Compact Muon Solenoid
LHC, CERN
Visit us: CMS Public Website, CMS Physics ;
Contact us: CMS Publications Committee
| CMS-TAU-24-002 ; CERN-EP-2026-008 | ||
| High-level hadronic tau lepton triggers of the CMS experiment in proton-proton collisions at $ \sqrt{s} = $ 13.6 TeV | ||
| CMS Collaboration | ||
| 11 February 2026 | ||
| Submitted to the Journal of Instrumentation | ||
| Abstract: The trigger system of the CMS detector is pivotal in the acquisition of data for physics measurements and searches. Studies of final states characterized by hadronic decays of tau leptons require the reconstruction and the identification of genuine tau leptons against quark- and gluon-initiated jets at the trigger level. This is a difficult task, particularly as improvements to the LHC have resulted in an increased number of interactions per bunch crossing in recent years. To address this challenge, a series of machine-learning algorithms with high identification efficiency and low computational cost have been incorporated into the high-level trigger for hadronically decaying tau leptons. In this paper, these developments and the trigger performance are summarized using data collected by the CMS experiment in proton-proton collisions at $ \sqrt{s} = $ 13.6 TeV in 2022--2023, corresponding to an integrated luminosity of 62 fb$ ^{-1} $. | ||
| Links: e-print arXiv:2602.11359 [hep-ex] (PDF) ; CDS record ; inSPIRE record ; CADI line (restricted) ; | ||
| Figures | |
png pdf |
Figure 1:
Workflows for $ \tau_\mathrm{h} $ candidate reconstruction at the HLT in Run 2 [21]. |
png pdf |
Figure 2:
Workflows for $ \tau_\mathrm{h} $ candidate reconstruction at the HLT in Run 3, since 2022. |
png pdf |
Figure 3:
Performance of the L2TAUNNTAG in the di-$ \tau_\mathrm{h} $ HLT path, using simulated $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ events. The absolute efficiency of the reconstructed L2 $ \tau_\mathrm{h} $ candidates as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $ (left) and $ \eta $ (right) are shown, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 3-a:
Performance of the L2TAUNNTAG in the di-$ \tau_\mathrm{h} $ HLT path, using simulated $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ events. The absolute efficiency of the reconstructed L2 $ \tau_\mathrm{h} $ candidates as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $ (left) and $ \eta $ (right) are shown, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 3-b:
Performance of the L2TAUNNTAG in the di-$ \tau_\mathrm{h} $ HLT path, using simulated $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ events. The absolute efficiency of the reconstructed L2 $ \tau_\mathrm{h} $ candidates as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $ (left) and $ \eta $ (right) are shown, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 4:
Performance of the L2TAUNNTAG in the single-$ \tau_\mathrm{h} $ HLT path, using simulated $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ events. The absolute efficiency of the reconstructed L2 $ \tau_\mathrm{h} $ candidates as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $ (left) and $ \eta $ (right) are shown, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 4-a:
Performance of the L2TAUNNTAG in the single-$ \tau_\mathrm{h} $ HLT path, using simulated $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ events. The absolute efficiency of the reconstructed L2 $ \tau_\mathrm{h} $ candidates as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $ (left) and $ \eta $ (right) are shown, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 4-b:
Performance of the L2TAUNNTAG in the single-$ \tau_\mathrm{h} $ HLT path, using simulated $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ events. The absolute efficiency of the reconstructed L2 $ \tau_\mathrm{h} $ candidates as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $ (left) and $ \eta $ (right) are shown, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 5:
Total L1$ + $ HLT path efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), single-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h} $ (lower right) HLT paths as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ samples are used in the evaluation. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 5-a:
Total L1$ + $ HLT path efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), single-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h} $ (lower right) HLT paths as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ samples are used in the evaluation. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 5-b:
Total L1$ + $ HLT path efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), single-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h} $ (lower right) HLT paths as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ samples are used in the evaluation. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 5-c:
Total L1$ + $ HLT path efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), single-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h} $ (lower right) HLT paths as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ samples are used in the evaluation. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 5-d:
Total L1$ + $ HLT path efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), single-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h} $ (lower right) HLT paths as a function of the visible generator-level $ \tau_\mathrm{h} p_{\mathrm{T}} $, where ``visible'' refers to the fact that the contribution of neutrinos is not taken into account. The $ \mathrm{H}\to\tau\tau $ and BSM $ \mathrm{Z}^{'}\to\tau\tau $ samples are used in the evaluation. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 6:
A comparison of the L1$ + $ HLT efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 6-a:
A comparison of the L1$ + $ HLT efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 6-b:
A comparison of the L1$ + $ HLT efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 6-c:
A comparison of the L1$ + $ HLT efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 6-d:
A comparison of the L1$ + $ HLT efficiency of the $ \mathrm{e}\tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 7:
A comparison of the L1$ + $ HLT efficiency of the $ \mu\tau_\mathrm{h} $ HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 7-a:
A comparison of the L1$ + $ HLT efficiency of the $ \mu\tau_\mathrm{h} $ HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 7-b:
A comparison of the L1$ + $ HLT efficiency of the $ \mu\tau_\mathrm{h} $ HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 7-c:
A comparison of the L1$ + $ HLT efficiency of the $ \mu\tau_\mathrm{h} $ HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 7-d:
A comparison of the L1$ + $ HLT efficiency of the $ \mu\tau_\mathrm{h} $ HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 8:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 8-a:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 8-b:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 8-c:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 8-d:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 9:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h}{+}\text{jet} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 9-a:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h}{+}\text{jet} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 9-b:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h}{+}\text{jet} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 9-c:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h}{+}\text{jet} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 9-d:
A comparison of the L1$ + $ HLT efficiency of the di-$ \tau_\mathrm{h}{+}\text{jet} $ monitoring HLT path in 2022 and 2023 as a function of offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ (upper left), $ \eta $ (upper right), and $ \phi $ (lower left). The dependence on the $ N_{\text{PV}} $ is also shown (lower right). The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. |
png pdf |
Figure 10:
Efficiencies and scale factors of the HLT monitoring paths using 2022--2023 data as functions of the offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ for the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), di-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h}{+}\text{jet} $ (lower right) HLT paths. The measured efficiencies for data are shown with black markers, and for simulation with green markers. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. The systematic uncertainties are negligible with respect to the statistical uncertainties and are not included in the bands. The corresponding dotted lines of the same color display the best fit results together with the statistical uncertainty bands. The scale factors, defined as the ratios of efficiencies between data fitted and simulation fitted from the upper panels, are shown in the lower panels as blue lines with associated blue uncertainty bands. Only values to the right of the red dotted line are used in physics studies, to avoid large statistical fluctuations, as they are sufficiently above the turn-on threshold of the trigger. |
png pdf |
Figure 10-a:
Efficiencies and scale factors of the HLT monitoring paths using 2022--2023 data as functions of the offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ for the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), di-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h}{+}\text{jet} $ (lower right) HLT paths. The measured efficiencies for data are shown with black markers, and for simulation with green markers. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. The systematic uncertainties are negligible with respect to the statistical uncertainties and are not included in the bands. The corresponding dotted lines of the same color display the best fit results together with the statistical uncertainty bands. The scale factors, defined as the ratios of efficiencies between data fitted and simulation fitted from the upper panels, are shown in the lower panels as blue lines with associated blue uncertainty bands. Only values to the right of the red dotted line are used in physics studies, to avoid large statistical fluctuations, as they are sufficiently above the turn-on threshold of the trigger. |
png pdf |
Figure 10-b:
Efficiencies and scale factors of the HLT monitoring paths using 2022--2023 data as functions of the offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ for the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), di-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h}{+}\text{jet} $ (lower right) HLT paths. The measured efficiencies for data are shown with black markers, and for simulation with green markers. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. The systematic uncertainties are negligible with respect to the statistical uncertainties and are not included in the bands. The corresponding dotted lines of the same color display the best fit results together with the statistical uncertainty bands. The scale factors, defined as the ratios of efficiencies between data fitted and simulation fitted from the upper panels, are shown in the lower panels as blue lines with associated blue uncertainty bands. Only values to the right of the red dotted line are used in physics studies, to avoid large statistical fluctuations, as they are sufficiently above the turn-on threshold of the trigger. |
png pdf |
Figure 10-c:
Efficiencies and scale factors of the HLT monitoring paths using 2022--2023 data as functions of the offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ for the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), di-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h}{+}\text{jet} $ (lower right) HLT paths. The measured efficiencies for data are shown with black markers, and for simulation with green markers. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. The systematic uncertainties are negligible with respect to the statistical uncertainties and are not included in the bands. The corresponding dotted lines of the same color display the best fit results together with the statistical uncertainty bands. The scale factors, defined as the ratios of efficiencies between data fitted and simulation fitted from the upper panels, are shown in the lower panels as blue lines with associated blue uncertainty bands. Only values to the right of the red dotted line are used in physics studies, to avoid large statistical fluctuations, as they are sufficiently above the turn-on threshold of the trigger. |
png pdf |
Figure 10-d:
Efficiencies and scale factors of the HLT monitoring paths using 2022--2023 data as functions of the offline $ \tau_\mathrm{h} $ candidate $ p_{\mathrm{T}} $ for the $ \mathrm{e}\tau_\mathrm{h} $ (upper left), $ \mu\tau_\mathrm{h} $ (upper right), di-$ \tau_\mathrm{h} $ (lower left), and di-$ \tau_\mathrm{h}{+}\text{jet} $ (lower right) HLT paths. The measured efficiencies for data are shown with black markers, and for simulation with green markers. The uncertainties shown by the vertical bars are from the number of events available in the sample, while the horizontal bars show the bin width. Some of the vertical bars are smaller than the markers and are not shown. The systematic uncertainties are negligible with respect to the statistical uncertainties and are not included in the bands. The corresponding dotted lines of the same color display the best fit results together with the statistical uncertainty bands. The scale factors, defined as the ratios of efficiencies between data fitted and simulation fitted from the upper panels, are shown in the lower panels as blue lines with associated blue uncertainty bands. Only values to the right of the red dotted line are used in physics studies, to avoid large statistical fluctuations, as they are sufficiently above the turn-on threshold of the trigger. |
| Tables | |
png pdf |
Table 1:
Decay modes and branching fractions ($ \mathcal{B} $) of the tau lepton alongside the mesonic resonances primarily involved in hadronic tau lepton decays [15]. |
png pdf |
Table 2:
Rate estimation and observation for the cut-based and L2TAUNNTAG algorithms. Column A scales Run 2 data collected by the cut-based algorithm to Run 3 conditions, column B re-emulates Run 2 data using the L2TAUNNTAG algorithm and scales the result to Run 3 conditions, and column C is the evaluated rate in Run 3. The instantaneous luminosities used were 1.68, 2.00, and 2.20 $ \times 10^{34} \text{cm}^{-2} \text{s}^{-1} $, respectively. The rates are inclusive calculations not excluding shared contributions from other algorithms or paths. The statistical uncertainties are negligible with respect to the significant digits reported. |
png pdf |
Table 3:
Rate estimation and observation for several $ \tau_\mathrm{h} $ HLT paths. Similarly to Table 2, column A scales Run 2 data collected by the cut-based algorithm to Run 3 conditions, column B re-emulates Run 2 data using the L2TAUNNTAG algorithm and scales the result to Run 3 conditions, and column C is the evaluated rate in Run 3. The instantaneous luminosities used for the Run 2 estimation, Run 3 projection, and Run 3 evaluation were 1.68, 2.00, and 2.20 $ \times 10^{34} \text{cm}^{-2} \text{s}^{-1} $, respectively. The rates are inclusive calculations not excluding shared contributions from other algorithms or paths. The statistical uncertainties are negligible with respect to the significant digits reported. |
| Summary |
| Two online machine-learning algorithms, L2TAUNNTAG, described here for the first time, and DEEPTAU, have been deployed into the high level trigger (HLT) to select hadronically decaying tau lepton ($ \tau_\mathrm{h} $) candidates. Their performance has been evaluated using the data collected by the CMS experiment in proton-proton collisions at $ \sqrt{s} = $ 13.6 TeV in 2022--2023, corresponding to an integrated luminosity of 62 fb$ ^{-1} $. Comparisons to simulation were performed and show good agreement with the collected data, validating the current understanding of the HLT paths involving $ \tau_\mathrm{h} $ candidates. The updated HLT paths are found to deliver improved $ \tau_\mathrm{h} $ candidate identification efficiency without significantly increasing computational cost or event rate, allowing more genuine hadronic tau lepton decays to be collected at roughly the same resource cost as in 2018. These improvements benefit physics studies targeting final states with hadronically decaying tau leptons, including precision measurements of the Higgs boson, and searches beyond the standard model. |
| References | ||||
| 1 | CMS Collaboration | Observation of the Higgs boson decay to a pair of $ \tau $ leptons with the CMS detector | PLB 779 (2018) 283 | CMS-HIG-16-043 1708.00373 |
| 2 | CMS Collaboration | Search for Higgs boson pair production in events with two bottom quarks and two tau leptons in proton--proton collisions at $ \sqrt{s} = $ 13 TeV | PLB 778 (2018) 101 | CMS-HIG-17-002 1707.02909 |
| 3 | ATLAS Collaboration | Cross-section measurements of the Higgs boson decaying into a pair of $ \tau $-leptons in proton-proton collisions at $ \sqrt{s} = $ 13 TeV with the ATLAS detector | PRD 99 (2019) 072001 | 1811.08856 |
| 4 | ATLAS Collaboration | Search for resonant and non-resonant Higgs boson pair production in the $ {\text{b}\bar{\text{b}}\tau^+\tau^-} $ decay channel in pp collisions at $ \sqrt{s} = $ 13 TeV with the ATLAS detector | [Erratum: \DOI10.1103/PhysRevLett.122.089901, Phys. Rev. Lett. 122 () 089901], 2018 PRL 121 (2018) 191801 |
1808.00336 |
| 5 | ATLAS Collaboration | Test of CP invariance in vector-boson fusion production of the Higgs boson in the $ \text{H}\rightarrow\tau\tau $ channel in proton-proton collisions at $ \sqrt{s} = $ 13 TeV with the ATLAS detector | PLB 805 (2020) 135426 | 2002.05315 |
| 6 | CMS Collaboration | Measurement of the production cross section of a Higgs boson with large transverse momentum in its decays to a pair of $ \tau $ leptons in proton-proton collisions at $ \sqrt{s} = $ 13 TeV | PLB 857 (2024) 138964 | CMS-HIG-21-017 2403.20201 |
| 7 | CMS Collaboration | Search for Higgs boson pairs decaying to WW*WW*, WW*$ \tau\tau $, and $ \tau\tau\tau\tau $ in proton-proton collisions at $ \sqrt{s} = $ 13 TeV | JHEP 07 (2023) 095 | CMS-HIG-21-002 2206.10268 |
| 8 | CMS Collaboration | Observation of $ \gamma\gamma\to\tau\tau $ in proton-proton collisions and limits on the anomalous electromagnetic moments of the $ \tau $ lepton | Rept. Prog. Phys. 87 (2024) 107801 | CMS-SMP-23-005 2406.03975 |
| 9 | ATLAS Collaboration | Observation of the $ \gamma\gamma\to\tau\tau $ process in Pb+Pb collisions and constraints on the $ \tau $-lepton anomalous magnetic moment with the ATLAS detector | PRL 131 (2023) 151802 | 2204.13478 |
| 10 | CMS Collaboration | Measurement of the $ \tau $ lepton polarization in Z boson decays in proton-proton collisions at $ \sqrt{s} = $ 13 TeV | JHEP 01 (2024) 101 | CMS-SMP-18-010 2309.12408 |
| 11 | ATLAS Collaboration | Measurement of $ \tau $ polarisation in $ {Z}/\gamma^{*}\rightarrow \tau \tau $ decays in proton-proton collisions at $ \sqrt{s} = $ 8 TeV with the ATLAS detector | EPJC 78 (2018) 163 | 1709.03490 |
| 12 | ATLAS Collaboration | Measurement of $ \tau $ polarization in $ {W} \rightarrow \tau \nu $ decays with the ATLAS detector in pp collisions at $ \sqrt{s} = $ 7 TeV | EPJC 72 (2012) 2062 | 1204.6720 |
| 13 | CMS Collaboration | Searches for additional Higgs bosons and for vector leptoquarks in $ \tau\tau $ final states in proton-proton collisions at $ \sqrt{s} = $ 13 TeV | JHEP 07 (2023) 073 | CMS-HIG-21-001 2208.02717 |
| 14 | ATLAS Collaboration | Search for heavy Higgs bosons decaying into two tau leptons with the ATLAS detector using pp collisions at $ \sqrt{s} = $ 13 TeV | PRL 125 (2020) 051801 | 2002.12223 |
| 15 | Particle Data Group , S. Navas et al. | Review of particle physics | PRD 110 (2024) 030001 | |
| 16 | CMS Collaboration | The CMS experiment at the CERN LHC | JINST 3 (2008) S08004 | 1003.4038 |
| 17 | CMS Collaboration | Development of the CMS detector for the CERN LHC Run 3 | JINST 19 (2024) P05064 | CMS-PRF-21-001 2309.05466 |
| 18 | CMS Collaboration | Performance of reconstruction and identification of $ \tau $ leptons decaying to hadrons and $ \nu_\tau $ in pp collisions at $ \sqrt{s} = $ 13 TeV | JINST 13 (2018) P10005 | CMS-TAU-16-003 1809.02816 |
| 19 | CMS Collaboration | Performance of the CMS Level-1 trigger in proton-proton collisions at $ \sqrt{s} = $ 13 TeV | JINST 15 (2020) P10017 | CMS-TRG-17-001 2006.10165 |
| 20 | CMS Collaboration | The CMS trigger system | JINST 12 (2017) P01020 | CMS-TRG-12-001 1609.02366 |
| 21 | CMS Collaboration | Performance of the CMS high-level trigger during LHC Run 2 | JINST 19 (2024) P11021 | CMS-TRG-19-001 2410.17038 |
| 22 | CMS Collaboration | Electron and photon reconstruction and identification with the CMS experiment at the CERN LHC | JINST 16 (2021) P05014 | CMS-EGM-17-001 2012.06888 |
| 23 | CMS Collaboration | Performance of the CMS muon detector and muon reconstruction with proton-proton collisions at $ \sqrt{s} = $ 13 TeV | JINST 13 (2018) P06015 | CMS-MUO-16-001 1804.04528 |
| 24 | CMS Collaboration | Description and performance of track and primary-vertex reconstruction with the CMS tracker | JINST 9 (2014) P10009 | CMS-TRK-11-001 1405.6569 |
| 25 | CMS Collaboration | CMS technical design report for the Phase 1 upgrade of the hadron calorimeter | Technical Report CERN-LHCC-2012-015, CMS-TDR-010, 2012 CDS |
|
| 26 | CMS Collaboration | Performance of the CMS Phase-1 pixel detector with Run 3 data | CMS Detector Performance Note CMS-DP-2022-047, 2022 CDS |
|
| 27 | CMS Collaboration | Commissioning CMS online reconstruction with GPUs | CMS Detector Performance Note CMS-DP-2023-004, 2022 CDS |
|
| 28 | CMS BRIL Collaboration | The pixel luminosity telescope: a detector for luminosity measurement at CMS using silicon pixel sensors | EPJC 83 (2023) 673 | 2206.08870 |
| 29 | N. Karunarathna | Run 3 luminosity measurements with the pixel luminosity telescope | CMS Collaboration, in st International Conference on High Energy physics, 2022 Proceedings of 4 (2022) 936 |
|
| 30 | J. Wa \'n czyk | Upgraded CMS fast beam condition monitor for LHC Run 3 online luminosity and beam induced background measurements | CMS Collaboration, in th Int. Beam Instrum. Conf, 2022 Proc. 1 (2022) 540 |
|
| 31 | CMS Collaboration | Particle-flow reconstruction and global event description with the CMS detector | JINST 12 (2017) P10003 | CMS-PRF-14-001 1706.04965 |
| 32 | D. Contardo et al. | Technical proposal for the Phase-II upgrade of the CMS detector | CMS Technical Design Review CMS-TDR-15-02, 2015 link |
|
| 33 | M. Cacciari, G. P. Salam, and G. Soyez | The anti-$ k_{\mathrm{T}} $ jet clustering algorithm | JHEP 04 (2008) 063 | 0802.1189 |
| 34 | M. Cacciari, G. P. Salam, and G. Soyez | FastJet user manual | EPJC 72 (2012) 1896 | 1111.6097 |
| 35 | CMS Collaboration | Pileup mitigation at CMS in 13 TeV data | JINST 15 (2020) P09018 | CMS-JME-18-001 2003.00503 |
| 36 | D. Bertolini, P. Harris, M. Low, and N. Tran | Pileup per particle identification | JHEP 10 (2014) 059 | 1407.6013 |
| 37 | CMS Collaboration | Jet energy scale and resolution in the CMS experiment in pp collisions at 8 TeV | JINST 12 (2017) P02014 | CMS-JME-13-004 1607.03663 |
| 38 | CMS Collaboration | Performance of electron reconstruction and selection with the CMS detector in proton-proton collisions at $ \sqrt{s} = $ 8TeV | JINST 10 (2015) P06005 | CMS-EGM-13-001 1502.02701 |
| 39 | CMS Collaboration | Performance of missing transverse momentum reconstruction in proton-proton collisions at $ \sqrt{s} = $ 13 TeV using the CMS detector | JINST 14 (2019) P07004 | CMS-JME-17-001 1903.06078 |
| 40 | CMS Collaboration | Performance of $ \tau $-lepton reconstruction and identification in CMS | JINST 7 (2012) P01001 | CMS-TAU-11-001 1109.6034 |
| 41 | CMS Collaboration | Reconstruction and identification of $ \tau $ lepton decays to hadrons and $ \nu_\tau $ at CMS | JINST 11 (2016) P01019 | CMS-TAU-14-001 1510.07488 |
| 42 | CMS Collaboration | Performance of the DeepTau algorithm for the discrimination of taus against jets, electron, and muons | CMS Detector Performance Note CMS-DP-2019-033, 2019 CDS |
|
| 43 | CMS Collaboration | Identification of hadronic tau lepton decays using a deep neural network | JINST 17 (2022) P07023 | CMS-TAU-20-001 2201.08458 |
| 44 | CMS Collaboration | Luminosity measurement in proton-proton collisions at 13.6 TeV in 2022 at CMS | CMS Physics Analysis Summary, 2024 CMS-PAS-LUM-22-001 |
CMS-PAS-LUM-22-001 |
| 45 | CMS Collaboration | Measurement of the offline integrated luminosity for the CMS proton-proton collision dataset recorded in 2023 | CMS Detector Performance Note CMS-DP-2024-068, 2024 CDS |
|
| 46 | CMS Collaboration | Luminosity determination using Z boson production at the CMS experiment | EPJC 84 (2024) 26 | CMS-LUM-21-001 2309.01008 |
| 47 | CMS Collaboration | Precision luminosity measurement in proton-proton collisions at $ \sqrt{s} = $ 13 TeV in 2015 and 2016 at CMS | EPJC 81 (2021) 800 | CMS-LUM-17-003 2104.01927 |
| 48 | J. Alwall et al. | The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations | JHEP 07 (2014) 079 | 1405.0301 |
| 49 | J. Alwall et al. | Comparative study of various algorithms for the merging of parton showers and matrix elements in hadronic collisions | EPJC 53 (2008) 473 | 0706.2569 |
| 50 | T. Sjöstrand et al. | An introduction to PYTHIA 8.2 | Comput. Phys. Commun. 191 (2015) 159 | 1410.3012 |
| 51 | P. Nason | A new method for combining NLO QCD with shower Monte Carlo algorithms | JHEP 11 (2004) 040 | hep-ph/0409146 |
| 52 | S. Frixione, P. Nason, and C. Oleari | Matching NLO QCD computations with parton shower simulations: the POWHEG method | JHEP 11 (2007) 070 | 0709.2092 |
| 53 | S. Alioli, P. Nason, C. Oleari, and E. Re | A general framework for implementing NLO calculations in shower Monte Carlo programs: the POWHEG BOX | JHEP 06 (2010) 043 | 1002.2581 |
| 54 | J. M. Campbell, R. K. Ellis, P. Nason, and E. Re | Top-pair production and decay at NLO matched with parton showers | JHEP 04 (2015) 114 | 1412.1828 |
| 55 | S. Frixione, P. Nason, and G. Ridolfi | A positive-weight next-to-leading-order Monte Carlo for heavy flavour hadroproduction | JHEP 09 (2007) 126 | 0707.3088 |
| 56 | P. Nason and C. Oleari | NLO Higgs boson production via vector-boson fusion matched with shower in POWHEG | JHEP 02 (2010) 037 | 0911.5299 |
| 57 | CMS Collaboration | Extraction and validation of a new set of CMS PYTHIA8 tunes from underlying-event measurements | EPJC 80 (2020) 4 | CMS-GEN-17-001 1903.12179 |
| 58 | N. Davidson et al. | Universal interface of TAUOLA technical and physics documentation | Comput. Phys. Commun. 183 (2012) 821 | 1002.0543 |
| 59 | NNPDF Collaboration | Parton distributions for the LHC Run II | JHEP 04 (2015) 040 | 1410.8849 |
| 60 | NNPDF Collaboration | Parton distributions from high-precision collider data | EPJC 77 (2017) 663 | 1706.00428 |
| 61 | GEANT4 Collaboration | GEANT 4---A simulation toolkit | NIM A 506 (2003) 250 | |
| 62 | CMS Collaboration | Performance of Level-1 trigger e/gamma and tau in Run 3 | CMS Detector Performance Note CMS-DP-2023-008, 2023 CDS |
|
| 63 | A. Bocci et al. | Heterogeneous reconstruction of tracks and primary vertices with the CMS pixel tracker | Front. Big Data 3 (2020) 601728 | 2008.13461 |
| 64 | CMS Collaboration | Performance of Run-3 HLT track reconstruction | CMS Detector Performance Note CMS-DP-2022-014, 2022 CDS |
|
| 65 | D. P. Kingma and J. Ba | Adam: A method for stochastic optimization | in 3rd International Conference on Learning Representations, ICLR, 2014 | 1412.6980 |
| 66 | CMS Collaboration | Identification of tau leptons using a convolutional neural network with domain adaptation | JINST 20 (2025) P12032 | CMS-TAU-24-001 2511.05468 |
| 67 | I. J. Goodfellow, Y. Bengio, and A. Courville | Deep Learning | MIT Press, Cambridge, MA, USA, ISBN 978-026613, 2016 link |
|
| 68 | CMS Collaboration | CMS RPC efficiency measurement using the tag-and-probe method | JINST 14 (2019) C10020 | |
|
|
Compact Muon Solenoid LHC, CERN |
|
|
|
|
|
|