AtlasPublicTopicHeader.png

Public b-Jet Trigger Plots for Collision Data

Introduction

Approved plots that can be shown by ATLAS speakers at conferences and similar events. Please do not add figures on your own. Contact the responsible project leader in case of questions and/or suggestions. Follow the guidelines on the trigger public results page.

2022 @ 13.6 TeV

Performance of Run 3 HLT b-tagging with fast tracking

Receiver operating characteristic (ROC) curves for the various b-jet taggers implemented in the High Level Trigger (HLT), obtained on a $t\bar{t}$ MC sample. A fast b-tagging pre-selection (fastDIPS) [ATL-PHYS-PUB-2020-014] is applied on topocluster-based (EMTopo) jets calibrated at the electromagnetic scale, with a pT requirement of 20 GeV. Inputs for the pre-selection are tracks reconstructed with “fast” trigger algorithms in regions of interest [Eur. Phys. J. C 82 (2022) 206] centered on the jet axis, without any primary vertex information. DL1d [Eur. Phys. J. C 79 (2019) 970] is the b-jet identification algorithm, running on precision tracks and jet level quantities after primary vertexing, on which the HLT decision is taken. The GN1 is a novel version of the latter based on a graph neural network architecture [ATL-PHYS-PUB-2022-027]. A special case is the one of fastDIPS on particle flow jets (PFlow jets), which is a tagger dedicated for the trigger level analysis (TLA), attaching the b-jet information to the TLA objects without applying any trigger selection. Statistical uncertainties for each ROC curve are represented with shaded regions around the curves. The purple vertical dashed lines represent the most common working points used for b-tagging. The bottom panel displays the ratio of all the ROC curves with respect to DL1d performance.

[png]
[pdf]

Fast b-tagging trigger selection performance and impact on the expected 𝐻𝐻 → 𝑏𝑏̄𝑏𝑏̄ acceptance. The rejection factors are estimated from Run-2 enhanced bias data sample [ATL-DAQ-PUB-2016-002], while the trigger acceptances are calculated using simulation. The L1 selection considered everywhere is common and requires three central (|𝜂| < 2.5) jets with pT > 15 GeV, out of which the leading one is required to have pT > 45 GeV. The High Level Trigger (HLT) selection used for comparison (‘selection (𝐻𝐻 → 𝑏𝑏̄𝑏𝑏̄)’ in the first column of the table) contains requirements on four jets (pT > 80 GeV, 55 GeV, 28 GeV, 20 GeV) as well as two b-tagged jets reconstructed with a high-level algorithm and selected with 77% efficiency per b-tagged jet [Eur. Phys. J. C 79 (2019) 970]. The fast b-tagging preselection (‘presel.’ in the first column of the table) applies a DIPS-based [ATL-PHYS-PUB-2020-014] selection on fast tracking [Eur. Phys. J. C 82 (2022) 206] in jet regions of interest, without any primary vertex information. Two b-jets with pT > 20 GeV must be identified by the DIPS-based preselection. Compared to a trigger without the preselection but the same selection (𝐻𝐻 → 𝑏𝑏̄𝑏𝑏̄), the preselection has a negligible impact in the 𝐻𝐻 → 𝑏𝑏̄𝑏𝑏̄ trigger acceptance (2-4%), but is necessary to reduce the rate at which the HLT event-wide tracking is run to fit within the online CPU constraints to run at the LHC peak luminosity of the Run-3 data taking (2 × 1034 cm−2s−1).

[png]
[pdf]

Conditional efficiencies of the fast b-tagging step (fastDIPS, DIPS-based b-jet tagger used as a pre-selection) [ATL-PHYS-PUB-2020-014] on the final High Level Trigger (HLT) tagger (DL1d) [Eur. Phys. J. C 79 (2019) 970], $P( fastDIPS \mid DL1d )$, for b-jets. The efficiencies are computed at jet level, having matched jets within a radius of 0.3 in the η-ɸ plane. The purple vertical dashed lines represent the most common working points used for b-tagging. The study was carried out on a $t\bar{t}$ MC sample.

[png]
[pdf]

CPU time required to reconstruct tracks evaluated in $t\bar{t}$ MC - with trigger dedicated “fast” algorithms [Eur. Phys. J. C 82 (2022) 206] - in a region of interest (RoI) centered on the jet axis, normalized to the time employed for running the same tracking algorithms over the full inner detector acceptance (Full Scan Tracking). The ratio is shown for different RoI sizes (η and ɸ half-widths, varied equally) and track transverse momenta requirement. The final working point used for Run-3 is: RoI half-widths=0.3, and track pT≥ 1 GeV.

[png]
[pdf]

Receiver operating characteristic (ROC) curves for DIPS-based b-jet taggers [ATL-PHYS-PUB-2020-014] using fast tracking in jet regions of interest (RoIs) [Eur. Phys. J. C 82 (2022) 206], obtained on a $t\bar{t}$ MC sample. The figure shows the b-tagging performance depending on the minimum pT of the tracks used for b-tagging. A minimum pT cut of 1 GeV performs slightly worse than 0.5 GeV, while a minimum of 1.5 GeV results in a large b-tagging degradation. Statistical uncertainties for each ROC curve are represented with shaded regions around the curves.

[png]
[pdf]

Receiver operating characteristic (ROC) curves for DIPS-based b-jet taggers [ATL-PHYS-PUB-2020-014] using fast tracking in jet regions of interest (RoIs) [Eur. Phys. J. C 82 (2022) 206], obtained on a $t\bar{t}$ MC sample. The impact of changing the RoI size (η and ɸ half-widths, varied equally) on the b-tagging performance is shown; only a small loss of performance is observed for a RoI size of 0.3 compared to 0.5, but a substantial loss is seen for smaller RoIs. Statistical uncertainties for each ROC curve are represented with shaded regions around the curves.

[png]
[pdf]

Distributions of the longitudinal component of the tracks impact parameter (IP), associated to light jets, obtained for the precision (offline-like) tracking and full detector acceptance (Full Scan) “fast” tracking steps in b-jet trigger chains [ Eur. Phys. J. C 82 (2022) 206, Eur. Phys. J. C 81 (2021) 087] , in a $t\bar{t}$ MC sample.

[png]
[pdf]

Distributions of the transverse component of the tracks impact parameter (IP), associated to light jets, obtained by the various tracking steps in b-jet trigger chains [ Eur. Phys. J. C 82 (2022) 206, Eur. Phys. J. C 81 (2021) 087] in a $t\bar{t}$ MC sample. The transverse IP for “fast” track reconstruction in jet regions of interest (RoI tracks) is determined with respect to the position of the beam-spot, while in the other cases it is determined with respect to the hard-scatter primary vertex, reconstructed via Full Scan Fast Tracks.

[png]
[pdf]

Expected Run 3 Trigger Efficiencies for $HH \to 4b$ and Comparison to Run 2

Trigger efficiencies in a simulated event sample of Standard Model Higgs boson pair production at a centre-of-mass energy of 13.6 TeV, with decays to four b-quarks, versus the generated Higgs pair mass $m_{HH}$. Three triggers providing events to the main physics stream are displayed: an asymmetric four-jet selection with three b-tags at an 85% efficiency operating point; a symmetric selection on four jets with two b-tags at 60% efficiency; and a three-jet selection requiring one high transverse momentum jet and two softer jets, of which two must be b-tagged at 70% efficiency. These triggers record events at 30 Hz, 25 Hz and 13 Hz respectively ($L_{inst}$ = 2e34 cm-2 s-1). Vertical error bars indicate the statistical uncertainty. Also shown is the efficiency for selecting $HH \to 4b$ events in the main physics stream, determined from the union of events selected by these three triggers. Assuming the Standard Model $m_{HH}$ distribution, the total signal efficiency of this combination is 53%.

[png]
[pdf]

Trigger efficiencies in a simulated event sample of Standard Model Higgs boson pair production at a centre-of-mass energy of 13.6 TeV, with decays to four b-quarks, versus the generated Higgs pair mass $m_{HH}$. The efficiency of a combination of three triggers providing events to the main physics stream is plotted, as described in Figure 1. Vertical error bars indicate the statistical uncertainty. The contribution of an asymmetric four-jet trigger with two b-tags at 77% efficiency is also shown. This trigger records events at a rate of 150 Hz ($L_{inst}$ = 2e34 cm-2 s-1) to a delayed physics stream, which is reconstructed using opportunistic computing resources. Combining the delayed stream trigger with the three main stream triggers improves the differential trigger efficiency in $m_{HH}$ by up to 20%. The signal efficiency integrated over all mass bins for the Standard Model assumption is increased from 53% for the main stream triggers alone to 59%.

[png]
[pdf]

Trigger efficiencies in a simulated event sample of Standard Model Higgs boson pair production, with decays to four b-quarks, versus the generated Higgs pair mass $m_{HH}$. The efficiencies of two Run 3 trigger combinations, as evaluated in simulation at 13.6 TeV with Run 3 detector conditions, are shown: three triggers contribute to the main physics stream combination (Figure 1), which is augmented by a higher rate trigger recording events to a delayed stream (Figure 2). Vertical error bars indicate the statistical uncertainty. In Run 2, events were recorded only to the main physics stream with the symmetric four-jet trigger and the asymmetric three-jet trigger, whose collective efficiency on a 13 TeV event sample, simulated with 2018 conditions from Run 2 is shown. The asymmetric four-jet trigger selections used in Run 3 increase the total efficiency for selecting Standard Model $HH \to 4b$ events from 41% to 59%.

[png]
[pdf]

Expected Performance of the ATLAS b-jet Trigger Algorithms for the Start of LHC Run 3

Distributions of the b-jet discriminant score for the DL1d algorithm $D = \log ( p_b / [ p_c f_c + p_u  (1-f_c) ] )$, where $p_b$, $p_c$ and $p_u$ are the b-, c-, and light-jet probability outputs of the DL1d algorithm, and $f_c$ is the effective c-jet fraction. The DL1d algorithm uses the DL1 architecture and the DIPS low-level tagger. Both were trained with simulated $t\bar{t}$ and $Z'$ events. The algorithms are evaluated on HLT Particle Flow jets from a $t\bar{t}$ sample. The b-jet (blue), c-jet (orange) and light-flavor jet (green) distributions are shown for $f_c = 0.018$. Red vertical lines indicate the 85%, 77%, 70% and 60% b-jet efficiency working points. Jets are labeled as b, c, or light using simulation-based hadron matching as described in previous ATLAS results.

[png]
[pdf]

Light-flavor jet rejection as a function of b-jet efficiency of the DIPS (green dashed) and DL1d (blue dash-dotted) algorithms in comparison to the benchmark DL1r algorithm (brown solid), evaluated on HLT Particle Flow jets in a $t\bar{t}$ sample. Jets are labeled as b, c, or light using simulation-based hadron matching as described in previous ATLAS results. The 60%, 70%, 77% and 85% b-jet efficiency working points are indicated by vertical red lines. In all cases the discriminant is given by $D = \log ( p_b / [ p_c f_c + p_u  (1-f_c) ] )$, where $p_b$, $p_c$ and $p_u$ are the b-, c-, and light-jet probability outputs of the DL1d algorithm, and $f_c$ is the effective c-jet fraction.

[png]
[pdf]

c-jet rejection as a function of b-jet efficiency of the DIPS (green dashed) and DL1d (blue dash-dotted) algorithms in comparison to the benchmark DL1r algorithm (brown solid), evaluated on HLT Particle Flow jets in a $t\bar{t}$ sample. Jets are labeled as b, c, or light using simulation-based hadron matching as described in previous ATLAS results. The 60%, 70%, 77% and 85% b-jet efficiency working points are indicated by vertical red lines. In all cases the discriminant is given by $D = \log ( p_b / [ p_c f_c + p_u  (1-f_c) ] )$, where $p_b$, $p_c$ and $p_u$ are the b-, c-, and light-jet probability outputs of the DL1d algorithm, and $f_c$ is the effective c-jet fraction.

[png]
[pdf]

Distributions of the DL1d discriminant evaluated on Particle Flow HLT jets from $t\bar{t}$ and multijet samples. The 77% b-jet efficiency working point is indicated by a vertical red line. The discriminant distributions are shown separately for jets of various flavors for $f_c = 0.018$, with significant overlap observed in the discriminant distributions for jets containing exactly one b-hadron (b-jets) and those containing more than one (bb-jets). The discriminant is given by $D = \log ( p_b / [ p_c f_c + p_u  (1-f_c) ] )$, where $p_b$, $p_c$ and $p_u$ are the b-, c-, and light-jet probability outputs of the DL1d algorithm, and $f_c$ is the effective c-jet fraction. Jets are labeled as bb, b, c, or light using simulation-based hadron matching as described in previous ATLAS results, where bb requires two matched b-hadrons.

[png]
[pdf]

Distributions of discriminant scores for the DL1dbb algorithm, trained on a mixture of $t\bar{t}$ and multijet events. The discriminant is defined by $D = \log ( p_b / p_{bb} )$, where $p_b$ and $p_{bb}$ are the b-jet and bb-jet outputs for the DL1dbb algorithm which uses the DL1 architecture and the DIPS low-level tagger to discriminant between b and bb-jets. The tagger is evaluated on Particle Flow HLT jets from $t\bar{t}$ and multijet samples. The 77% b-jet efficiency working point is indicated by a vertical red line. Jets are labeled as bb, b, c, or light using simulation-based hadron matching as described in previous ATLAS results, where bb requires two matched b-hadrons.

[png]
[pdf]

bb-jet rejection as a function of b-jet efficiency of the DL1d (blue) and DL1dbb (orange) algorithms, evaluated on HLT Particle Flow jets from $t\bar{t}$ and multijet samples. Jets are labeled as bb, b, c, or light using simulation-based hadron matching as described in previous ATLAS results, where bb requires two matched b-hadrons. The discriminant is defined by $D = \log ( p_b / p_{bb} )$, where $p_b$ and $p_{bb}$ are the b-jet and bb-jet outputs for the DL1dbb algorithm. The DL1dbb performance is conditional on a DL1d discriminant cut at the 85% working point. The 60%, 70%, 77% and 85% b-jet efficiency working points are indicated by vertical red lines.

[png]
[pdf]

Expected trigger rates as a function of b-tagging efficiency using the DL1d algorithm when requiring at least four Particle Flow jets, three of which are requirement to be above the b-tagging threshold. Jets are labeled as b, c, or light using simulation-based hadron matching as described in previous ATLAS results. Rates are estimated with Run 2 Enhanced Bias Data, and the b-jet efficiencies are estimated for $t\bar{t}$ and multijet samples using trigger Particle Flow jets. The relative errors on trigger rates are between 10% and 21%.

[png]
[pdf]

Expected trigger rates as a function of b-tagging efficiency while requiring both a variable DL1dbb threshold and a fixed DL1d threshold at the 85% working point. Jets are labeled as bb, b, c, or light using simulation-based hadron matching as described in previous ATLAS results, where bb requires two matched b-hadrons. Here at least four Particle Flow jets are required, three of which should be above the b-tagging threshold. Rates are estimated with Run 2 Enhanced Bias Data, and the b-jet efficiencies are estimated for $t\bar{t}$ and multijet samples using trigger Particle Flow jets. The relative errors on the trigger rates are between 10% and 17%.

[png]
[pdf]

2018 @ 13 TeV

Performance of the ATLAS muon-jet trigger in Run2 heavy-ion collision data ATL-COM-DAQ-2020-058 (July 2020)

The conditional muon-jet trigger efficiency measured in data as a function of offline jet $p_T$. The conditional muon-jet trigger efficiency is defined as the number of offline muon-jets in events for which a given muon-jet trigger fired divided by the total number of offline muon-jets that fired a single-muon trigger with the same $p_T$ threshold as the muon-jet trigger used in the numerator. The muon-jet triggers shown require a muon with $p_T$ > 4 GeV geometrically matched ($\Delta R < 0.5$) to a trigger jet with $E_T$ thresholds of 40, 50, and 60 GeV. The muon-jet trigger efficiency is estimated with respect to the single-muon trigger requiring $p_T$($\mu$) > 4 GeV. Events are also required to have an offline muon with $p_T$ > 12 GeV, and both offline muon and jet are required to have $|\eta| < 1.05$. The performance is shown in the barrel region of the detector. The centrality of a collision is assessed on an event-by-event basis using the $E_T$ deposited in the forward calorimeters. In this analysis, central collisions are defined as those in the 0-40% centrality interval where the contribution from the underlying event effects is the largest.

[png]
[eps]

The conditional muon-jet trigger efficiency measured in data as a function of offline jet $p_T$. The conditional muon-jet trigger efficiency is defined as the number of offline muon-jets in events for which a given muon-jet trigger fired divided by the total number of offline muon-jets that fired a single-muon trigger with the same $p_T$ threshold as the muon-jet trigger used in the numerator. The muon-jet triggers shown require a muon with $p_T$ > 4 GeV geometrically matched ($\Delta R < 0.5$) to a trigger jet with $E_T$ thresholds of 40, 50, and 60 GeV. The muon-jet trigger efficiency is estimated with respect to the single-muon trigger requiring $p_T$($\mu$) > 4 GeV. Events are also required to have an offline muon with $p_T$ > 12 GeV, and both offline muon and jet are required to have $|\eta| < 1.05$. The performance is shown in the barrel region of the detector. The centrality of a collision is assessed on an event-by-event basis using the $E_T$ deposited in the forward calorimeters. In this analysis, central collisions are defined as those in the 40-80% centrality interval where the contribution from the underlying event effects is the largest.

[png]
[eps]

The conditional muon-jet trigger efficiency measured in data as a function of offline jet $p_T$. The conditional muon-jet trigger efficiency is defined as the number of offline muon-jets in events for which a given muon-jet trigger fired divided by the total number of offline muon-jets that fired a single-muon trigger with the same $p_T$ threshold as the muon-jet trigger used in the numerator. The muon-jet triggers shown require a muon with $p_T$ > 4 GeV geometrically matched ($\Delta R < 0.5$) to a trigger jet with $E_T$ thresholds of 40, 50, and 60 GeV. The muon-jet trigger efficiency is estimated with respect to the single-muon trigger requiring $p_T$($\mu$) > 4 GeV. Events are also required to have an offline muon with $p_T$ > 12 GeV, and both offline muon and jet are required to have $1.05 < |\eta| < 2.40$. The performance is shown in the endcap regions of the detector. The centrality of a collision is assessed on an event-by-event basis using the $E_T$ deposited in the forward calorimeters. In this analysis, central collisions are defined as those in the 0-40% centrality interval where the contribution from the underlying event effects is the largest.

[png]
[eps]

The conditional muon-jet trigger efficiency measured in data as a function of offline jet $p_T$. The conditional muon-jet trigger efficiency is defined as the number of offline muon-jets in events for which a given muon-jet trigger fired divided by the total number of offline muon-jets that fired a single-muon trigger with the same $p_T$ threshold as the muon-jet trigger used in the numerator. The muon-jet triggers shown require a muon with $p_T$ > 4 GeV geometrically matched ($\Delta R < 0.5$) to a trigger jet with $E_T$ thresholds of 40, 50, and 60 GeV. The muon-jet trigger efficiency is estimated with respect to the single-muon trigger requiring $p_T$($\mu$) > 4 GeV. Events are also required to have an offline muon with $p_T$ > 12 GeV, and both offline muon and jet are required to have $1.05 < |\eta| < 2.40$. The performance is shown in the endcap regions of the detector. The centrality of a collision is assessed on an event-by-event basis using the $E_T$ deposited in the forward calorimeters. In this analysis, central collisions are defined as those in the 40-80% centrality interval where the contribution from the underlying event effects is the largest.

[png]
[eps]

The two-dimensional absolute trigger efficiency, measured in data, of the muon-jet trigger that requires a muon with $p_T$ > 4 GeV geometrically matched ($\Delta R < 0.5$) to a trigger jet with $E_T$>40 GeV, as a function of the offline muon and jet $p_T$. The absolute muon-jet trigger efficiency is defined as the product of the conditional trigger efficiency and the single-muon trigger efficiency. The measurements are performed integrated across the full collision centrality range (0-80\%) for the barrel region. The centrality of a collision is assessed on an event-by-event basis using the $E_T$ deposited in the forward calorimeters.

[png]
[eps]

The two-dimensional absolute trigger efficiency, measured in data, of the muon-jet trigger that requires a muon with $p_T$ > 4 GeV geometrically matched ($\Delta R < 0.5$) to a trigger jet with $E_T$>40 GeV, as a function of the offline muon and jet $p_T$. The absolute muon-jet trigger efficiency is defined as the product of the conditional trigger efficiency and the single-muon trigger efficiency. The measurements are performed integrated across the full collision centrality range (0-80\%) for the endcap region. The centrality of a collision is assessed on an event-by-event basis using the $E_T$ deposited in the forward calorimeters.

[png]
[eps]

Calibration of the ATLAS Trigger b-jet identification efficiency ATL-COM-DAQ-2020-035 (May, 2020)

Scale factors (data/MC) of online $b$-jet identification efficiency ($SF_b^{trig}$) measurement for 60\% single-cut working point of the online MV2 tagger as a function of offline jet $p_T$. Vertical error bars include data statistical uncertainties only, while the green bands correspond to the sum in quadrature of statistical and systematic uncertainties. The measurement is performed using a data sample enriched with $t\bar{t}$ events with $e\mu$ opposite charge leptons, targeting a phase space with a high purity of $b$-jets. Events are selected with performance triggers requiring a lepton and at least two jets on which $b$-jet identification algorithm is applied without taking decision. Further, events are required to have exactly 2 offline jets matched to online jets with a geometrical requirement. Events passing the selection are classified according to jet $p_T$ and the output of the online $b$-jet identification algorithm, then the $b$-jet identification efficiency is extracted from data with a likelihood fit method. [\texttt{arXiv:1907.05120 hep-ex}]

[png]
[eps]

Scale factors (data/MC) for the conditional online $b$-jet identification efficiency ($SF_b^{trig|tag}$) measurement for 60\% single-cut working point of the online MV2 tagger as a function of jet $p_T$ after applying the 60\% single-cut working point on the DL1r offline tagger. Vertical error bars include data statistical uncertainties only, while the green bands correspond to the sum in quadrature of statistical and systematic uncertainties. The measurement is performed using a data sample enriched with $t\bar{t}$ events with $e\mu$ opposite charge leptons, targeting a phase space with a high purity of $b$-jets. Events are selected with performance triggers requiring a lepton and at least two jets on which the $b$-jet identification algorithm is applied without taking a decision. Further, events are required to have exactly 2 offline jets matched to online jets with a geometrical requirement. Events passing the selection are classified according to jet $p_T$ and the output of the online $b$-jet identification algorithm. The $b$-jet identification efficiency is extracted from data with a likelihood fit method. [\texttt{arXiv:1907.05120 hep-ex}]

[png]
[eps]

Decomposition of the $t\bar{t}$ modeling uncertainty on the online $b$-jet identification efficiency measurement of the online MV2 tagger for 60\% single-cut working point as a function of jet $p_T$. The total $t\bar{t}$ modeling uncertainty is shown as a green band while the various components are shown as points. The modeling uncertainty on $t\bar{t}$ were computed based on the difference between the “nominal” $t\bar{t}$ MC sample and other $t\bar{t}$ MC samples generated with alternative parton shower and hadronization model, additional initial-state radiation (ISR) and final-state radiation (FSR), and alternative parton density function (PDF). The measurement is performed using a data sample enriched with $t\bar{t}$ events with $e\mu$ opposite charge leptons, targeting a phase space with a high purity of $b$-jets. Events are selected with performance triggers requiring a lepton and at least two jets on which the $b$-jet identification algorithm is applied without taking a decision. Further, events are required to have exactly 2 offline jets matched to online jets with a geometrical requirement. Events passing the selection are classified according to jet $p_T$ and the output of the online $b$-jet identification algorithm. The $b$-jet identification efficiency is extracted from data with a likelihood fit method. [\texttt{arXiv:1907.05120 hep-ex}]

[png]
[eps]

Decomposition of the $t\bar{t}$ modeling uncertainty on the online $b$-jet identification efficiency measurement of the online MV2 tagger for 60\% single-cut working point as a function of jet $p_T$. Output discriminant of the online MV2 algorithm for the $p_T$-leading jet in events passing the selection applied for the online $b$-jet identification efficiency measurement, targeting a sample enriched with $t\bar{t}$ dilepton events, thus with a high purity of $b$-jets. Events are required to have exactly 2 offline jets matched to online jets with a geometrical requirement. Simulated events are split according to the physics process. Simulated background processes considered are $V+$jets, Dibosons and Single top. %from light blue to dark blue: $t\bar{t}$, Single top, $Z+$jets, Diboson, $W+$jets. No significant difference is observed in the behavior of the $p_T$ subleading jet. The lower panels show the data-to-simulation ratio as well as the fraction of $t\bar{t}$ events among the simulated events. The green bands in the first ratio panel show MC statistical uncertainty and the total uncertainty. [\texttt{arXiv:1907.05120 hep-ex}]

[png]
[eps]

Decomposition of the $t\bar{t}$ modeling uncertainty on the online $b$-jet identification efficiency measurement of the online MV2 tagger for 60\% single-cut working point as a function of jet $p_T$. Output discriminant of the online MV2 algorithm for the $p_T$-leading jet in events passing the selection applied for the trigger $b$-jet identification efficiency measurement, targeting a sample enriched with $t\bar{t}$ dilepton events, thus with a high purity of $b$-jets. Events are required to have exactly 2 offline jets matched to online jets with a geometrical requirement. Simulated events are classified according to the flavor composition of the two jets, but only the leading jet is shown: the two colors on top are for truth $b$-jets while the other two colors on the bottom are for light jets. No significant difference is observed in the behavior of the $p_T$ subleading jet. The lower panels show the data-to-simulation ratio as well as the fraction of bb events among the simulated events. The green bands in the first ratio panel show MC statistical uncertainty and the total uncertainty. [\texttt{arXiv:1907.05120 hep-ex}]

[png]
[eps]

Trigger b-tagging hybrid performances ATL-COM-DAQ-2019-018 (May, 2019)

The ATLAS \textit{b}-jet trigger uses a BDT algorithm to separate \textit{b}-jets from light and \textit{c}-jet backgrounds. The BDT algorithm used for \textit{b}-jet triggers in 2018 data taking has been trained on a $t\bar{t}$ Monte Carlo simulation. The same BDT algorithm has been trained on an alternative (Hybrid) training sample consisting of a mixture of $t\bar{t}$ and $Z'$ Monte Carlo samples, in the same way the BDT algorithm for offline jets is trained. Performance of \textit{b}-tagging algorithms (measured using $t\bar{t}$ Monte Carlo events), with respect to the true flavour of jets, is shown in terms of light-jet rejection as a function of \textit{b}-jet efficiency. Expected performance of \textit{b}-tagging algorithm (MV2c10) for \textit{b}-jet triggers in 2018 data-taking (blue solid line) is compared to the same \textit{b}-tagging algorithm trained on the Hybrid training sample (red solid line). [png]
[pdf], [eps]

The ATLAS \textit{b}-jet trigger uses a BDT algorithm to separate \textit{b}-jets from light and \textit{c}-jet backgrounds. The BDT algorithm used for \textit{b}-jet triggers in 2018 data taking has been trained on a $t\bar{t}$ Monte Carlo simulation. The same BDT algorithm has been trained on an alternative (Hybrid) training sample consisting of a mixture of $t\bar{t}$ and $Z'$ Monte Carlo samples, in the same way the BDT algorithm for offline jets is trained. Performance of \textit{b}-tagging algorithms (measured using $t\bar{t}$ Monte Carlo events), with respect to the true flavour of jets, is shown in terms of \textit{c}-jet rejection as a function of \textit{b}-jet efficiency. Expected performance of \textit{b}-tagging algorithm (MV2c10) for \textit{b}-jet triggers in 2018 data-taking (blue solid line) is compared to the same \textit{b}-tagging algorithm trained on the Hybrid training sample (red solid line). [png]
[pdf], [eps]

The ATLAS \textit{b}-jet trigger uses a BDT algorithm to separate \textit{b}-jets from light and \textit{c}-jet backgrounds. The BDT algorithm used for \textit{b}-jet triggers in 2018 data taking has been trained on a $t\bar{t}$ Monte Carlo simulation. The same BDT algorithm has been trained on an alternative (Hybrid) training sample consisting of a mixture of $t\bar{t}$ and $Z'$ Monte Carlo samples, in the same way the BDT algorithm for offline jets is trained. Performance of \textit{b}-tagging algorithms (measured using $Z'$ Monte Carlo events), with respect to the true flavour of jets, is shown in terms of light-jet rejection as a function of \textit{b}-jet efficiency. Expected performance of \textit{b}-tagging algorithm (MV2c10) for \textit{b}-jet triggers in 2018 data-taking (blue solid line) is compared to the same \textit{b}-tagging algorithm trained on the Hybrid training sample (red solid line). [png]
[pdf], [eps]

The ATLAS \textit{b}-jet trigger uses a BDT algorithm to separate \textit{b}-jets from light and \textit{c}-jet backgrounds. The BDT algorithm used for \textit{b}-jet triggers in 2018 data taking has been trained on a $t\bar{t}$ Monte Carlo simulation. The same BDT algorithm has been trained on an alternative (Hybrid) training sample consisting of a mixture of $t\bar{t}$ and $Z'$ Monte Carlo samples, in the same way the BDT algorithm for offline jets is trained. Performance of \textit{b}-tagging algorithms (measured using $Z'$ Monte Carlo events), with respect to the true flavour of jets, is shown in terms of \textit{c}-jet rejection as a function of \textit{b}-jet efficiency. Expected performance of \textit{b}-tagging algorithm (MV2c10) for \textit{b}-jet triggers in 2018 data-taking (blue solid line) is compared to the same \textit{b}-tagging algorithm trained on the Hybrid training sample (red solid line). [png]
[pdf], [eps]

Performance of the ATLAS b-jet trigger in 2018 data at high pile-up ATL-COM-DAQ-2018-124 (August 2, 2018)

The $b$-jet trigger efficiency with respect to the offline $b$-tagging algorithm (\texttt{MV2c10}) at the 70\% efficiency operating point for various online efficiency operating points vs. the mean number of interactions per crossing. The relative $b$-jet trigger efficiency is measured in high purity di-lepton $t\bar{t}$ events collected in the 2018 data-set using dedicated single-lepton$+$jets triggers, which are unbiased with respect to the online $b$-tagging. Operating point efficiencies are defined using offline-reconstructed jets from an unbiased sample of Monte Carlo simulated $t\bar{t}$ events, where jets are labeled according to their hadron content. The online operating points were defined to have roughly the quoted efficiency using online-reconstructed jets matched to offline-reconstructed jets. The same $b$-tagging algorithm is used offline and online and retuned for the online environment. Statistical uncertainties only are shown.
[pdf], [eps]

The $b$-jet trigger efficiency with respect to the offline $b$-tagging algorithm (\texttt{MV2c10}) at the 70\% efficiency operating point for various online efficiency operating points vs. the mean number of interactions per crossing. The relative $b$-jet trigger efficiency is measured in high purity di-lepton $t\bar{t}$ events collected in the 2018 data-set using dedicated single-lepton$+$jets triggers, which are unbiased with respect to the online $b$-tagging. Operating point efficiencies are defined using offline-reconstructed jets from an unbiased sample of Monte Carlo simulated $t\bar{t}$ events, where jets are labeled according to their hadron content. The online operating points were defined to have roughly the quoted efficiency using online-reconstructed jets matched to offline-reconstructed jets. The same $b$-tagging algorithm is used offline and online and retuned for the online environment. Statistical uncertainties only are shown.
[pdf], [eps]

2017 @ 13 TeV

Measurement of the ATLAS b-jet trigger efficiency in 2017 data ATL-COM-DAQ-2019-077 (July 2019)

The b-jet trigger efficiency at the 40\% online operating point (OP) with respect to offline b-tagging at the 70\% offline operating point as a function of offline jet $p_{\mathrm T}$. This set of online and offline OPs is one of the least used combinations in analyses. The b-jet trigger efficiency is measured using a high b-jet purity di-lepton $t\bar t$ selection in the 2017 data. Events in data are required to have passed data quality selection for the b-jet trigger. Offline jets are required to match a trigger-level jet, using $\Delta R < 0.4$ matching scheme. Total uncertainties on the measured efficiencies are due to data statistics and systematic uncertainties to account for non b-jet contamination, namely purity of b-jets and efficiency of b-jet trigger for non b-jets. [png]
[eps]

The b-jet trigger efficiency at the 60\% online operating point (OP) with respect to offline b-tagging at the 70\% offline operating point as a function of offline jet $p_{\mathrm T}$. This set of online and offline OPs is one of the most used combinations in analyses. The b-jet trigger efficiency is measured using a high b-jet purity di-lepton $t\bar t$ selection in the 2017 data. Events in data are required to have passed data quality selection for the b-jet trigger. Offline jets are required to match a trigger-level jet, using $\Delta R < 0.4$ matching scheme. Total uncertainties on the measured efficiencies are due to data statistics and systematic uncertainties to account for non b-jet contamination, namely purity of b-jets and efficiency of b-jet trigger for non b-jets. [png]
[eps]

The b-jet trigger efficiency at the 70\% online operating point (OP) with respect to offline b-tagging at the 70\% offline operating point as a function of offline jet $p_{\mathrm T}$. This set of online and offline OPs is one of the most used combinations in analyses. The b-jet trigger efficiency is measured using a high b-jet purity di-lepton $t\bar t$ selection in the 2017 data. Events in data are required to have passed data quality selection for the b-jet trigger. Offline jets are required to match a trigger-level jet, using $\Delta R < 0.4$ matching scheme. Total uncertainties on the measured efficiencies are due to data statistics and systematic uncertainties to account for non b-jet contamination, namely purity of b-jets and efficiency of b-jet trigger for non b-jets. [png]
[eps]

Performance of the ATLAS b-jet trigger in 2017 data at high pile-up ATL-COM-DAQ-2017-182 (November 28, 2017)

The $b$-jet trigger efficiency with respect to the offline $b$-tagging algorithm (\texttt{MV2c10}) at the 70\% efficiency operating point for various online efficiency operating points vs. the mean number of interactions per crossing. The relative $b$-jet trigger efficiency is measured in high purity di-lepton $t\bar{t}$ events collected in the 2017 data-set using dedicated single-lepton$+$jets triggers, which are unbiased with respect to the online $b$-tagging. The online operating points were defined to have roughly the quoted efficiency for $b$-jets in an unbiased sample of Monte Carlo simulated $t\bar{t}$ events. The uncertainty bars shown only represent statistical uncertainties.
[pdf], [eps]

The fraction of trigger jets with Global Sequential Calibration (GSC)-corrected $E_T > 55~\textrm{GeV}$ that pass the online $b$-tagging algorithm (\texttt{MV2c10}) at various online efficiency operating points vs. the mean number of interactions per crossing. The pass fraction is measured in a subset of the 2017 data-set with events containing at least one jet with GSC-corrected $E_T > 55~\textrm{GeV}$. The trigger used is unbiased with respect to the online $b$-tagging. The online operating points were defined to have roughly the quoted efficiency for $b$-jets in an unbiased sample of Monte Carlo simulated $t\bar{t}$ events. The uncertainty bars shown only represent statistical uncertainties.
[pdf], [eps]

Expected performance of the ATLAS b-jet trigger in 2017

* http://cds.cern.ch/record/2271945

The ATLAS b-jet trigger uses a BDT algorithm to separate b-jets from light and c-jet backgrounds. The BDT algorithm is re-optimized to improve b-tagging performance. Performance of b-tagging algorithms (measured using ttbar Monte Carlo events) is shown in terms of light-jet rejection as a function of b-jet efficiency. Expected performance of b-tagging algorithm (MV2c10) for b-jet triggers in 2017 data-taking (green solid line) is compared to b-tagging algorithms used for b-jet triggers in 2016 (MV2c20) and 2015 (IP3D+SV1) data taking. Performance of b-tagging algorithm MV2c10 for offline jets is shown in purple dotted curve.
[pdf], [eps]

The ATLAS b-jet trigger uses a BDT algorithm to separate b-jets from light and c-jet backgrounds. The BDT algorithm is re-optimized to improve b-tagging performance. Performance of b-tagging algorithms (measured using ttbar Monte Carlo events) is shown in terms of c-jet rejection as a function of b-jet efficiency. Expected performance of b-tagging algorithm (MV2c10) for b-jet triggers in 2017 data-taking (green solid line) is compared to b-tagging algorithms used for b-jet triggers in 2016 (MV2c20). Performance of b-tagging algorithm MV2c10 for offline jets is shown in purple dotted curve.
[pdf], [eps]

2016 @ 13 TeV

Measurement of the ATLAS b-jet trigger efficiency in 2016 data ATL-COM-DAQ-2017-009 (February 8, 2017)

The $b$-jet trigger efficiency at the 60\% online operating point with respect to offline $b$-tagging at the 70\% offline operating point as a function of offline jet-$p_{T}$. The $b$-jet trigger efficiency is measured using a high $b$-jet purity di-lepton $t\bar{t}$ selection in the 2016 data-set and $t\bar{t}$ simulation is used to extrapolate for jet-$p_{T} >$ 240 GeV. Events in data are required to have passed data quality selection for the $b$-jet trigger. Offline jets are required to match a trigger-level jet, where matching is done exclusively using $\Delta R$ matching. Systematics account for non $b$-jet contamination and the simulation-based extrapolation to high jet-$p_{T}$.
[pdf], [eps]

2015 @ 13 TeV

Impact Parameter Significance: Online vs Offline

Transverse impact parameter significance for tracks associated to light-flavour (black) and b-quark (red) jets measured on a sample of simulated ttbar events. The solid lines show the distribution for the offline tracks. The points show the corresponding distribution for tracks used in the b-jet trigger. The impact parameter significance is defined as the impact parameter divided by the associated uncertainty. The impact parameters are signed such that track displacements in the direction of the jet have positive values, while tracks with displacements opposite of the jet direction are negative. The impact parameter significance is used to identify jets originating from decays of $b$-quarks.
[pdf], [eps]

Longitudinal impact parameter significance for tracks associated to light-flavour (black) and b-quark (red) jets measured on a sample of simulated ttbar events. The solid lines show the distribution for the offline tracks. The points show the corresponding distribution for tracks used in the b-jet trigger. The impact parameter significance is defined as the impact parameter divided by the associated uncertainty. The impact parameters are signed such that track displacements in the direction of the jet have positive values, while tracks with displacements opposite of the jet direction are negative. The impact parameter significance is used to identify jets originating from decays of $b$-quarks.
[pdf], [eps]

Trigger rates of the ATLAS b-jet trigger in Run 2

* https://cds.cern.ch/record/2051479

Output rates of ATLAS multi-b-jet triggers as a function of the instantaneous luminosity during 2015 proton-proton data taking with a center-of-mass energy of 13 TeV and an LHC bunch-crossing interval of 50 ns. These triggers consist of hardware-based first-level (L1) and software-based high-level trigger (HLT) selections. At L1, one jet with a transverse energy (ET) > 100 GeV is required. At the HLT, either one b-tagged jet with ET> 175 GeV, 225 GeV or 300 GeV, or two b-tagged jets with E_T thresholds of 175 and 60 GeV are requested. The operating points, `bloose' and `bmedium' are optimized such that the efficiency for selecting b-jets is 79% and 72%, respectively.
[pdf], [eps]

Expected performance of the ATLAS b-jet trigger in Run 2

* http://cds.cern.ch/record/2032280

The ATLAS b-jet trigger will use the same tools as are used for offline reconstruction to select heavy flavour jets during Run 2. This new tagger (MV2c20) uses a BDT to separate b-jets from light and c-jet backgrounds. The expected online performance in terms of light-jet rejection of the MV2c20 tagger (solid black line) is shown together with the expected performance of the IP3D+SV1 tagger in Run 2 (dashed blue line) and the actual performance of the IP3D+SV1 tagger that was achieved during Run 1 (red stars). The improvement in IP3D+SV1 for Run 2 compared to Run 1 is due to several effects: inclusion of the Insertable B-Layer (IBL), improved tracking performance and tagger re-tuning for Run 2 conditions. The b-jet trigger performance in Run 1 was also affected by overly conservative pixel cluster errors used by the tracking that caused a mis-measurement of the track d0 uncertainty. The tuning is performed on ttbar simulation with √s = 13 TeV. Jets used are required to have pT > 55 GeV and |η| < 2.5. The points illustrating the Run 1 performance were derived using ttbar simulation with √s =8 TeV.
[pdf], [eps]

The ATLAS b-jet trigger will use the same tools as are used for offline reconstruction to select heavy flavour jets during Run 2. This new tagger (MV2c20) uses a BDT to separate b-jets from light and c-jet backgrounds. The expected online performance in terms of c-jet rejection of the MV2c20 tagger (solid black line) is shown together with the expected performance of the IP3D+SV1 tagger in Run 2 (dashed blue line). The tuning is performed on ttbar simulation with √s = 13 TeV. Jets used are required to have pT > 55 GeV and |η| < 2.5.
[pdf], [eps]

2012 @ 8 TeV

L2 and EF Trigger b-tagging weights in jets containing D* mesons for 2012 data ATL-COM-PHYS-2014-631 (June 26, 2014)

* https://cds.cern.ch/record/1710286

Comparison between the L2 Trigger $b$-tagging weight distribution on a background-subtracted jets, containing $D^{*+}$ mesons, data sample with the corresponding simulated PYTHIA sample. The statistical uncertainty of the simulation is below a few percent and not shown in this figure. The data sample is collected in 2012 using single jet triggers, and the transverse momentum of the selected jets is required to be above 20 GeV. The beauty to charm jet fraction in the simulation is constrained to the value obtained by a pseudo-proper time fit on data for the $D^0$ mesons arising from the $D^{*+}$ decays. More complete description of the method is available in \href{https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2012-039/}{ATLAS-CONF-2012-039}.
[pdf], [eps]

Comparison between the EF Trigger $b$-tagging weight distribution on a background-subtracted jets, containing $D^{*+}$ mesons, data sample with the corresponding simulated PYTHIA sample. The statistical uncertainty of the simulation is below a few percent and not shown in this figure. The data sample is collected in 2012 using single jet triggers, and the transverse momentum of the selected jets is required to be above 20 GeV. The beauty to charm jet fraction in the simulation is constrained to the value obtained by a pseudo-proper time fit on data for the $D^0$ mesons arising from the $D^{*+}$ decays. More complete description of the method is available in \href{https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2012-039/}{ATLAS-CONF-2012-039}.
[pdf], [eps]

b-jet tagger data to simulation comparison using 8 TeV data ATL-COM-DAQ-2013-018 (May 9, 2013)

* https://cds.cern.ch/record/1541260

Jet weight distribution for the tagger based on the combination of the impact parameter significance and the secondary vertex likelihood-based taggers, calculated from prescaled Level 2 tracks in Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Jet weight distribution for the tagger based on the combination of the impact parameter significance and the secondary vertex likelihood-based taggers, calculated from prescaled Event Filter tracks in Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Jet weight distribution for the tagger based on the combination of the impact parameter significance and the secondary vertex likelihood-based taggers, calculated from prescaled Level 2 tracks in Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Jet weight distribution for the tagger based on the combination of the impact parameter significance and the secondary vertex likelihood-based taggers, calculated from Event Filter tracks in Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Jet weight distribution for the likelihood-ratio tagger based on the longitudinal and transverse impact parameter significance [\href{https://cds.cern.ch/record/1369219/files/ATLAS-CONF-2011-102.pdf}{ATLAS-CONF-2011-102}] of prescaled Level 2 tracks in Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Jet weight distribution for the likelihood-ratio tagger based on the longitudinal and transverse impact parameter significance [\href{https://cds.cern.ch/record/1369219/files/ATLAS-CONF-2011-102.pdf}{ATLAS-CONF-2011-102}] of prescaled Event Filter tracks in Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Jet weight distribution for the likelihood-ratio tagger based on the longitudinal and transverse impact parameter significance [\href{https://cds.cern.ch/record/1369219/files/ATLAS-CONF-2011-102.pdf}{ATLAS-CONF-2011-102}] of prescaled Level 2 tracks in Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Jet weight distribution for the likelihood-ratio tagger based on the longitudinal and transverse impact parameter significance [\href{https://cds.cern.ch/record/1369219/files/ATLAS-CONF-2011-102.pdf}{ATLAS-CONF-2011-102}] of prescaled Event Filter tracks in Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Ratio of energy sum of quality tracks associated with the prescaled Level 2 jets' secondary vertex and the energy sum of all quality tracks in the jet for Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Ratio of energy sum of quality tracks associated with the prescaled Event Filter jets' secondary vertex and the energy sum of all quality tracks in the jet for Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Ratio of energy sum of quality tracks associated with the prescaled Level 2 jets' secondary vertex and the energy sum of all quality tracks in the jet for Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Ratio of energy sum of quality tracks associated with the prescaled Event Filter jets' secondary vertex and the energy sum of all quality tracks in the jet for Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Track multiplicity at the prescaled Level 2 jets' secondary vertex for Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Track multiplicity at the prescaled Event Filter jets' secondary vertex for Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Track multiplicity at the prescaled Level 2 jets' secondary vertex for Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Track multiplicity at the prescaled Event Filter jets' secondary vertex for Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Invariant mass at the prescaled Level 2 jets' secondary vertex for Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Invariant mass at the prescaled Event Filter jets' secondary vertex for Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$. Only statistical errors are shown.
[pdf], [eps]

Invariant mass at the prescaled Level 2 jets' secondary vertex for Level 2 jets with $p_T > 50$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

Invariant mass at the Event Filter jets' secondary vertex for Event Filter jets with $p_T > 55$ GeV and $|\eta| < 2.5$ matched to a muon with $p_T > 4$ GeV to enhance the b-jet component. Only statistical errors are shown.
[pdf], [eps]

2011 @ 7 TeV

b-Jet triggering in ATLAS ATL-COM-DAQ-2012-050 (May 16, 2012)

* http://cdsweb.cern.ch/record/1446650

Resolution for the primary vertex Z position estimate as a function of the number of online tracks at L2 and EF. The z coordinate of the primary vertex (PV) is calculated by histogramming the $z_0$ of all selected tracks in the event and using a sliding window algorithm to select the largest local maximum.
[pdf], [eps]

b-Jet Trigger rate plots with 2011 data ATL-COM-DAQ-2011-067 (August 25, 2011)

* http://cdsweb.cern.ch/record/1375816

Trigger rate for 1b/4j topology. LVL1, LVL2 and EF rate of a b-jet trigger requiring at least four jets in the event and at least one b-tagged jet. The jet thresholds correspond to 10, 25 and 30 GeV at LVL1, LVL2 and EF with energies measured at the electromagnetic scale. The b-jet requirement is applied at LVL2 and EF and is tuned to give 70\% efficiency on a b-tagged jet sample using top MC simulation.
[pdf], [eps]

Trigger rate for 2b/4j topology. LVL1, LVL2 and EF rate of a b-jet trigger requiring at least four jets in the event and at least two b-tagged jet. The jet thresholds correspond to 10, 25 and 30 GeV at LVL1, LVL2 and EF with energies measured at the electromagnetic scale. The b-jet requirement is applied at LVL2 and EF is tuned to give 70\% efficiency on a b-tagged jet sample using top MC simulation.
[pdf], [eps]

Trigger rate for 2b/2j topology. LVL1, LVL2 and EF rate of a b-jet trigger requiring at least two jets in the event and at least one b-tagged jet. The jet thresholds are asymmetric and correspond to 50(10), 70(25) and 75(30) GeV at LVL1, LVL2 and EF for the leading (second leading) jet with energies measured at the electromagnetic scale. The b-jet requirement is applied at LVL2 and EF and is tuned to give 70\% efficiency on a b-tagged jet sample using top MC simulation.
[pdf], [eps]

b-Jet Trigger plots with 2011 data ATL-COM-DAQ-2011-052 (July 19, 2011)

* http://cdsweb.cern.ch/record/1364854

Signed transverse impact parameter significance of reconstructed tracks at the Event Filter level. Tracks are reconstructed starting from a low pT jet identified by the Level 1 and are requested to fulfill online b-tagging criteria.
[pdf], [eps]

Track multiplicity at the Event Filter level. Tracks are reconstructed starting from a low pT jet identified by the Level 1 and are requested to fulfill online b-tagging criteria.
[pdf], [eps]

Track transverse momentum at the Event Filter level. Tracks are reconstructed starting from a low pT jet identified by the Level 1 and are requested to fulfill online b-tagging criteria.
[pdf], [eps]

Offline JetProb distribution in data and simulation and the same distribution in data when a b-jet requirement is added at the second trigger level. The three working points correspond to 90%, 70% and 50% b-tagging efficiency on a b-tagged jet sample using top MC simulation.
[pdf], [eps]

Offline JetProb distribution in data and simulation and the same distribution in data when a b-jet requirement is added at the trigger level (both Level 2 and Event Filter). The three working points correspond to 90%, 70% and 50% b-tagging efficiency on a b-tagged jet sample using top MC simulation.
[pdf], [eps]

muon-in-jet trigger distribution as a function of the jet transverse momentum in offline muon-in-jet candidates. This class of triggers includes a geometrical matching (Delta R < 0.4) between the muon and the jet objects. Several triggers with various jet thresholds select events in order to collect a sample of muon-in-jet candidates in the entire jet transverse momentum spectrum.
[pdf], [eps]

2010 @ 7 TeV

b-Jet Trigger dependence on the Beam Spot Determination in 2010 Data ATL-COM-DAQ-2010-197 (November 4, 2010)

* http://cdsweb.cern.ch/record/1304878

Probability for the LVL2 tracks to originate from primary vertex before and after the HLT beam spot update. The probability for a track to originate from the primary vertex is computed on the signed transverse impact parameter significance of each selected track in the jet RoI using the resolution function for prompt tracks. A jet probability is then defined by considering the probabilities of all the selected tracks. By construction, it ensures a uniform distribution between 0 and 1 for tracks originating from the primary vertex while tracks from displaced B decays tend to accumulate near 0. This plot demonstrates the usefulness of having the beam spot parameters accessible within the HLT farm in real time since the b-jet triggers rely on the beam spot parameters to compute the track transverse impact parameter significance. Before the update the beam spot position w.r.t. the real position is 45um (170um) away in x (y) while after the update the position is 5um (10um).
[pdf], [eps], [jpg]

FTK Public Results

Some plots related to the B-jet trigger are included in the FTK public results: https://twiki.cern.ch/twiki/bin/view/AtlasPublic/FTKPublicResults

https://twiki.cern.ch/twiki/bin/view/AtlasPublic/FTKPublicResults


Major updates:
-- AndreaCoccaro - 06-Jun-2011

Responsible: AndreaCoccaro
Subject: public

  • b-jet trigger calibration in 2018:
    onlinesfPrel.png

  • b-jet trigger calibration in 2018:
    combsfPrel.png

  • b-jet trigger calibration in 2018:
    ttsystPrel.png

  • b-jet trigger calibration in 2018:
    sampFillPrel.png

  • b-jet trigger calibration in 2018:
    flavFillPrel.png

  • Muon-Jet Trigger Performance in Run2 Data:
    Eff_2dim_abs_BAR.png

  • Muon-Jet Trigger Performance in Run2 Data:
    Eff_2dim_abs_EC.png

  • Muon-Jet Trigger Performance in Run2 Data:
    Eff_central_barrel.png

  • Muon-Jet Trigger Performance in Run2 Data:
    Eff_central_endcaps.png

  • Muon-Jet Trigger Performance in Run2 Data:
    Eff_peripheral_barrel.png

  • Muon-Jet Trigger Performance in Run2 Data:
    Eff_peripheral_endcaps.png

Topic attachments
I Attachment History Action Size Date Who Comment
Unknown file formateps BJetTrigg_efficiency_vs_mu-2018.eps r1 manage 99.4 K 2018-08-02 - 14:10 StephenSekula  
PDFpdf BJetTrigg_efficiency_vs_mu-2018.pdf r1 manage 23.2 K 2018-08-02 - 14:09 StephenSekula  
PNGpng BJetTrigg_efficiency_vs_mu-2018.png r1 manage 114.2 K 2018-08-02 - 14:09 StephenSekula  
Unknown file formateps BJetTrigg_efficiency_vs_mu-yzoom-2018.eps r1 manage 103.2 K 2018-08-02 - 14:10 StephenSekula  
PDFpdf BJetTrigg_efficiency_vs_mu-yzoom-2018.pdf r1 manage 24.0 K 2018-08-02 - 14:10 StephenSekula  
PNGpng BJetTrigg_efficiency_vs_mu-yzoom-2018.png r1 manage 120.5 K 2018-08-02 - 14:10 StephenSekula  
Unknown file formateps BJetTrigg_efficiency_vs_mu.eps r1 manage 21.4 K 2017-12-11 - 22:00 MatthewFeickert The plots from ATL-COM-DAQ-2017-182
PDFpdf BJetTrigg_efficiency_vs_mu.pdf r1 manage 20.0 K 2017-12-11 - 22:00 MatthewFeickert The plots from ATL-COM-DAQ-2017-182
PNGpng BJetTrigg_efficiency_vs_mu.png r1 manage 24.6 K 2017-12-11 - 22:00 MatthewFeickert The plots from ATL-COM-DAQ-2017-182
Unknown file formateps BJetTrigg_pass_fraction_vs_mu.eps r1 manage 22.9 K 2017-12-11 - 22:00 MatthewFeickert The plots from ATL-COM-DAQ-2017-182
PDFpdf BJetTrigg_pass_fraction_vs_mu.pdf r1 manage 21.3 K 2017-12-11 - 22:00 MatthewFeickert The plots from ATL-COM-DAQ-2017-182
PNGpng BJetTrigg_pass_fraction_vs_mu.png r1 manage 24.0 K 2017-12-11 - 22:00 MatthewFeickert The plots from ATL-COM-DAQ-2017-182
PDFpdf CPU_scan_RoI-tracking.pdf r1 manage 185.3 K 2022-11-07 - 17:52 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng CPU_scan_RoI-tracking.png r1 manage 67.6 K 2022-11-07 - 17:52 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng DL1d_bb_flavour_DL1d20211216_ICHEP_299-1.png r1 manage 87.8 K 2022-06-29 - 15:08 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PDFpdf DL1d_bb_flavour_DL1d20211216_ICHEP_299.pdf r1 manage 29.8 K 2022-06-29 - 14:20 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PNGpng DL1d_cjet_ROC-1.png r1 manage 66.5 K 2022-06-29 - 15:08 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PDFpdf DL1d_cjet_ROC.pdf r1 manage 36.8 K 2022-06-29 - 14:20 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PNGpng DL1d_disc-1.png r1 manage 55.8 K 2022-06-29 - 15:09 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PDFpdf DL1d_disc.pdf r1 manage 35.2 K 2022-06-29 - 14:12 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PNGpng DL1d_ujet_ROC-1.png r1 manage 74.1 K 2022-06-29 - 15:09 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PDFpdf DL1d_ujet_ROC.pdf r1 manage 37.6 K 2022-06-29 - 14:20 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
Unknown file formateps Eff_2dim_abs_BAR.eps r1 manage 42.5 K 2020-07-26 - 00:09 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
PNGpng Eff_2dim_abs_BAR.png r1 manage 19.3 K 2020-07-26 - 00:08 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
Unknown file formateps Eff_2dim_abs_EC.eps r1 manage 45.5 K 2020-07-26 - 00:09 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
PNGpng Eff_2dim_abs_EC.png r1 manage 20.3 K 2020-07-26 - 00:08 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
Unknown file formateps Eff_central_barrel.eps r1 manage 26.1 K 2020-07-26 - 00:09 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
PNGpng Eff_central_barrel.png r1 manage 15.6 K 2020-07-26 - 00:08 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
Unknown file formateps Eff_central_endcaps.eps r1 manage 27.0 K 2020-07-26 - 00:09 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
PNGpng Eff_central_endcaps.png r1 manage 16.0 K 2020-07-26 - 00:08 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
Unknown file formateps Eff_peripheral_barrel.eps r1 manage 26.9 K 2020-07-26 - 00:09 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
PNGpng Eff_peripheral_barrel.png r1 manage 15.5 K 2020-07-26 - 00:08 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
Unknown file formateps Eff_peripheral_endcaps.eps r1 manage 26.2 K 2020-07-26 - 00:09 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
PNGpng Eff_peripheral_endcaps.png r1 manage 15.4 K 2020-07-26 - 00:08 SuyogShrestha Muon-Jet Trigger Performance in Run2 Data
Unknown file formateps HLT_bjet_for_pub.eps r1 manage 11.9 K 2015-09-21 - 16:20 LidijaZivkovic b jet trigger rates
PDFpdf HLT_bjet_for_pub.pdf r1 manage 22.5 K 2015-09-21 - 16:20 LidijaZivkovic b jet trigger rates
PNGpng HLT_bjet_for_pub.png r1 manage 23.2 K 2015-09-21 - 16:20 LidijaZivkovic b jet trigger rates
PDFpdf Presel_HH4b_rej_eff.pdf r1 manage 138.2 K 2022-11-07 - 18:17 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng Presel_HH4b_rej_eff.png r1 manage 71.1 K 2022-11-07 - 18:17 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
Unknown file formateps ROC_cb.eps r1 manage 22.0 K 2019-05-02 - 10:27 CarloVarni Hybrid tuning performances in 2018
PDFpdf ROC_cb.pdf r1 manage 27.9 K 2019-05-02 - 10:27 CarloVarni Hybrid tuning performances in 2018
PNGpng ROC_cb.png r1 manage 102.3 K 2019-05-02 - 10:27 CarloVarni Hybrid tuning performances in 2018
Unknown file formateps ROC_cb_Zprime.eps r1 manage 23.3 K 2019-05-02 - 10:46 CarloVarni Hybrid tuning performances in 2018
PDFpdf ROC_cb_Zprime.pdf r1 manage 27.0 K 2019-05-02 - 10:46 CarloVarni Hybrid tuning performances in 2018
PNGpng ROC_cb_Zprime.png r1 manage 93.8 K 2019-05-02 - 10:46 CarloVarni Hybrid tuning performances in 2018
Unknown file formateps ROC_lb.eps r1 manage 24.2 K 2019-05-02 - 10:27 CarloVarni Hybrid tuning performances in 2018
PDFpdf ROC_lb.pdf r1 manage 28.2 K 2019-05-02 - 10:27 CarloVarni Hybrid tuning performances in 2018
PNGpng ROC_lb.png r1 manage 93.7 K 2019-05-02 - 10:27 CarloVarni Hybrid tuning performances in 2018
Unknown file formateps ROC_lb_Zprime.eps r1 manage 24.9 K 2019-05-02 - 10:46 CarloVarni Hybrid tuning performances in 2018
PDFpdf ROC_lb_Zprime.pdf r1 manage 27.4 K 2019-05-02 - 10:46 CarloVarni Hybrid tuning performances in 2018
PNGpng ROC_lb_Zprime.png r1 manage 110.5 K 2019-05-02 - 10:46 CarloVarni Hybrid tuning performances in 2018
PDFpdf Run-3_Trigger-taggers_comparison.pdf r1 manage 290.9 K 2022-11-07 - 18:17 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng Run-3_Trigger-taggers_comparison.png r1 manage 145.5 K 2022-11-07 - 18:17 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PDFpdf Run-3_trks_d0.pdf r1 manage 189.3 K 2022-11-07 - 17:27 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng Run-3_trks_d0.png r1 manage 55.0 K 2022-11-07 - 17:24 StefanoFranchellucci Tracks d0 for all tracking steps in b-jet trigger chains
PDFpdf Run-3_trks_z0.pdf r1 manage 188.3 K 2022-11-07 - 17:27 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng Run-3_trks_z0.png r1 manage 46.2 K 2022-11-07 - 17:27 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng Trigger_rates_16_05_DL1d-1.png r1 manage 85.4 K 2022-06-29 - 15:10 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PDFpdf Trigger_rates_16_05_DL1d.pdf r1 manage 25.4 K 2022-06-29 - 14:20 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PNGpng Trigger_rates_16_05_DL1dbb-1.png r1 manage 88.2 K 2022-06-29 - 15:10 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PDFpdf Trigger_rates_16_05_DL1dbb.pdf r1 manage 27.1 K 2022-06-29 - 14:20 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
Unknown file formateps bJetTriggerPerf.eps r1 manage 16.7 K 2015-07-17 - 14:18 KatharineLeney  
PDFpdf bJetTriggerPerf.pdf r1 manage 58.2 K 2015-07-17 - 14:18 KatharineLeney  
PNGpng bJetTriggerPerf.png r1 manage 179.6 K 2015-07-17 - 14:25 KatharineLeney  
Unknown file formateps bJetTriggerPerf_cRej.eps r1 manage 15.4 K 2015-07-17 - 14:18 KatharineLeney  
PDFpdf bJetTriggerPerf_cRej.pdf r1 manage 56.1 K 2015-07-17 - 14:18 KatharineLeney  
PNGpng bJetTriggerPerf_cRej.png r1 manage 153.0 K 2015-07-17 - 14:25 KatharineLeney  
Unknown file formateps charm.eps r1 manage 23.8 K 2017-07-04 - 10:17 RuchiGupta bjet Trigger Expected Performance in 2017
PDFpdf charm.pdf r1 manage 34.2 K 2017-07-04 - 10:17 RuchiGupta bjet Trigger Expected Performance in 2017
PNGpng charm.png r1 manage 72.0 K 2017-07-04 - 10:17 RuchiGupta bjet Trigger Expected Performance in 2017
PNGpng combsfPrel.png r1 manage 18.8 K 2020-05-28 - 13:02 SuyogShrestha b-jet trigger calibration in 2018
Unknown file formateps condsfPrel.eps r1 manage 14.8 K 2020-05-30 - 01:37 SuyogShrestha  
PDFpdf effSyst_offJets70_match_hlt60_jetPt_00-02-00.pdf r1 manage 14.5 K 2017-02-14 - 21:56 LidijaZivkovic efficiency for 2016
PNGpng effSyst_offJets70_match_hlt60_jetPt_00-02-00.png r1 manage 21.0 K 2017-02-14 - 21:55 LidijaZivkovic efficiency for 2016
Unknown file formateps eff_Off70_matched_HLT40_18June2019.eps r1 manage 14.8 K 2019-07-25 - 12:52 CarloVarni  
PNGpng eff_Off70_matched_HLT40_18June2019.png r1 manage 113.8 K 2019-07-25 - 14:42 CarloVarni  
Unknown file formateps eff_Off70_matched_HLT60_18June2019.eps r1 manage 14.2 K 2019-07-25 - 12:52 CarloVarni  
PNGpng eff_Off70_matched_HLT60_18June2019.png r1 manage 109.6 K 2019-07-25 - 14:42 CarloVarni  
Unknown file formateps eff_Off70_matched_HLT70_18June2019.eps r1 manage 14.3 K 2019-07-25 - 12:52 CarloVarni  
PNGpng eff_Off70_matched_HLT70_18June2019.png r1 manage 109.5 K 2019-07-25 - 14:42 CarloVarni  
PDFpdf fastDIPS_DR_training.pdf r1 manage 342.3 K 2022-11-07 - 17:52 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng fastDIPS_DR_training.png r1 manage 109.7 K 2022-11-07 - 17:52 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PDFpdf fastDIPS_conditional.pdf r1 manage 273.6 K 2022-11-07 - 18:17 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng fastDIPS_conditional.png r1 manage 102.7 K 2022-11-07 - 18:17 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PDFpdf fastDIPS_track-pT_training.pdf r1 manage 342.2 K 2022-11-07 - 17:52 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
PNGpng fastDIPS_track-pT_training.png r1 manage 103.0 K 2022-11-07 - 17:52 StefanoFranchellucci Performance of Run 3 HLT b-tagging with fast tracking
Unknown file formateps flavFillPrel.eps r1 manage 44.1 K 2020-05-30 - 01:37 SuyogShrestha  
PNGpng flavFillPrel.png r1 manage 32.0 K 2020-05-28 - 13:02 SuyogShrestha b-jet trigger calibration in 2018
Unknown file formateps light.eps r2 r1 manage 28.0 K 2017-07-04 - 10:21 RuchiGupta  
PDFpdf light.pdf r2 r1 manage 35.3 K 2017-07-04 - 10:21 RuchiGupta  
PNGpng light.png r2 r1 manage 49.2 K 2017-07-04 - 10:21 RuchiGupta  
Unknown file formateps onlinesfPrel.eps r1 manage 14.0 K 2020-05-30 - 01:37 SuyogShrestha  
PNGpng onlinesfPrel.png r1 manage 17.8 K 2020-05-28 - 13:02 SuyogShrestha b-jet trigger calibration in 2018
PDFpdf run3-effs.pdf r1 manage 32.7 K 2022-09-09 - 12:26 ChrisPollard  
PNGpng run3-effs.png r1 manage 76.9 K 2022-09-09 - 12:26 ChrisPollard  
PDFpdf run3main-effs.pdf r1 manage 30.3 K 2022-09-09 - 12:26 ChrisPollard  
PNGpng run3main-effs.png r1 manage 80.9 K 2022-09-09 - 12:26 ChrisPollard  
PDFpdf runcomp-effs.pdf r1 manage 30.4 K 2022-09-09 - 12:26 ChrisPollard  
PNGpng runcomp-effs.png r1 manage 71.0 K 2022-09-09 - 12:26 ChrisPollard  
Unknown file formateps sampFillPrel.eps r1 manage 45.2 K 2020-05-30 - 01:37 SuyogShrestha  
PNGpng sampFillPrel.png r1 manage 32.9 K 2020-05-28 - 13:02 SuyogShrestha b-jet trigger calibration in 2018
PNGpng scores_DL1d20211216_bb_ICHEP_299-1.png r1 manage 86.6 K 2022-06-29 - 15:05 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PDFpdf scores_DL1d20211216_bb_ICHEP_299.pdf r1 manage 31.8 K 2022-06-29 - 14:20 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PNGpng scores_DL1dbb_ICHEP_299-1.png r1 manage 74.4 K 2022-06-29 - 15:07 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
PDFpdf scores_DL1dbb_ICHEP_299.pdf r1 manage 29.4 K 2022-06-29 - 14:20 MeiqiChen Expected Performance of b-jet Trigger Algorithms for Run 3
Unknown file formateps ttsystPrel.eps r1 manage 17.3 K 2020-05-30 - 01:37 SuyogShrestha  
PNGpng ttsystPrel.png r1 manage 21.0 K 2020-05-28 - 13:02 SuyogShrestha b-jet trigger calibration in 2018
Edit | Attach | Watch | Print version | History: r31 < r30 < r29 < r28 < r27 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r31 - 2022-11-07 - StefanoFranchellucci
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Atlas All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback