The Higgs-to-diphoton decay channel is a versatile tool for exploring physics beyond the Standard Model. First, its decay rate is sensitive to potential new physics contributions in the loop of virtual particles, allowing for tests of a wide range of new physics models. Second, this channel provides a clean signature of the Higgs boson, making it ideal for precise tests of the Standard Model’s predictions about its properties. Finally, because diphoton events are rare at hadron colliders, this channel is particularly well-suited for identifying unusual Higgs boson production.

Over the past decade, our group has made significant contributions to expanding the potential of the Higgs-to-diphoton decay channel at the LHC. Early efforts helped develop the ATLAS analysis strategy that contributed to the discovery of the Higgs boson and further measurements of its properties. The group’s expertise in this channel has positioned us as leaders in the field. Between 2018 and 2020, members of our group co-led the ATLAS HGamma working group, coordinating over 100 physicists from more than 30 research institutions worldwide. This effort focused on measuring the Higgs boson's diphoton and Z+photon decays, as well as related searches for new physics.

Higgs boson coupling and STXS measurements

Our research group has conducted one of the most detailed single-channel measurements of Higgs boson properties using the Higgs-to-diphoton (γγ) decay channel, published in the Journal of High Energy Physics. This comprehensive study measured event rates across 28 distinct regions of the Higgs boson production phase space, based on Simplified Template Cross Sections (STXS). These measurements have been instrumental in constraining potential Beyond the Standard Model (BSM) interactions, which can be framed within the Standard Model Effective Field Theory.

Our group contributed to nearly every aspect of this analysis. This full Run-2 measurement of Higgs boson properties using the diphoton channel is one of the key inputs to the ATLAS Run-2 flagship publication, "A detailed map of Higgs boson interactions by the ATLAS experiment ten years after the discovery".

As the LHC enters Run-3, our current focus is on developing innovative machine learning techniques to further improve the experimental sensitivity to the diphoton channel. Our novel generative machine learning-enabled detector effect modeling method has already been applied in the first ATLAS Run-3 publication, which measured the inclusive Higgs boson production cross section at 13.6 TeV.

Model-independent search for new physics phenomena

Using Run-2 data, our group launched a novel, model-independent search for new physics phenomena using the Higgs boson diphoton decay channel. This innovative approach provided the first comprehensive and systematic examination of Higgs boson events at the LHC. The results, published in 2023, analyzed 23 distinct diphoton final states to investigate any deviations from Standard Model predictions. By presenting the findings in a model-independent format, we enabled a broad evaluation of various beyond the Standard Model (BSM) theories that propose anomalous Higgs boson production.

Search for Higgs boson decaying to delayed and non-pointing photons

Despite a decade of detailed measurements of the Higgs boson's properties, there remains the possibility that this particle could decay into final states not predicted by the Standard Model. One compelling scenario within supersymmetry suggests that the Higgs boson could decay into neutralinos with finite lifetimes. These neutralinos would then decay into photons and the lightest supersymmetric particles, which would escape detection. This model not only offers potential dark matter candidates but also predicts a striking signature: the production of delayed and non-pointing photons.

Our group has conducted an ATLAS project to investigate this type of exotic Higgs boson decay. While no significant deviations from the Standard Model were observed, the search set the most stringent constraints to date on the production of non-pointing photons originating from Higgs boson decays. The results were published and highlighted for their contribution to advancing the understanding of this rare and intriguing possibility.

Search for the Higgs boson decaying to a Z boson and a photon

The Higgs boson's decay into a Z boson and a photon provides a unique opportunity to explore the fundamental interactions of particles. Similar to the Higgs-to-diphoton decay, this process is mediated by a virtual particle loop. Any deviation from Standard Model predictions in either channel would require a detailed correlation analysis to uncover the underlying physics.

Our group played a key role in the ATLAS search for the Higgs to Z boson and photon process and contributed significantly to the statistical combination of the search results from the ATLAS and CMS experiments. In 2023, the combined efforts of both collaborations led to the first evidence of the Higgs boson decaying to a Z boson and a photon. This work, which built on earlier analyses, has been published by Physical Review Letters.

The interaction between the Higgs boson and the top quark, known as the Higgs-top Yukawa interaction, provides a crucial window into one of the seventeen fundamental parameters of the Standard Model. This interaction is particularly noteworthy because the top quark is the heaviest elementary particle, more than tens of thousands of times heavier than the lightest quark. As a result, the Higgs-top Yukawa interaction is the strongest coupling between the Higgs boson and any elementary particle, making it a key focus for numerous beyond Standard Model (BSM) theories attempting to explain its origin. A systematic characterization of the Higgs-top Yukawa interaction, including its strength and CP properties, is essential for probing BSM physics. Recognizing the potential of this area, our group has established a comprehensive research program aimed at systematically characterizing the Higgs-top Yukawa interaction. We employ a combination of complementary measurements and innovative machine-learning techniques to explore this interaction and its potential connections to new physics beyond the Standard Model.

Probing the Higgs-top Yukawa Interaction through the ttH Process

At the LHC, the production of the Higgs boson in association with a top-quark pair, known as the ttH process, provides the most direct way to study the Higgs-top Yukawa interaction. Observing the ttH process is a key step in unlocking deeper insights into this fundamental interaction. In 2018, our group played a pivotal role in the ATLAS effort to observe ttH production by combining results from multiple decay channels. This work marked a significant milestone in understanding the Higgs-top interaction.

Building on this achievement, we further strengthened the evidence by observing ttH production in the diphoton channel. This observation, based on the complete dataset from the LHC Run-2, was one of the first major results from the ATLAS collaboration during this period. This important finding was presented at the prestigious 2019 Moriond Electroweak Conference and highlighted in an ATLAS Physics Briefing for its impact on our understanding of the Higgs coupling to the top quark.

First Test of the CP Properties of the Higgs-top Yukawa Interaction

After the observation of ttH production, our group's focus shifted to characterizing the Higgs-top Yukawa interaction in greater detail. We conducted the first direct experimental test of the CP properties of this interaction by analyzing both the rate and kinematics of ttH production, as well as the associated production of the Higgs boson with a single top quark (tH production).

Our findings, published in Physical Review Letters, provide the most stringent direct constraint on the potential CP-violating component of the Higgs-top interaction, and offer the strongest constraint to date on the cross section of the tH production process. This result, alongside a corresponding CMS result, was highlighted by CERN Courier for its significance in testing the CP symmetry of the Higgs-top Yukawa interaction.

Discovering the Simultaneous Production of Four Top Quarks

The simultaneous production of four top quarks is an exceptionally rare process within the Standard Model. This unique event offers a valuable opportunity to investigate the "off-shell" Higgs-top Yukawa interaction, as well as explore numerous possibilities for new physics beyond the Standard Model. Our group played a central role in the ATLAS effort to observe this process. We developed an advanced event classifier using Graph Neural Networks to capture complex relational information within the four-top final states, significantly improving the ability to isolate these elusive signal events. Our innovation resulted in a substantial 40% enhancement in the ATLAS experiment's sensitivity to the four-top process. In March 2023, we reported the observation of this rare production with a significance exceeding six standard deviations. Building on this achievement, our group also led efforts to use the four-top measurement to constrain the Higgs-top Yukawa interaction strength, its CP properties, and investigate the anomalous production of three top quarks.

Exploring the Associated Production of the Higgs Boson with a Single Top Quark in the Higgs to Bottom Quarks (H → bb) Channel (Ongoing)

This project is ongoing. No public information is available.

The total width of the Higgs boson is an intrinsic property that accounts for all its decay modes, including potential contributions from beyond-the-Standard-Model (BSM) physics. Measuring the total width provides a comprehensive approach for searching for BSM physics. While it was once thought that an electron-positron collider was necessary for precise determination of the Higgs boson’s total width, theoretical advancements over the last decade have demonstrated that the LHC can also achieve model-independent measurements through a combination of on-shell and off-shell Higgs production, though some model-dependent assumptions are still required.

As the LHC continues to gather a vast amount of collision data, the coming years may see the measurement of the Higgs boson’s total width reach new levels of precision, driven by the development of innovative analysis techniques. This progress opens up opportunities to explore new physics scenarios. In 2022, our group expanded its research efforts to include the measurement of the Higgs boson’s total width, focusing on areas that can benefit most from cutting-edge analysis strategies and refined methods. Currently, we are leading two major efforts in this area, aiming to push the boundaries of what is possible in the study of the Higgs boson.

Measurement of the Higgs Boson Total Width Through a Combined Measurement of On-Shell and Off-Shell Processes Involving Higgs-Top Yukawa Interaction

Our group is currently leading a novel approach to determining the total width of the Higgs boson by leveraging processes involving the Higgs-top Yukawa interaction. This method builds on our recent observation of the simultaneous production of four top quarks and our measurements of Higgs boson properties in the diphoton channel. The four-top quark measurement allows us to determine the Higgs-top Yukawa coupling strength without making assumptions about the total width, providing a unique opportunity to combine this with measurements of on-shell Higgs boson production to further constrain the total width.

We have completed a proof-of-principle measurement that combines the on-shell processes of gluon-gluon fusion (ggH) and top-associated production (ttH) with the four-top quark measurement. The results of this effort have been submitted for publication and represent a significant advancement in our ability to constrain the Higgs boson’s total width. Building on this success, our group plans to further expand the measurement by incorporating all on-shell Higgs boson processes, aiming to achieve the highest possible sensitivity to the Higgs boson’s total width.

Ongoing Measurement of the Off-Shell H → WW Production and the Higgs Boson Total Width Constraint

This project is ongoing. No public information is available.

Our group has been at the forefront of developing and applying innovative machine learning architectures tailored to the unique challenges of particle physics. We were among the early adopters of Graph Neural Networks (GNNs), devising a method to represent collision events as graphs and utilizing GNNs for enhanced event classification. This work culminated in the groundbreaking discovery of the simultaneous production of four top quarks in 2023, where our GNN classifier played a pivotal role in establishing the signal.

In addition, we have pioneered a novel approach to measuring particles with complex decay final states, such as top quarks and Higgs bosons. Central to this is the Covariant Particle Transformer (CPT), a transformer-based architecture that incorporates collider physics principles like rotational symmetry and Lorentz invariance, to improve training performance. The CPT allows for the prediction of decaying particle properties by analyzing the entire collision event, bypassing the traditional parton-jet matching method often used in top quark reconstruction.

Recently, our efforts have shifted towards employing GNNs to improve detector performance for photon measurements, a crucial component of several high-impact physics projects in ATLAS. By constructing a graph of interconnected calorimeter cells to represent the electromagnetic shower, we use GNNs to predict the photons' identity, direction, and energy. This project is currently under active development, with the goal of demonstrating superior performance of GNN-based photon measurements and integrating this approach into the ATLAS analysis workflow for Run 3 measurements and photon-involving searches.

Generative machine learning is a rapidly evolving field with immense potential to enhance computational efficiency in detector simulation and introduce new approaches for modeling detector performance and physics processes in particle physics experiments. Our team has been actively developing novel generative machine learning applications for the ATLAS experiment.

Recently, we developed a normalizing flow-based architecture to model correlated and non-symmetric detector responses to photon measurements with the ATLAS detector. This innovation enables the creation of ultra-high statistics background Monte Carlo samples for modeling the background invariant mass distribution in the Higgs-to-diphoton measurement. The normalizing flow architecture serves as a surrogate for detector simulation, bypassing computationally intensive detector response simulations. This approach was successfully applied in the first Run-3 ATLAS publication, which reported the fiducial cross-section measurement in the H to diphoton and H to 4-lepton channels.

Currently, our generative machine learning research program is supported by the DOE's inaugural AI for HEP grant. We are actively exploring novel architectures and investigating the use of generative machine learning as surrogates for modeling collision processes and detector responses.

This project is ongoing. No public information is available.

The full potential of the LHC physics program will be realized with the High Luminosity LHC (HL-LHC) program, scheduled to run from 2029 to 2040. The HL-LHC will enable high-precision measurements of the Higgs boson, including the possibility of observing Higgs pair production, which would offer insights into the Higgs boson’s self-interaction. It will also greatly extend experimental coverage for signals from "natural" supersymmetric models. To support the physics goals of the HL-LHC, the ATLAS detector is undergoing a "Phase-2" upgrade, an international program with significant contributions from U.S. institutions, funded by the Department of Energy (DOE) and the National Science Foundation (NSF). Contributions from faculty, laboratory scientific staff, postdoctoral researchers, and graduate students are key to the success of this upgrade. Our research group collaborates closely with scientists at LBNL and other institutions to contribute to the Inner Tracker (ITk) upgrade, the largest component of the overall Phase-2 upgrade.

Our group has played a significant role in the ATLAS Inner Tracker (ITk) upgrade, focusing on radiation tolerance, quality control, and data acquisition software development. We developed specialized software and tools to assess the radiation hardness of the Strip tracker frontend ASICs, including testing at the TRIUMF facility and analyzing Single Event Upset effects. Additionally, we contributed to quality control efforts at LBNL, ensuring the production of approximately 2,000 Strip tracker barrel modules and developing tests for Strip module powerboards and pixel module construction. To support ITk integration and commissioning, we have been actively involved in developing the YARR framework for data acquisition, enabling advanced testing of Strip tracker modules and staves to meet the stringent requirements of the HL-LHC program.