressing the Standard Model process and therefore, we are looking at the CNC 14 00:01:20,640 --> 00:01:25,320 cabling looking for possible deviation that may come from beyond standard models 15 00:01:25,320 --> 00:01:31,950 scenarios. And the the analysis targets especially the production mode from the 16 00:01:31,980 --> 00:01:37,710 from the left hand diagram in the single top production but also include signals 17 00:01:37,710 --> 00:01:44,310 from from the decay mode and the goal is to estimate EFT operators next lingo to 18 00:01:44,340 --> 00:01:50,310 use D of course with the bosun WB and here right handed and left handed couplings 19 00:01:50,460 --> 00:01:56,160 been treated independently. And the data set has been used is the partial from 20 00:01:56,160 --> 00:02:04,980 2020 15 to 2017 of 2018. One inverse The barn and the selection tends to select 21 00:02:04,980 --> 00:02:10,350 production mode Indeed, we have single top topology in electronic case. So, one 22 00:02:10,350 --> 00:02:16,170 laptop one v jets are required and more than 330 gb of missing dt plus one photon. 23 00:02:16,560 --> 00:02:21,630 And indeed when we look at the production to the, to the decay modes of the 24 00:02:21,630 --> 00:02:25,080 coupling, we can see the need especially in the T q gamma, we have a factor of five 25 00:02:25,080 --> 00:02:31,230 enhancements. When we look at the background, the composition here I will 26 00:02:31,230 --> 00:02:35,850 focus here more on the experimental point of view we see more w plus Jetsons that 27 00:02:35,850 --> 00:02:40,020 was jets event has their useful backgrounds and here we address fitting 28 00:02:40,020 --> 00:02:44,730 generalization in dedicated control region that we can see here. And assuming 29 00:02:44,730 --> 00:02:49,980 generalization here the shapes instead from the Monte Carlo. So, we made them 30 00:02:49,980 --> 00:02:54,990 with the template method but we can see as well that's a gamma feature gamma faces a 31 00:02:54,990 --> 00:03:01,260 large contribution and this is estimated with a data driven decision. Do method we 32 00:03:01,260 --> 00:03:06,210 estimate the to gamma conversion efficiency in both domains in Monte Carlo 33 00:03:06,210 --> 00:03:10,740 and data and then we apply the scale factor to the simulation to have the 34 00:03:10,740 --> 00:03:16,020 prediction and the shape of the prediction is checked in dedicated validation region 35 00:03:16,020 --> 00:03:21,450 where we can see excellent modeling of of the state type of background to be more 36 00:03:21,450 --> 00:03:26,490 sensitive to different coupling also in your network have been trained and used as 37 00:03:26,490 --> 00:03:31,110 a discriminant variable in senior region and for different neural networks have 38 00:03:31,110 --> 00:03:34,410 been trained for all the possible for all the cabling. 39 00:03:35,700 --> 00:03:41,640 Consider so do you gamma, light if the light left and right handed and TC gamma 40 00:03:41,640 --> 00:03:48,210 left and right handed, going to the results very, very brief. We have a 41 00:03:48,210 --> 00:03:51,840 profile on your feet has been done in the country. In the seagull region. Here we 42 00:03:51,840 --> 00:03:58,020 can see an example of two gamma to a left handed scenario. And since we don't see 43 00:03:58,020 --> 00:04:03,120 any deviation from the Standard Model Background we set limits at the 95% 44 00:04:03,120 --> 00:04:08,700 confidence level improving the limit the already existing limits from two times up 45 00:04:08,700 --> 00:04:13,590 to one order of magnitude better. Setting branching limits on the branching ratio 46 00:04:13,590 --> 00:04:18,930 for example, up to the order of 10 to minus five. The limiting factor of the 47 00:04:18,930 --> 00:04:23,910 analysis of course is statistics but also some calibration such as Judge energy 48 00:04:23,910 --> 00:04:30,180 resolution, and gamma scale factor on PT and then modeling of GD bar and single top 49 00:04:30,210 --> 00:04:36,180 that comes into play gamma. fakes go instead into the Standard Model production 50 00:04:36,180 --> 00:04:42,930 mode is processes I will focus in on the tissue production. This is the production 51 00:04:42,930 --> 00:04:46,590 of a single top producer associate with the set bows on him expected to end 52 00:04:46,590 --> 00:04:52,920 predicted by the standard model with a cross section of 102 femto bars at level 53 00:04:52,920 --> 00:04:58,590 of 5% next year in order to CD in a diner Face Face. Official face to face we 54 00:04:58,590 --> 00:05:03,060 developed a muscle by both 30 gV that's very interesting because it's more 55 00:05:03,060 --> 00:05:07,530 sensitive to the desert on the desert cabling because of this very small next to 56 00:05:07,530 --> 00:05:12,900 the encoder using the correction. But it's also sensitive to W that kelvins through 57 00:05:13,200 --> 00:05:18,990 Fineman diagrams such as the one on the right. When we the we use the full brand 58 00:05:18,990 --> 00:05:25,920 new data set with 125 39 inverse femto barn and we selected one top decaying one 59 00:05:25,920 --> 00:05:30,690 top the single top decaying lip tonically clauses that decay lip Donnie Canada as 60 00:05:30,690 --> 00:05:36,000 well. So we expect free lectins in the final Saito's or relapse free leptons in 61 00:05:36,000 --> 00:05:41,040 the final states and in terms of number of jets and V jets. We expect one the Jets 62 00:05:41,040 --> 00:05:46,110 from the top and two and three v jets have been considered because we want to take 63 00:05:46,110 --> 00:05:48,240 into account the curlicue CD radiation as well. 64 00:05:50,040 --> 00:05:55,230 And but the while the B jets is expected to be central from the top, the extra 65 00:05:55,230 --> 00:06:01,140 jetties BK is expected to be so the untag jets as a B jet is expected to be forward. 66 00:06:01,140 --> 00:06:08,130 So, we extended the EDA and accelerate PDD acceptance of 2.4 4.5 the mass of the 67 00:06:08,220 --> 00:06:12,930 electronic mass opposite sign are expected are required to be in desert within that 68 00:06:12,930 --> 00:06:18,660 peak and in the seagull region the violet one, we want the pair to be E or mu mu of 69 00:06:18,660 --> 00:06:23,520 course from the Zed boson while we the same requirements we can set up the 70 00:06:23,520 --> 00:06:28,140 control region containing emu pairs to control this kind of background. But when 71 00:06:28,140 --> 00:06:32,040 we look at the background composition, the most important backgrounds are a TD set in 72 00:06:32,040 --> 00:06:39,150 Diebold simplicity flavor, and therefore, we set it up to call for different for 73 00:06:39,150 --> 00:06:43,770 control region to take into or to control them because that control region at higher 74 00:06:43,770 --> 00:06:49,770 jet multiplicity and to number to be jets, while the bozo dipole zone instead at the 75 00:06:49,770 --> 00:06:55,200 same requirements in terms of jets as the signal region where we bet zero BJ midgets 76 00:06:55,470 --> 00:06:59,790 and we can see that the modeling is quite good for both for both the backgrounds 77 00:07:00,000 --> 00:07:05,700 When we look at the DT bar, let's let's come to this a bit. Tiki Bar Zed cos jets 78 00:07:05,700 --> 00:07:10,620 and also dw can enter the signal region just via fake or non prompt lectins. 79 00:07:10,890 --> 00:07:15,150 studies on the truth level I've seen that most of the contribution come from leptons 80 00:07:15,150 --> 00:07:20,700 from non prompt, non or non prompt leptons from heavy flavor decays. And those are 81 00:07:20,700 --> 00:07:24,960 estimated with the replacement method and normalization of these predictions left 82 00:07:24,960 --> 00:07:30,450 free floating as a free floating parameters in the fit. Here we have, we 83 00:07:30,450 --> 00:07:34,530 can see the control region for the D bar where minimization only is fitted. 84 00:07:35,160 --> 00:07:41,490 Finally, the the discriminant variable need a single region in purple region is 85 00:07:41,520 --> 00:07:46,140 the output of a neural network that has been trained for the signal against the 86 00:07:46,140 --> 00:07:50,190 background, the total background and the most important variables for the training 87 00:07:50,220 --> 00:07:54,420 as we can expect are the mass from the untagged and objects because we expect to 88 00:07:54,420 --> 00:07:58,140 Central forward to apology. They thought they must have the top the reconstructed 89 00:07:58,140 --> 00:08:05,340 top and the suitability of the untagged jets. Finally, the results the fitted 90 00:08:05,370 --> 00:08:12,570 confusional cross region in the same space as the prediction is 94 7% as 997 femto 91 00:08:12,570 --> 00:08:18,090 barn with an accuracy of 13% dominated by statistical uncertainty. There is also 92 00:08:18,090 --> 00:08:24,360 while in agreement with Sundar model within one sigma and the significance over 93 00:08:24,360 --> 00:08:29,160 the background of the signal is above five sigma and we can see as well from the plot 94 00:08:29,190 --> 00:08:34,710 how the, the access is very compatible with these two features. So we can claim 95 00:08:34,710 --> 00:08:39,270 observation of the standard model is a cube reduction. The limiting factors here 96 00:08:39,600 --> 00:08:44,400 are data statistics and concerning the systematics already luminosity intervene 97 00:08:44,400 --> 00:08:50,010 and also and especially modeling of non prompt button backgrounds is quite 98 00:08:50,010 --> 00:08:57,120 important. Going to the latest to the latest analysis I wanted to talk about is 99 00:08:57,120 --> 00:09:01,980 a production of the fourth tops in this system. Model process with a cross section 100 00:09:01,980 --> 00:09:06,450 that is to edit is predicted the next leading order QC D plus electroweak of 101 00:09:07,020 --> 00:09:15,960 correction at 12pm to 120 percent level of accuracy. And this has been very strongly 102 00:09:15,960 --> 00:09:21,390 studied in the past because, yes are many videos under many scenarios predicts 103 00:09:21,450 --> 00:09:27,210 announcement of the cross section for this process, some example are here below. And 104 00:09:27,210 --> 00:09:32,100 the latest as the results combining for possible detection channel. So very 105 00:09:32,100 --> 00:09:37,770 complex combination of analysis using the partial data set of 20 1516 data sets 106 00:09:38,340 --> 00:09:43,500 observed a limit on the Four Tops cross section that is doubled than the expected 107 00:09:43,950 --> 00:09:49,890 so it's worth it for this reason to look at, look at it with the full data to run 108 00:09:49,890 --> 00:09:57,990 two data sets. And that's indeed what we have done looking at it using audio 139% 109 00:09:57,990 --> 00:10:04,680 are born and here we are is not Oh no, we don't consider all the possible detection 110 00:10:04,680 --> 00:10:09,840 channels coming from all the possibilities of the top decay modes, but we specialize 111 00:10:09,840 --> 00:10:14,160 on the same sign the electron in three or more lepton colored so move the electron 112 00:10:14,160 --> 00:10:22,110 channel corresponding to the 12% of the total of the total events. The Moreover, 113 00:10:22,110 --> 00:10:26,070 we were required for six or more jets because the other tops are expected to 114 00:10:26,100 --> 00:10:32,880 decay ironically, and two or more v jets not to decrease too much efficiencies on 115 00:10:32,880 --> 00:10:37,770 the signal. Also in cut on this HD, that is some of the PTO though the objectives 116 00:10:37,800 --> 00:10:42,300 in the final stages are required to be above 500 gV. And we can see from the plot 117 00:10:42,660 --> 00:10:47,940 that this really cut down most of the background while keep a good efficiency on 118 00:10:47,940 --> 00:10:48,450 the signal. 119 00:10:49,860 --> 00:10:53,550 Looking at the background composition here in this region, we can see that the most 120 00:10:53,550 --> 00:10:58,740 important are of course, tGwU that we'll come to this later. And here I want to 121 00:10:58,740 --> 00:11:03,930 focus a bit more on the Freaking nonprofit lepton backgrounds that is very complex 122 00:11:03,930 --> 00:11:07,770 because comes from many different resources such as charging 123 00:11:07,770 --> 00:11:12,780 misidentification, nonprofit leptons from heavy flavor the case, but also conversion 124 00:11:12,780 --> 00:11:19,140 of photons into material or low mass c plus d minus production. And those have 125 00:11:19,140 --> 00:11:23,760 been addressed differently depending on the sources for most of them. So, for the 126 00:11:23,760 --> 00:11:30,120 lessons from the heavy flavor and material conversion and low mass plus minus and 127 00:11:30,150 --> 00:11:35,160 backgrounds, the template method has been has been used. So, MonteCarlo shapes are 128 00:11:35,160 --> 00:11:38,850 assumed and normalization of the backgrounds has been fitted in indicated 129 00:11:38,850 --> 00:11:42,240 control region as we can see just normalization in the control region for 130 00:11:42,270 --> 00:11:48,300 behavioral flavor Latins have been fitted well for the material conversion and low 131 00:11:48,300 --> 00:11:54,000 mass beaker II plus the minus background. The mass of the E pair at primary vertex 132 00:11:54,000 --> 00:11:57,990 has been used because that's sensitive to different sources of the background in 133 00:11:57,990 --> 00:12:03,870 different bits of the distribution. We can see also in this distribution a greenish 134 00:12:04,080 --> 00:12:10,080 contribution is there and this comes from charge misidentification and this is 135 00:12:10,110 --> 00:12:15,960 estimated with a data driven method based on data as the two events. Basically we we 136 00:12:15,960 --> 00:12:21,450 estimated the charge misidentification efficiencies and then we apply that to 137 00:12:21,480 --> 00:12:26,790 opposite sign events to estimate same sign contributions. But the most challenging 138 00:12:26,790 --> 00:12:33,090 part is of course, the estimation or the TT w background in the in our sigma region 139 00:12:34,200 --> 00:12:39,540 and for that, we use the template method we can see a contour region here we can 140 00:12:39,540 --> 00:12:44,550 discuss later if you want about the definition of the country region and with 141 00:12:44,550 --> 00:12:48,900 the chosen good modeling Pacific with the with the background but with the with the 142 00:12:48,900 --> 00:12:54,630 data, but we also check the shapes and the modeling actually of the DW exploiting the 143 00:12:54,630 --> 00:12:59,850 VW charges inventory in invalidation region. So we have the difference between 144 00:12:59,850 --> 00:13:04,560 those In between the number of events we did total lip tonic charge positive and 145 00:13:04,560 --> 00:13:10,740 negative and what is the difference we obtained a very, very pure entity by 146 00:13:10,740 --> 00:13:17,370 entity w region very pure entity W and where we can check the modeling and here 147 00:13:17,400 --> 00:13:22,470 we looked at it and we saw a good a large discrepancy between data MonteCarlo 148 00:13:22,470 --> 00:13:27,330 especially we did with the DW associated with seven or eight jets. And from this 149 00:13:27,330 --> 00:13:35,730 discrepancy, we associated the uncertainty of 154 25 and 300% respectively however, 150 00:13:35,970 --> 00:13:40,140 is excessive we can see in the number of jets doesn't reflect in our BDD score that 151 00:13:40,140 --> 00:13:45,270 we'll come to in the next slides and that is the discriminant variable so and that's 152 00:13:45,270 --> 00:13:48,900 a good news indeed number of jets is not one of the most important variable 153 00:13:48,900 --> 00:13:55,140 instability another answer another sir uncertainty included in the in the fetus 154 00:13:55,140 --> 00:14:00,420 and btw with three and four v jets so associated with Devin farraj we If we 155 00:14:00,420 --> 00:14:06,360 favor jets, and we assume the 50% coming from studies often favor ideation in Tiki 156 00:14:06,360 --> 00:14:13,950 Bar, concerning the signal discrimination ability in our discriminating variable is 157 00:14:13,950 --> 00:14:16,980 the output of BDD that has been trained with 158 00:14:18,270 --> 00:14:23,400 for for the signal to separate with background and for stops in all the 159 00:14:23,400 --> 00:14:27,660 backgrounds have very different topologies. So, we can expect that many 160 00:14:27,660 --> 00:14:32,280 variables are actually important, but the most important one is is the besides game 161 00:14:32,280 --> 00:14:38,190 information. And we can see here the most important variable is the sum over the 162 00:14:38,190 --> 00:14:43,920 tagging score of digits in the events. We can see first a large separation between 163 00:14:43,920 --> 00:14:49,200 shapes that say between the Four Tops and all the background picking into different 164 00:14:49,200 --> 00:14:58,170 places, but also a good modeling of the of this variable. Finally, we we show the 165 00:14:58,170 --> 00:15:03,090 results. So we may To profile likelihood fit in country region, a single region 166 00:15:03,120 --> 00:15:08,250 Sima standardly in a single region, we can see our BTT score in the excellent data to 167 00:15:08,250 --> 00:15:12,990 monta to predict a Monte Carlo agreement after the fit and the normalization 168 00:15:13,200 --> 00:15:19,080 factors are have been fitted coherent as well in the profile it fit as a free 14 169 00:15:19,080 --> 00:15:25,500 parameters for all the backgrounds for you that uses the template method and we can 170 00:15:25,500 --> 00:15:30,270 see for example, their normalization for the material conversion is fitted 1.6 plus 171 00:15:30,270 --> 00:15:37,200 minus 0.5. But very much more importantly a Tw background is fitted to be 1.6 plus 172 00:15:37,230 --> 00:15:43,830 minus 0.3 times the standard error the the cross section, the Standard Model cross 173 00:15:43,830 --> 00:15:49,590 section. And this is indeed very consistent with what has been observed for 174 00:15:49,590 --> 00:15:53,850 example, in detachment 11 is similar, not exactly the same but similar face base. 175 00:15:54,660 --> 00:16:02,490 Finally, coming to the results on in terms of Four Tops Wi Fi signal sent is double 176 00:16:02,520 --> 00:16:07,440 the standard model, the Standard Model value but compatible with the sender model 177 00:16:07,440 --> 00:16:13,530 within 1.7 sigma and a cross section of 24 femto barn with a certainty of 27% 178 00:16:13,530 --> 00:16:18,600 approximately, the observed significance of the background is 4.3 standard 179 00:16:18,600 --> 00:16:25,380 deviation were 2.5 point 2.4 were expected. And when we look at the features 180 00:16:25,380 --> 00:16:29,550 of the data access with respect to the background, we see that the access is very 181 00:16:29,550 --> 00:16:33,750 compatible with four stops feature, features and signature. So, we can claim 182 00:16:33,750 --> 00:16:42,510 finally evidence for for this process. So, conclusions, we shown the latest results 183 00:16:42,540 --> 00:16:49,530 from others in terms of rate of work processes. If mC mC dq gamma coupling has 184 00:16:49,530 --> 00:16:54,090 been studied and best limits have been set concerning operators and renting ratio, 185 00:16:54,450 --> 00:16:59,520 the standard monitors to production has been observed and with a cross section 186 00:16:59,520 --> 00:17:04,380 compatible With the southern model within one sigma and the sun are more and we have 187 00:17:04,380 --> 00:17:08,220 also evidence of the standard model for stops core production with a significance 188 00:17:08,220 --> 00:17:13,260 of 4.3 standard deviation and the cross section compatible with the Standard Model 189 00:17:13,290 --> 00:17:14,610 within one point 70. 190 00:17:15,510 --> 00:17:16,320 And thank you. 191 00:17:17,430 --> 00:17:22,740 Okay, thank you Paulo for this very nice talk. Are there questions for Paulo? 192 00:17:22,800 --> 00:17:24,960 Please raise your hand if you have questions. 193 00:17:32,250 --> 00:17:38,370 Okay, I see for the moment no hands, but I do have a question. At the end, I have 194 00:17:38,580 --> 00:17:44,280 multiple questions, but I will start with the first one. You have shown the T gamma 195 00:17:44,280 --> 00:17:50,850 q in the context of SNC. Did I miss whether you mentioned Do you have also a 196 00:17:50,850 --> 00:17:54,060 cross section extraction from this or a separate analysis? 197 00:17:56,610 --> 00:18:01,800 You mean limits on the FC NC cross For the gamma or 198 00:18:02,310 --> 00:18:07,020 EQ gamma, the Standard Model, cross section mission. Do you have that? 199 00:18:07,350 --> 00:18:14,400 You know, we don't have it's I think it's ongoing effort as far as I know. 200 00:18:15,660 --> 00:18:17,130 There is no public results on it. 201 00:18:18,420 --> 00:18:20,790 Okay, let's see the standard model. Thank you. 202 00:18:21,990 --> 00:18:28,740 I see still no hands, but I saw I have, I will post the second question about the TT 203 00:18:28,740 --> 00:18:34,710 w control region. It looks pretty nice and pure. Can you tell a bit more about the 204 00:18:34,710 --> 00:18:35,550 control region? 205 00:18:36,870 --> 00:18:39,570 How do you select is just the number of jets? 206 00:18:41,370 --> 00:18:44,820 Do you mean the validation or the control region for the Tw backroom? 207 00:18:45,330 --> 00:18:51,030 Actually both looks very pure the maybe the middle plot it's rather 208 00:18:51,389 --> 00:18:56,549 okay. Yeah, this is this is used to check basically the the modeling of the DW. 209 00:18:57,209 --> 00:19:03,539 Apart from the control region that is more closer to To our signal signal region by 210 00:19:03,539 --> 00:19:09,539 the validation region it's made by subtracting the number event of events and 211 00:19:09,539 --> 00:19:13,649 it has the total charge of the electrons of the leptons. So the sum of over over 212 00:19:13,649 --> 00:19:20,189 all the charge the electronic charge positive minus all the the events with the 213 00:19:20,189 --> 00:19:26,819 total electronic charge that is negative. And it's basically discard so the charge 214 00:19:26,819 --> 00:19:32,639 symmetric processes and you get a very purity GW set. This is where we can check 215 00:19:32,639 --> 00:19:33,359 the modeling Yeah. 216 00:19:33,780 --> 00:19:36,030 Okay. So this is sort of background subtracted. 217 00:19:37,470 --> 00:19:38,460 Innocence sense. 218 00:19:40,140 --> 00:19:47,700 Yeah, it's it's basically it comes it comes by by the construction that all the 219 00:19:47,700 --> 00:19:51,630 charge symmetric that is basically very, very similar. 220 00:19:52,980 --> 00:19:56,370 It has basically symmetrical in charge gets gets 221 00:19:57,000 --> 00:20:02,250 discarded from from how the region has been. has been constructed. So yeah. 222 00:20:04,259 --> 00:20:05,399 It's kind of 223 00:20:06,660 --> 00:20:10,110 a way to to to reduce the background in the in the region. Yeah. 224 00:20:10,710 --> 00:20:18,300 Okay. Thank you. Um, so are there other questions for follow?