28th Conference on Computing in High Energy and Nuclear Physics (CHEP 2026)

Session

Poster

25 May 2026, 15:15

Chulalongkorn University

644. 40 MHz Data Scouting at CMS during LHC Run 3 and beyond

Thomas Owen James (CERN)

Track 2 - Online and real-time computing

Poster Presentation

The Compact Muon Solenoid (CMS) experiment at the CERN LHC has traditionally relied on a highly selective Level-1 trigger to reduce the 40 MHz LHC collision rate to a rate more manageable for data-reading and recording. This selection inherently limits access to event types with large irreducible backgrounds or with unconventional signatures. During LHC Run 3, CMS deployed a novel 40 MHz data...

666. A Blueprint for Emerging Centers: The NNU HEP Farm's Evolution from Local Cluster to Grid Integration

Kai Yi (Nanjing Normal University (CN))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

For many university-based HEP groups, the path to becoming a production-ready WLCG Tier-3 center can seem daunting, often constrained by limited budgets, small teams, and a steep learning curve for grid middleware. This poster presents the evolution of the NNU HEP Farm not just as a site report, but as a practical case study and blueprint for other groups embarking on a similar journey.
We...

305. A Geometric Graph Node Converter for PMT Non Linearity Correction in GRAPES-3 Experiment

Sambit Sarkar (Tata Institute of Fundamental Research Mumbai)

Track 3 - Offline data processing

Poster Presentation

The GRAPES-3 experiment aims to study high-energy cosmic rays through their production mechanisms, propagation, and sources. Located in Ooty at an altitude of 2200 m, it spans an area of 25,000 m$^2$ and comprises about 400 plastic scintillator detectors (SDs) arranged with 8 m spacing to measure the charged component of extensive air showers, along with a dedicated muon detector consisting of...

620. A Graph-Pipeline Framework for ALICE MC Workflows in Run 3 and Beyond

Sandro Christian Wenzel (CERN)

Track 4 - Distributed computing

Poster Presentation

ALICE has undergone a substantial software transformation from Run 2 to Run 3, embracing a message-passing, distributed-computing paradigm that unifies online and offline processing. Building on this shift, we present the Monte Carlo (MC) production framework developed within the O2DPG environment, which orchestrates full Run 3 and Run 4 simulation workflows across the heterogeneous computing...

251. A Lightweight Engine for Automated WLCG Resource Inclusion and Exclusion in ATLAS

Dr Brij Kishor Jashal (Rutherford Appleton Laboratory)

Track 4 - Distributed computing

Poster Presentation

From Probes to Policy: Harmonising ATLAS Resource Health Signals

The operational status of WLCG resources in ATLAS is determined through several parallel mechanisms: probe results and declared downtimes (Switcher), persistent failures in functional or performance tests (HammerCloud), and data transfer or storage exclusion conditions managed by distributed data management (DDM). ATLAS...

340. A Low-Latency Muon-Shower trigger for Long-Lived Particle Signatures

Javier Prado Pico (Universidad de Oviedo (ES))

Track 2 - Online and real-time computing

Poster Presentation

In preparation for the High-Luminosity LHC, the CMS experiment is upgrading its Level-1 Trigger system to handle increased luminosity and pile-up. The new trigger system opens up a plethora of possibilities to detect non-conventional signatures such as those arising from long-lived particles (LLPs). In particular such LLPs may decay far away from the interaction point and decay to hadrons on...

457. A new era for the management of 10,000 Windows devices at CERN

Ben Jones (CERN)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

CERN manages over 10,000 Windows devices – from simulation-heavy workstations to security-hardened desktops and servers critical for accelerator controls. For two decades, this was done with CMF, the CERN-built device management solution. Today, we are gradually moving to mainstream solutions such as Microsoft Intune and Configuration Manager, aiming to leverage industry standard off-the-shelf...

494. A novel library for Monte-Carlo simulation of relativistic nucleus-nucleus collisions systematization

Savva Savenkov (INR RAS, MIPT(NRU))

Track 6 - Software environment and maintainability

Poster Presentation

The integration of diverse high-energy collision Monte Carlo models into a unified simulation workflow is usually time-consuming. This is primarily because these models are conventionally developed as monolithic applications with heterogeneous data input and output formats. As a result, a need for multiple converters and auxiliary scripts arises, which not only impedes the modeling process but...

458. A Path Toward Scalable FPGA Offloading in Federated Research Infrastructures

Daniele Spiga, Diego Ciangottini (INFN, Perugia (IT)), Francesco Brivio (Universita & INFN, Milano-Bicocca (IT)), Giulio Bianchini (Universita e INFN, Perugia (IT)), Massimo Sgaravatto (Universita e INFN, Padova (IT)), Mirko Mariotti (Universita e INFN, Perugia (IT)), Paolo Dini, Simone Gennai (Universita & INFN, Milano-Bicocca (IT))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

In the last years, INFN has consolidated and expanded its distributed computing infrastructure toward heterogeneous hardware systems, also thanks to the ICSC and TeraBIT projects, funded in the context of the Italian National Recovery and Resilience Plan. Among the most innovative components of the national federation are the specialized hardware clusters known as HPC Bubbles, in particular...

623. A Phase-Space Inclusive Figure of Merit based in Optimal Transport for Validating Monte Carlo Reweightings

Lauren Meryl Hay (SUNY Buffalo), Rishabh Jain (Brown University (US))

Track 5 - Event generation and simulation

Poster Presentation

Validating that a full phase-space reweighting of a Monte Carlo prediction preserves the physical fidelity of the underlying model can be challenging, and often relies on comparisons to marginalized 1D histograms of kinematic variables that can mask subtle biases of the original high-dimensional unbinned prediction. In this poster, we present a novel, unbinned approach to comparing the...

360. A Snakemake-Based Workflow for Reproducible Particle Physics Analyses

Oxana Smirnova, Oxana Smirnova (Lund University)

Track 9 - Analysis software and workflows

Poster Presentation

We present a prototype implementation of a particle physics analysis workflow using Snakemake for an ATLAS anomaly detection search. Snakemake provides a flexible and scalable workflow for managing thousands of jobs with complex dependencies, supporting execution both locally and across different HPC environments. The workflow cleanly separates small-scale tasks, such as plotting, histogram...

455. Accelerating charged-particle track reconstruction with deep neural networks for point-cloud segmentation

Minh-Tuan Pham (University of Wisconsin Madison (US))

Track 3 - Offline data processing

Poster Presentation

Charged-particle track reconstruction is an important part of modern collider experiments such as ATLAS and CMS that will face challenging conditions in the future High Luminosity phase of the LHC due to high pile-up. The increasing time and compute costs associated with the current tracking algorithm have spurred the development of machine learning (ML) alternatives to high degrees of...

437. Accelerating Diffusion Models for High-Fidelity Calorimeter Simulation

Anwar Ibrahim

Track 5 - Event generation and simulation

Poster Presentation

Detailed simulation of particle interactions in calorimeters represents a major computational bottleneck for high-energy physics experiments, particularly in the upcoming High-Luminosity LHC (HL-LHC) era. While Generative Adversarial Networks (e.g., CaloGAN) have demonstrated the potential of ML-based fast simulation, they often suffer from mode collapse and limited precision in modeling...

628. Adding a New Facility to IceCube's Long Term Archive

David Schultz (University of Wisconsin-Madison)

Track 1 - Data and metadata organization, management and access

Poster Presentation

The IceCube Neutrino Observatory has accumulated over 15 years of science data, with more years to come. This data has previously been archived in a distributed setup according to accessibility needs and processing level. Trigger-level data is stored at NERSC’s tape system for “online” storage and on physical hard drives stored in a climate controlled room on a shelf for “offline” storage at...

384. Advances in Awareness of Sustainability in the community of Research on Universe and Matter

Stefan Krischer (RWTH Aachen University)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

As global greenhouse gas emissions continue to rise, a significant share originates from growing resource consumption in research environments. Beyond energy use, this includes human resources, infrastructure, equipment, and material life cycles. In research on Universe and Matter, the increasing reliance on large-scale infrastructures and complex digital workflows further amplifies this...

194. Agentic Assistant in the LHCb Real-Time Analysis

Pawel Kopciewicz (CERN)

Track 6 - Software environment and maintainability

Poster Presentation

We present a suite of applications for an agentic chatbot to enhance workflows in the LHCb Real-Time Analysis (RTA). The first presented use case allows experiment operators to request, via natural language on the Mattermost platform, the automated generation of monitoring plots—such as trigger rate or detector temperature versus time—from live or historical subsystem data. This functionality...

245. AI Framework for Constructing High-Quality Knowledge Bases and Agent Applications in Large-Scale Scientific Research

Dr Peng Hu (Institute of High Energy Physics, Chinese Academy of Sciences)

Track 1 - Data and metadata organization, management and access

Poster Presentation

In large-scale scientific research, experimental data faces high acquisition costs and a shortage of high-quality data, while a significant amount of critical data is scattered in unstructured forms across various scientific literature. To address this issue, this study proposes an artificial intelligence framework for constructing high-quality knowledge bases from literature corpora and its...

366. AI Inference and Document Intelligence Service for GSI/FAIR

Alexey Rybalchenko (G)

Track 6 - Software environment and maintainability

Poster Presentation

Large Language Models (LLMs) are transforming software development and data analysis workflows in many fields, including nuclear and particle physics experiments.
However, deploying LLMs in production research environments requires careful attention to scalability, security, and resource efficiency.
This work presents a versatile production-grade LLM inference and document intelligence...

484. AI-Driven Operations and Digital Twin for the LHCb Data Centre

Pierfrancesco Cifra (CERN)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

In the exabyte era, physical science research infrastructures will have to deal with massive quantities of raw data by relying on large heterogeneous computing facilities. In the LHCb context, the ODISSEE project aims to maximize the computational performance and reliability of those systems while reducing the required energy and the total cost of ownership by using AI tools and techniques. By...

383. AI_INFN Platform - a distributed Virtual Research Environment

Rosa Petrini (Universita e INFN, Firenze (IT))

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

Over the past decade, Machine Learning and Artificial Intelligence technologies have evolved at an extraordinary pace, making collaboration among geographically distributed experts and students more critical than ever.
The AI_INFN Platform is designed to play a key role in providing access to hardware accelerators for research communities in both fundamental and applied physics.
The platform...

609. An XRootD Agent as A Unified Storage Interface for INK Platform

Prof. Haibo Li (Institute of High Energy Physics, Chinese Academy of Sciences), Xuantong Zhang (Institute of High Enegry Physics, Chinese Academy of Sciences (CN)), Dr Yujiang BI (Institute of High Energy Physics, Chinese Academy of Sciences)

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

With the emergence and continuous evolution of various new development and analysis tools like Jupyter and VSCode, the demand for interactive data analysis has been steadily increasing, leading to significant changes in the traditional high-energy physics data analysis workflow.
To meet the growing and evolving needs of high-energy physics users in data analysis and processing, an all-in-one...

312. An Adaptive Serialization and Compression Optimization Framework Based on Data Feature Perception

Ms 世园 符 (高能物理研究所)

Track 9 - Analysis software and workflows

Poster Presentation

Large scientific facilities such as the High Energy Photon Source (HEPS) generate massive volumes of heterogeneous experimental data during operation. These data exhibit remarkable diversity in terms of scale, structure, and distribution characteristics, imposing extremely high requirements on the real-time response capability and long-term archive storage efficiency of data processing...

191. An Automated Data Processing Workflow Platform for Space Astronomy Satellites

Shiyuan Li (Nanyang Normal University)

Track 9 - Analysis software and workflows

Poster Presentation

Space astronomy satellites serve as critical infrastructure in the field of astrophysics, and data processing is one of the most essential processes for conducting scientific research. The Institute of High Energy Physics (IHEP) of the Chinese Academy of Sciences has undertaken the development and construction of multiple space astronomy satellites, including HXMT, GECAM, SVOM, eXTP and CATCH....

329. An IaC-Framework for the future of CERN’s Infrastructure and Configuration Management

Ben Jones (CERN)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

CERN operates a large and distributed computing environment in which provisioning, configuration, and operational state are handled by different systems. Since 2012, the IT department has invested heavily in bridging these areas under the Agile Infrastructure project. Open-source projects such as OpenStack, Puppet, and Foreman have been integrated with in-house services to offer a cohesive...

171. An online data filtering and monitoring kit with ERSAP and streaming CODA

Torri Jeske (Jefferson Lab)

Track 2 - Online and real-time computing

Poster Presentation

At Jefferson Lab, the CEBAF Online Data Acquisition (CODA) kit and the commonly used front end electronics modules are recently upgraded to support streaming readout data acquisition (DAQ). Depending on the use case, the streaming DAQ data may consist primarily of empty time frames during cosmic runs, or it may be dominated by background signals. A tool kit that applies user-defined online...

155. Anomaly Detection in dCache Billing Data

Christian Voss, Marina Sahakyan, Mr Tigran Mkrtchyan (DESY)

Track 1 - Data and metadata organization, management and access

Poster Presentation

The dCache project provides an open-source, highly scalable distributed storage system deployed at numerous laboratories worldwide. Its modular architecture supports high-rate data ingestion, WAN data distribution, efficient HPC access, and long-term archival storage. Although initially developed for high-energy physics, dCache now serves a broad range of scientific communities with diverse...

402. Application of Unity-based Visualization in HEP experiments

Minghua Liao (Sun Yat-Sen University (CN))

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

Visualization tools are used to display the detector geometry and information of event hits. It plays an important role in physics analysis, data quality monitoring, algorithm optimization, physics education, and public outreach. Unity, as a powerful game engine, exhibits advantages such as high-performance rendering, multi-platform support, and rich tools and features, making it suitable for...

149. Attention-Driven Context Enrichment for Stabilized Real-Time Jet Tagging

Tarik Ourida

Track 2 - Online and real-time computing

Poster Presentation

Standard Level-1 trigger algorithms treat collision events as statistically independent, a design choice that simplifies implementation but prevents models from leveraging short term variations in detector performance. These fluctuations can transiently distort reconstructed features and weaken the stability of fast classification algorithms. To address this limitation, we introduce Context...

648. Automated Jet Energy Correction for Prompt Data in the CMS experiment

Patin Inkaew (Helsinki Institute of Physics (FI)), Punnawich Chokeprasert (Helsinki Institute of Physics (FI)), Theodoros Chatzistavrou (National Technical Univ. of Athens (GR))

Track 2 - Online and real-time computing

Poster Presentation

The LHC experiments have so far calibrated and re-reconstructed data typically years after the end of data-taking to make them usable for precision physics analyses, costing millions of CPU hours. This approach becomes untenable at the HL-LHC, with 10 times larger datasets. Jet energy corrections (JEC) are among the dominant sources of systematic uncertainty in many physics analyses and...

500. Background jobs for recovering CPU usage during draining phases: ATLAS use case

Sebastian Wozniewski (Georg August Universitaet Goettingen (DE))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

On batch systems with many jobs sharing a worker node, the draining of a node in order to terminate it for operational purposes without job abortions leads to idle CPU cores and a loss of compute time. This is becoming a prominent issue at German university-based Tier-2 centres, in particular. Towards the High-Luminosity LHC, they are undergoing a transformation and CPU will be provided via...

392. Benchmarking a PureStorage S500 Appliance for Physics Workflows

Gianmaria Del Monte (CERN)

Track 1 - Data and metadata organization, management and access

Poster Presentation

The continuous growth in data volumes and diversification of access patterns in high-energy physics (HEP) are gaining attention in storage systems that offer both extreme performance and ease of use. To explore the potential of modern flash technologies for scientific workloads, we conducted a comprehensive benchmarking campaign for a PureStorage all-flash appliance, focusing on its...

53. CERN STEAM Academy: Growing the future Trigger, DAQ, Software & ML experts

Anna Kravchenko (CERN)

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

The High-Luminosity LHC era is pushing experiments toward more complex software stacks, high throughput data processing, heterogeneous computing architectures, DAQ, and AI real-time decision making. To strengthen community capacity for next-generation trigger and data-processing systems, we present the CERN STEAM Academy: a 10-week, hands-on programme hosted at CERN developed within the Next...

287. CernVM-FS Unchained: An Update to Container Unpacking and Distribution with CVMFS

Valentin Volkl (CERN)

Track 4 - Distributed computing

Poster Presentation

The CernVM-Filesystem (CVMFS) is a global, read-only, on-demand filesystem optimized for software distribution. CVMFS is also a very efficient way of distributing container images and can be used with container runtimes such as Apptainer or Containerd to lazy-load images. The unpacked.cern.ch repository at CERN, a service that allows users to publish container images to CVMFS has become one of...

540. Charged Particle Tracking with Reinforcement Learning for Drift Chambers

Yao Zhang

Track 3 - Offline data processing

Poster Presentation

Charged particle tracking is a critical task for physics analysis. In this work, we propose applying reinforcement learning (RL) for reconstructing particle trajectories in drift chambers. Our designed workflow uses the output of a graph neural network (GNN) as the observation for RL. Agent training employs a reward metric derived from Monte Carlo truth information, with the objective of...

330. Closing the VM Capacity Gap: Native and Container Solutions to Improve Volunteer Computing Throughput

Lorenzo Valentini (CERN)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

For over two decades, the LHC@home volunteer computing project has provided additional opportunistic computing capacity to support the scientific research conducted at CERN. With the retirement of the SixTrack application, the only natively executable one, there has been a significant reduction in job throughput. This paper highlights the difference in job throughput between the native and...

529. CMS Site Monitoring and new Hammercloud Setup

Chan-anun Rungphitakchai (Chulalongkorn University (TH))

Track 4 - Distributed computing

Poster Presentation

The CMS collaboration operates a large distributed computing infrastructure to meet the computing requirements of the experiment. About half a million CPU cores and an exabyte of storage are utilized to reconstruct the recorded data, simulate signals of physics processes, and analyze data. Computing resources are located at about one hundred sites around the world.

Monitoring the...

56. CMS Storage Performance with RNTuple

Eric Vaandering (Fermi National Accelerator Lab. (US))

Track 3 - Offline data processing

Poster Presentation

CMS is transitioning to use ROOT’s new RNTuple data storage format for the files CMS will write in the HL-LHC era. Based on initial tests, CMS expects faster I/O and smaller files compared to the present TTree storage format. This contribution will show a comprehensive performance comparison between RNTuple and TTree I/O using CMS AOD and MiniAOD data formats as test cases for both simulation...

631. Coffea schemas modifications

Maksym Naumchyk, Maksym Naumchyk (Princeton University (US))

Track 6 - Software environment and maintainability

Poster Presentation

This presentation is about my recent project as an IRIS-HEP fellow. I was working on improving the Coffea 'schemas' by simplifying how they work internally. It eventually transitioned into making a new package that would include all the simplified schemas, separated from Coffea. Eventually Coffea will use them instead of its old schemas. This new package was given the name Zipper and...

503. CRISP: A Collaborative Research Information Sharing Platform

Eric Lancon (Brookhaven National Laboratory (US))

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

The Collaborative Research Information Sharing Platform (CRISP) provides an integrated system for managing scientific collaboration workflows, documentation, and institutional knowledge for the future Electron Ion Collider (EIC). CRISP is designed to address practical challenges in coordinating activities across a multi-thousand-user community distributed international collaboration, using a...

404. Data quality monitoring system in the JUNO experiment

Jize Yang (Sun Yat-Sen University)

Track 2 - Online and real-time computing

Poster Presentation

Jiangmen Underground Neutrino Observatory (JUNO) is a large neutrino experiment located in southern China, aiming at determining neutrino mass ordering, as well as other neutrino physics topics. JUNO has completed detector commission and started data taking since Aug. 22, 2025. Data quality monitoring (DQM) system is critical for data taking，data quality control, and data analysis in any high...

367. Data-Driven Validation of HS23 with CMS Grid Job Monitoring Data

Robin Hofsaess

Track 7 - Computing infrastructure and sustainability

Poster Presentation

With this contribution, a data-driven method for the performance comparison of Grid sites is presented.
While the WLCG sites with an MoU typically report their performance in HS23, opportunistic sites, such as HPC or Tier-3 centers, usually don't.
For the comparison of opportunistically used HPC clusters in Germany, a method was developed to asses the performance of these sites based on CMS...

244. Deep Learning–Based Multi-Photon Reconstruction for \(K^0_s\) Events in LHCf

Mr Andrea Paccagnella

Track 3 - Offline data processing

Poster Presentation

The LHCf experiment measures forward neutral particle production at the LHC, providing key inputs for the tuning of hadronic interaction models used in ultra-high-energy cosmic ray physics. The reconstruction of multi-photon final states in forward experimets represents a challenging offline computing problem, due to overlapping showers, non-uniform detector response, and strong correlations...

62. Demonstrating the Processing Chain for the Next Generation Triggers in the CMS Experiment

Dr Mateusz Zarucki (CERN)

Track 2 - Online and real-time computing

Poster Presentation

The Next Generation Triggers (NGT) R3 (Real-time Reconstruction Revolution) project in CMS aims to rethink the experiment’s data acquisition system, allowing its physics programme to process all collisions accepted by the Level-1 hardware-based trigger system (L1T), in view of the Phase-2 upgrade for the HL-LHC. Its main objective is to expand the High-Level Trigger (HLT) data scouting...

583. Deploying the new LHCb simulation framework based on Gaussino for distributed production of simulated samples

Wojciech Krupa (CERN)

Track 5 - Event generation and simulation

Poster Presentation

The Gauss software is the main simulation framework in LHCb and handles both the event generation step and the tracking of particles through the detector material. Gauss has recently been restructured as a thin LHCb-specific software layer above an experiment-independent HEP simulation framework (Gaussino). In this talk we report on the steps that were taken toward the deployement and...

87. Designing a robust and modular quality control framework for the construction of the SiPM-on-Tile section of the CMS High Granularity Calorimeter upgrade

Anurag Sritharan (Deutsches Elektronen-Synchrotron (DE))

Track 6 - Software environment and maintainability

Poster Presentation

The CMS experiment will upgrade its detectors to cope with higher luminosities and collision rates during the High-Luminosity era of the LHC. One key upgrade of the CMS is the High Granularity Calorimeter (HGCAL), which will completely replace the current end-cap calorimeter. The hadronic calorimeter is split into two sections using different technologies, depending on the expected amount of...

403. Detector identifier and geometry service in JUNO offline software

Yuning Su (Sun Yat-Sen University (CN))

Track 5 - Event generation and simulation

Poster Presentation

Detector identifier and geometry management system plays important role in the offline software of every nuclear and particle physics experiment. Jiangmen Underground Neutrino Observatory~(JUNO), a large neutrino experiment starting design in 2013, has completed detector construction and begins data taking in 2025. We will describe the design and implementation of the JUNO detector identifier...

511. Developing reusable systems in the Glance project at CERN

Pedro Glaser De Senna (Federal University of Rio de Janeiro (BR)), Pedro Glaser de Senna

Track 6 - Software environment and maintainability

Poster Presentation

Developing systems with reusability in mind is often a challenge. Even when a common context for system deployment is identified, some groundwork is required before it can be adopted by different teams. The Glance project at CERN addresses this challenge by implementing modular development and reuse across over 20 systems spanning four experiments: ALICE, ATLAS, CMS and LHCb. Originally...

405. Development of the JUNO-TAO event display based on PHOENIX

Xuesen Wang

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

Taishan Anti-neutrino Observatory (TAO) is a satellite experiment of Jiangmen Underground Neutrino Observatory (JUNO). It is located near the Taishan nuclear power plant (NPP) to monitor the neutrinos emitted from the NPP.
Event display is a critical tool in High Energy Physics (HEP) experiments. It helps monitoring data taking, data quality control, event simulation, reconstruction, and...

361. Development of the Metadex platform for DUNE conditions metadata management.

Mike Clymer (Colorado State University (US))

Track 1 - Data and metadata organization, management and access

Poster Presentation

DUNE is a next-generation neutrino oscillation experiment. During its decades-long operational lifetime, it is expected that many exabytes of data will be collected. It is critical that this data be correctly characterized with respect to its associated conditions metadata – the non-event data used to process event data during reconstruction and analysis. To meet the operational scale and...

217. Differentiable Everything: A JAX Ecosystem for Likelihood-Based Inference in HEP

Mohamed Aly (Princeton University (US))

Track 9 - Analysis software and workflows

Poster Presentation

The JAX framework provides automatic differentiation, JIT compilation, vectorization, and multi-hardware acceleration well-suited for statistical inference in HEP. In this contribution, we present an ecosystem of interoperable tools that leverage the power of JAX, with a focus on everwillow, an inference tool agnostic to the underlying statistical model. At the modelling layer of this...

364. Diffusion Models for ECAL Simulation in the LHCb experiment

Anwar Ibrahim

Track 5 - Event generation and simulation

Poster Presentation

In this work, we investigate diffusion-based generative models as a fast simulation alternative for modeling detector response on the example of the electromagnetic calorimeter response for the LHCb experiment. We consider both classical denoising diffusion probabilistic models with Gaussian noise and their extension based on Gamma-distributed noise, which is expected to be better suited for...

420. Digital Twin for Server Network Analysis with LLM-Driven RAG Integration

Mr Andrey Shevel (Petersburg Nuclear Physics Institute named by B.P. Konstantinov of National Research Centre «Kurchatov Institute» (NRC «Kurchatov Institute» - PNPI))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

Traditional server network monitoring relies on specialized tools and complex queries, demanding significant domain expertise and being time-consuming. We propose a Digital Twin (DT) framework that provides a real-time, unified model of network behavior, enabling intuitive natural-language interactions powered by large language models (LLMs).
The DT fuses live telemetry from monitoring...

83. Driving the CMS Submission Infrastructure Towards the HL-LHC Scales

Antonio Perez-Calero Yzquierdo (Centro de Investigaciones Energéticas Medioambientales y Tecnológicas)

Track 4 - Distributed computing

Poster Presentation

The CMS Submission Infrastructure (SI) provisions and orchestrates the compute resources used for CMS data processing, simulation, and analysis. While the SI has reliably supported Run-3 operations at scales of several hundred thousand concurrent jobs across Grid, HPC, and cloud sites, the computational demands of the HL-LHC era require a substantially more scalable and robust system. To...

508. DUNE Rucio Server Scalability Studies

Jacob Calcutt (Brookhaven National Laboratory (US))

Track 1 - Data and metadata organization, management and access

Poster Presentation

The DUNE collaboration has an ongoing production effort to simulate the full detectors and to analyze the various prototypes that are currently running. Rucio is used to manage the 40PB of files made to date. When 500 or more jobs were sending output to Rucio simultaneously via Rucio upload, we observed timeouts, unhandled exceptions, and Rucio server restarts due to slow performance. In...

606. Efficient Statistical Modeling for Particle Physics Using Computational Graphs in Python

Dr Giordon Holtsberg Stark (University of California,Santa Cruz (US))

Track 9 - Analysis software and workflows

Poster Presentation

Statistical modeling is central to discovery in particle physics, yet the tools commonly used to define, share, and evaluate these models are often complex, fragmented, or tightly coupled to legacy systems. In parallel, the scientific Python community has developed a variety of statistical modeling tools that have been widely adopted for their performance and ease of use, but remain...

453. Enabling Accessibility to CERN audiovisual content via Automated Speech Recognition

Pablo Saiz (CERN)

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

Diversity awareness requires that we provide all CERN-made multi-media content with subtitles and make them fully searchable, addressing in particular the needs of persons with impairments and speakers of foreign languages. The goal of the “Transcription and Translation as a Service” (TTaas) software [1] is to deliver a performant, privacy-preserving and cost-efficient Automated Speech...

291. Enhancement and Validation of Hadronic De-excitation Models in Geant4

Nikita Chalyi (Tomsk State University (TSU))

Track 5 - Event generation and simulation

Poster Presentation

In this work, we describe enhancements to the hadronic de-excitation models implemented in the Geant4 toolkit. We extend the comprehensive and independent validation system for these models, covering a wide range of tests in the moderate energy region, from the reaction threshold to 3 GeV. The underlying processes have a defining impact on the formation of hadronic showers and the resulting...

28. Enhancing ATLAS ITk Outer Barrel Loading Workflows through Guided Procedures and Automated Database Interactions with CLApp

Dr Maximilian Horzela (Georg August Universitaet Goettingen (DE))

Track 1 - Data and metadata organization, management and access

Poster Presentation

The future Inner Tracker (ITk) of the ATLAS experiment will replace the current Inner Detector to maintain excellent tracking and vertexing performance under the challenging conditions of the High-Luminosity LHC (HL-LHC). It must withstand significantly increased radiation levels and occupancy while handling higher data rates and extending forward coverage. At the same time, with more than 150...

375. Enhancing Secure Remote Access at CERN: Implementing Two-Factor Authentication for Windows Terminal Infrastructure

Ben Jones (CERN)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

For several years, CERN has provided access to Windows remote desktops through its Windows Terminal Infrastructure service. As the need for stronger security measures grew, particularly around connections using Microsoft’s Remote Desktop Protocol, we began exploring ways to integrate Two-Factor Authentication (2FA) into this critical service. This presented unique challenges in CERN’s academic...

646. Enhancing the ME0 Level-1 Trigger in the CMS Experiment Using Machine Learning

Woohyeon Heo (University of Seoul, Department of Physics (KR))

Track 2 - Online and real-time computing

Poster Presentation

The ME0 Gas Electron Multiplier (GEM) detector systems will be installed for the phase-2 upgrade of the Compact Muon Solenoid (CMS) experiment in the Large Hadron Collider (LHC). The ME0 detectors, located in each endcap of the muon system, are the only muon detectors that cover the range 2.4 < |eta| < 2.8. Due to the high background environment, keeping the trigger rate low while maintaining...

166. Enhancing the Performance of the SCAB SuperSearch in the ATLAS Speakers System

Leonardo Mira Marins (Federal University of Rio de Janeiro (BR))

Track 6 - Software environment and maintainability

Poster Presentation

The European Organization for Nuclear Research (CERN), home to the Large Hadron Collider, hosts one of the world’s largest particle physics experiments, the ATLAS experiment. To effectively support administration, workflow management, and scientific communication within ATLAS, the Glance project was established in 2003 to provide web-based automated solutions for membership, analysis tracking,...

433. Ensuring Long-Term Data Sustainability: Evaluating a Migration to Rucio for the LZ Dark Matter Experiment

Tyler Anderson (LBNL)

Track 1 - Data and metadata organization, management and access

Poster Presentation

Long-running high energy physics experiments often depend on legacy architectures for orchestrating their data. While these custom tools can be effective, the expertise to maintain them is often concentrated in limited personnel, which raises concerns of software sustainability and long-term maintenance. Transitioning to a community-supported standard like Rucio, created at CERN, offers a...

352. EveNet: A Foundation Model for Particle Collision Data Analysis

Ting-Hsiang Hsu (National Taiwan University (TW))

Track 9 - Analysis software and workflows

Poster Presentation

Foundation models are large neural networks pretrained on vast datasets and adapted to many downstream tasks with minimal task-specific training. In high-energy physics, precise Monte Carlo event generators allow the simulation of billions of events, but the enormous space of beyond-Standard-Model scenarios makes training specialized large models for each analysis computationally impractical....

125. Expanding CloudVeneto: seamless integration of High Performance Computing resources for enhanced scientific collaboration

Giovanni Zago (Universita e INFN, Padova (IT))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

CloudVeneto is a private cloud targeted to scientific communities, based on OpenStack software, designed in 2013 to support INFN projects, initially mostly Nuclear Physics and HEP ones. During the last 12 years it evolved by integrating resources and use cases of several Departments of the University of Padova. It currently supports several scientific disciplines of different domains, but it...

60. Extending the reconstruction of displaced tracks at CMS during the HL-LHC with Line Segment Tracking

Manos Vourliotis (Univ. of California San Diego (US))

Track 2 - Online and real-time computing

Poster Presentation

The upgrade of the Large Hadron Collider (LHC) to the High-Luminosity LHC (HL-LHC) will increase the number of proton-proton collisions by several-fold, and thus place a large demand on computing resources for charged particle tracking. The Line Segment Tracking (LST) algorithm is a novel, highly parallelizable algorithm that can run efficiently on GPUs and has been integrated into the CMS...

574. FairRoot an experiment independent framework for simulation, analysis and reconstruction

Florian Uhlig (GSI - Helmholtzzentrum fur Schwerionenforschung GmbH (DE))

Track 9 - Analysis software and workflows

Poster Presentation

FairRoot is a framework for simulation, reconstruction, and analysis of nuclear and high energy physics experiments. It provides the necessary building blocks that allow users to easily implement their specific experimental setup. Originally started as a project at GSI focused on a specific experiment, FairRoot has evolved into a widely used platform by various experiments worldwide,...

542. Fast simulation for scattering muography applications using generative adversarial neural networks

Javier Prado Pico (Universidad de Oviedo (ES))

Track 5 - Event generation and simulation

Poster Presentation

Muography is an emergent non-destructive testing technique that uses cosmic muons to probe the interior of objects and structures. This technique can be employed to perform preventive maintenance of critical equipment in the industry in order to test the structural integrity of the facility. Several muography imaging algorithms based on machine learning methods are being developed in the...

469. Fast Simulation Of FARICH Detector Using Generative Adversarial Networks

Dr Alexey Boldyrev

Track 5 - Event generation and simulation

Poster Presentation

Focusing Aerogel Ring Imaging CHerenkov detector (FARICH) is a promising particle identification technology for the SPD expertiment. A free-running (triggerless) data acquisition pipeline to be employed in the SPD results in a high data rate necessitating new approaches to event generation and simulation of detector responses. In this work, we propose several machine learning based approaches...

328. Federation of compute and storage resources for SPD offline computing

Mr Andrey Kirianov (A.Alikhanyan National Science Laboratory (AM))

Track 4 - Distributed computing

Poster Presentation

The Spin Physics Detector (SPD), currently under construction at the NICA complex at JINR, is expected to generate large volumes of data. It is therefore assumed that at least some members of the SPD Collaboration will contribute significant computing and storage resources. Unlike in large-scale grids, the number of participating sites is not so large and most of them will be located in Russia...

591. From database-backed to stateless OAuth2 access tokens in INDIGO IAM

Berk Balci (CERN), Francesco Giacomini (INFN CNAF)

Track 4 - Distributed computing

Poster Presentation

INDIGO IAM is a central Identity and Access Management service for distributed research infrastructures, supporting authentication and authorization at scale. As the number of relying services and users continues to grow, improving the performance and efficiency of IAM operations has become a key objective. One of the most significant performance bottlenecks identified in the current...

351. From Hits to Tracks: A Transformer Approach for Track Reconstruction in Drift Chambers

Yipu Liao (Institute of High Energy Physics, CAS, Beijing)

Track 3 - Offline data processing

Poster Presentation

Denoising and track reconstruction in drift chambers fundamental to particle identification and momentum measurement at electron-positron colliders. While Transformer architectures have revolutionized many sequence-processing domains, their potential for track reconstruction in high-energy physics has not been fully explored. In this work, we introduce Transformer-based methods at two stages...

266. From In-House Tools to HashiCorp Vault: CERN's Transition to Modern Scalable Secrets Management

Ben Jones (CERN)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

CERN's computing infrastructure manages thousands of services across a complex distributed environment, requiring robust secret management for application credentials, root accounts, certificates, and service tokens. This paper explores CERN's transition from puppet-oriented, in-house secrets management solutions to HashiCorp Vault as a centralized, enterprise-level secret management...

602. GlueX-Demon: a compact Python and JSROOT package for interactive data quality monitoring

Dr Naomi Jarvis (Carnegie Mellon University)

Track 3 - Offline data processing

Poster Presentation

GlueX is a hadronic physics photoproduction experiment based at Jefferson Lab. The GlueX spectrometer and beamline detectors include over a dozen individual detectors whose performance and calibrations are mostly independent. During data collection, the data are divided into a series of runs, lasting up to 2 hours each, with the run boundaries acting as calibration boundaries. Data quality...

406. GPT-like transformer model for silicon tracking detector simulation

Tadej Novak (Jozef Stefan Institute (SI))

Track 5 - Event generation and simulation

Poster Presentation

Simulating physics processes and detector responses is essential in high energy physics and represents significant computing costs. Generative machine learning has been demonstrated to be potentially powerful in accelerating simulations, outperforming traditional fast simulation methods. The efforts have focused primarily on calorimeters.

This contribution presents the very first studies on...

447. GPU-Accelerated Simulation of 3D Diamond Sensors Using the TeRABIT Infrastructure

Rosa Petrini (Universita e INFN, Firenze (IT))

Track 5 - Event generation and simulation

Poster Presentation

Diamond detectors with laser-graphitized electrodes orthogonal to the surface are emerging as fast, full-carbon sensors for applications ranging from High Energy Physics to Nuclear Medicine. Recent advances in low-resistance electrode fabrication have enabled sub-100 ps timing performance. However, accurately modeling signal formation remains challenging due to the intertwined effects of...

388. GUNTAM: A Transformer for seed reconstruction

Minh-Tuan Pham (University of Wisconsin Madison (US))

Track 3 - Offline data processing

Poster Presentation

Reconstructing particle trajectories is a significant challenge in most particle physics experiments and a major consumer of CPU resources. It can typically be divided into three steps: seeding, track finding, and track fitting. Seeding involves identifying potential trajectory candidates, while track finding entails associating detected hits with the corresponding particle. Finally, track...

566. Harmonizing Relativistic Nuclear Collision Models via Multi-Discriminator GANs

Aleksandr Svetlichnyi (INR RAS, MIPT(NRU))

Track 5 - Event generation and simulation

Poster Presentation

Relativistic heavy-ion collisions serve as a primary tool for investigating the fundamental properties of matter under extreme conditions. The theoretical modeling of these interactions relies on various computational models whose predictive power often fluctuates across different kinematic ranges and physical observables. Furthermore, the underlying complex phenomenological chains are...

41. HEPS Large-Scale Data Challenges and Corresponding Solutions

Siqi Hou

Track 1 - Data and metadata organization, management and access

Poster Presentation

China's High Energy Photon Source (HEPS) will complete facility construction and commissioning by the end of 2025. Data acquisition and analysis have already begun. The 14 beamlines of the first phase of HEPS will generate approximately 300PB of raw data annually, with further expansion expected in the future. This not only poses significant challenges for the reliability and read-write...

78. Improving GlideinWMS Factory Compute Resource Configuration with Automation Tools

Antonio Perez-Calero Yzquierdo (Centro de Investigaciones Energéticas Medioambientales y Tecnológicas)

Track 4 - Distributed computing

Poster Presentation

Sciences such as High Energy Physics, Computational Biology, and other communities use distributed computing facilities to find the solutions to problems that require the execution of computationally intensive algorithms. The Open Science Grid (OSG) enables access to over 100 individual compute clusters spanning the globe for scientists from these disciplines. These sites, primarily at...

154. In-memory compression of Awkward Arrays with Coffea

Manfred Peter Fackeldey (Princeton University (US))

Track 9 - Analysis software and workflows

Poster Presentation

Modern analyses in high-energy physics (HEP) have high memory requirements due to the sheer volume of data collected in experiments at the Large Hadron Collider (LHC) at CERN.
Awkward Array recently released a new version of lazy arrays (“virtual arrays”) that mitigates this problem by loading only the columns required for HEP analysis. Nevertheless, these columns can still add up in size,...

523. Innovations for hardware-level muon trigger for long-lived particles in CMS

Pelayo Leguina (Universidad de Oviedo (ES))

Track 2 - Online and real-time computing

Poster Presentation

Our world has witnessed a massive explosion of data and a surge of machine learning (ML) and AI applications. The result is an ever-increasing need for higher throughput and real-time computing capabilities. The Large Hadron Collider (LHC) and its experiments provide the perfect benchmark to bring the recent industry developments and explore beyond-the-state-of-the-art technologies to process...

165. Integrating the ATLAS Glance systems to CERN’s Group Management API

Rafaella Lenzi Romano, Rafaella Lenzi Romano (Federal University of Rio de Janeiro (BR))

Track 6 - Software environment and maintainability

Poster Presentation

The ATLAS experiment involves over 6,000 members, including students, physicists, engineers, and researchers. This dynamic CERN environment brings up some challenges, such as information centralisation, communication, and the continuity of workflows. To overcome these challenges, the ATLAS Glance Team has developed and maintained several automated systems that rely on CERN’s Group Management...

627. Integrating User Data Tokens in IceCube's Production Workflow Manager

David Schultz (University of Wisconsin-Madison)

Track 4 - Distributed computing

Poster Presentation

As part of the IceCube Neutrino Observatory's move to the Pelican Platform for data transfer, our production workflow management tools also needed to be updated. There were two major changes happening at the same time: moving from X.509 certificates to tokens, and gathering the tokens at the initial dataset submission rather than during the job processing. Some significant problems had to be...

157. Integration of JupyterHub with the Slurm cluster at GSI

Florian Uhlig (GSI - Helmholtzzentrum fur Schwerionenforschung GmbH (DE))

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

Jupyter is a powerful tool for data visualisation and interactive analysis with Python, and in particular JupyterHub offers a simplified way for users to run their workflows on dedicated HPC hardware. The use of JupyterHub is already widespread among many research centres and computing clusters. However, many of the existing deployments rely on specialised network setups such as a dedicated...

139. IntelliRTA: a dashboard to analyse trends and correlations for the LHCb triggers

Pawel Kopciewicz (CERN)

Track 2 - Online and real-time computing

Poster Presentation

The LHCb experiment in Run 3 features a full software trigger: the GPU-based HLT1 with O(100) and the CPU-based HLT2 with O(4000) trigger lines. Human control of every aspect of data quality in a complex system of this scale is extremely difficult and requires a high degree of automation. IntelliRTA is a monitoring dashboard that provides a holistic view of the trigger lines and the data...

398. Introducing ARES: the LHCb Analysis platform for Radiation Environment Simulations

Ryunosuke O'Neil (CERN)

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

High Energy Physics (HEP) experiments increasingly rely on large volumes of Monte Carlo (MC) simulation data to estimate radiation levels and activation scenarios. Within the LHCb collaboration, we present a new system developed to simplify the management and exploration of such MC simulation outputs as obtained with the FLUKA code: the Analysis platform for Radiation Environment Simulations...

45. IPVLAN Container Manager: A High-Performance, Fault-Tolerant Orchestration Framework for Distributed Scientific Computing

Dr Geonmo Ryu (Korea Institute of Science & Technology Information (KR))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

In small-scale scientific infrastructures typically consisting of 3–7 nodes, industry-standard orchestrators like Kubernetes often introduce an "operational gap" due to their resource-heavy control planes. Furthermore, traditional overlay networks such as VXLAN introduce significant latency and CPU overhead, which hinders the performance of data-intensive distributed scientific computing....

49. JobLens: A Lightweight Job Observability Collector for High-Throughput HEP Computing

Mr Zhenyuan Wang (Computing center, Institute of High Energy Physics, CAS, China)

Track 4 - Distributed computing

Poster Presentation

With the escalating processing demands of modern high-energy physics experiments, traditional monitoring tools are faltering under the dual pressures of cumbersome deployment and coarse-grained observability in high-throughput production environments. JobLens is a lightweight, one-click-deployable data collector designed to deliver fine-grained, job-level observability for HEP workloads. Its...

199. JUNO Raw Data Preprocessing

Tao Lin (Chinese Academy of Sciences (CN)), Dr Tao Lin

Track 3 - Offline data processing

Poster Presentation

The Jiangmen Underground Neutrino Observatory (JUNO) is a large-scale neutrino experiment with multiple physics goals. After its completion at the end of 2024, commissioning for data taking began, followed by the commencement of official data-taking on August 26, 2025. The raw data acquired by the JUNO DAQ system is stored in a custom binary format. After transmission to the data center, this...

326. Leveraging Visualization for Enhanced Physics Analysis in BESIII

Zhijun Li (Sun Yat-Sen University (CN))

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

In high energy physics experiments, visualization plays a crucial role in detector design, data quality monitoring, offline data processing, and has great potential for improving physics analysis. In addition to traditional physics data analysis based on statistical methods, visualization offers unique intuitive advantages in the search for rare signal events and in reducing background noise....

466. LGT4HEP: Computing Traineeship in Lattice Gauge Theory for High-Energy Physics

Huey-Wen Lin

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

LGT4HEP (High-Energy Physics Computing Traineeship for Lattice Gauge Theory) is a multidisciplinary training initiative designed to prepare the next generation of researchers in computational lattice field theory and high-performance computing. The program emphasizes rigorous coursework, including lattice QCD and advanced computational methods, paired with hands-on experience on...

333. Machine Learning Approaches for Mass Composition Estimation in Ultra-High-Energy Cosmic Rays at the Pierre Auger Observatory

Daniele Martello (Università del Salento & INFN Lecce)

Track 3 - Offline data processing

Poster Presentation

The Pierre Auger Observatory collects vast amounts of complex spatial-temporal data from extensive air showers induced by ultra-high-energy cosmic rays (UHECRs), i.e., those with energies above 10^18 eV. Determining the mass composition of the primary particle is a key challenge, as direct measurements are impossible and traditional analytical methods struggle with the complexity of shower...

175. Machine Learning-Based Enhanced Soft Muon Tagging in Atlas

Federico Andrea Corchia (Universita e INFN, Bologna (IT))

Track 3 - Offline data processing

Poster Presentation

Identification (“tagging") of hadronic jets associated with charm and bottom quarks is crucial for many experimental signatures explored with the ATLAS detector at the LHC. Soft Muon Tagging (SMT) is a tagging technique based on the identification of muons from b/c -> mu + X within hadronic jets, complementary to other jet-based algorithms. With the SMT algorithm, muons can be used as a proxy...

233. MC tuning from the commissioning to the physics data taking in JUNO

Peidong Yu (IHEP)

Track 5 - Event generation and simulation

Poster Presentation

The Jiangmen Underground Neutrino Observatory (JUNO) is a multi-purpose experiment featuring a 20,000-ton liquid scintillator central detector, water cerenkov detector and top track, primarily designed to determine neutrino mass ordering. Following detector construction completion in late 2024, the detector was successively filled with ultrapure water and then liquid scintillator (LS). As LS...

38. Migrating INSPIREHEP Document Harvesting and Curation Workflows to Apache Airflow

Ioannis Tsanaktsidis (CERN)

Track 6 - Software environment and maintainability

Poster Presentation

The continuous ingestion of scientific documents from external sources into INSPIREHEP created challenges in scalability, transparency, and long term maintenance. This contribution describes the migration of our document harvesting and curation pipeline to the open source workflow orchestrator Apache Airflow. The work involved re-engineering legacy scripts and cron based tasks into modular...

64. mkFit for track fitting with the CMS Phase-2 detector

Aashay Arora (Univ. of California San Diego (US))

Track 3 - Offline data processing

Poster Presentation

The mkFit algorithm offers an implementation of the Kalman filter-based track reconstruction algorithm that exploits both thread- and data-level parallelism. mkFit has been adopted by the CMS collaboration as the main track building algorithm for both the Run-3 offline and online track reconstruction, and it has been to speed up track building by 3.5x on average, while retaining or improving...

504. Modeling Systematic Uncertainty in Latent Representations for Anomaly Detection at Scale

Gaia Grosso (IAIFI, MIT)

Track 9 - Analysis software and workflows

Poster Presentation

Machine-learning-based anomaly detection (AD) offers a promising, model-agnostic alternative to traditional LHC analyses, allowing us to search for many signals at once. Recent AI advances in representation learning motivate the use of neural embeddings to map detector data into low-dimensional latent spaces, preserving critical features ([Metzger et al., Phys. Rev. D 112, 072011 (2025)][1])....

567. Modernizing Financial Processes at CERN through Industry-Standard Software Integration

Catalin Codreanu (Technical University of Cluj-Napoca (RO)), Cristian Schuszter (CERN)

Track 6 - Software environment and maintainability

Poster Presentation

Modern financial operations in large scientific organizations increasingly rely on sustainable, modular, and well-integrated software ecosystems. Over the past years, the FAP-BC group of CERN has focused on modernizing key financial processes by adopting service-oriented approaches, strengthening system integrations, and reducing long-term maintenance costs.

This paper presents recent work...

557. NetDRIVE: NetZero Digital Research Infrastructure Vision and Expertise. A UK perspective on NetZero Research Computing

Caterina Doglioni (The University of Manchester (GB))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

NetDRIVE (NetZero Digital Research Infrastructure Vision and Expertise) [1] is the UK Research and Innovation (UKRI) project developing plans and expertise to tackle NetZero issues around the UK’s government funded research computing or digital research infrastructure (DRI). Following on from the UKRI DRI NetZero Scoping project [2], NetDRIVE is a £4M project spread over the course of 2.5...

363. Neural network decorrelation in the LHCb topological b trigger

James Connaughton (University of Warwick (GB))

Track 2 - Online and real-time computing

Poster Presentation

The LHCb experiment at the LHC employs a fully-software trigger to reconstruct and select events in real time. Key to this approach is the topological beauty (b) trigger, a set of algorithms which select decays of hadrons containing b quarks based on their distinct topology, i.e., highly displaced candidates with a large momentum. For Run 3 of the LHC, these algorithms were reimplemented...

304. New Geant4 internal data base for neutron cross sections.

Andrei Berngardt (Tomsk State University)

Track 5 - Event generation and simulation

Poster Presentation

We present a new method for generating neutron cross-section (XS) data sets from evaluated nuclear data (HP) that improves the accuracy of XS approximation in resonance regions while maintaining computational efficiency for HEP applications in Geant4. Our approach supplements the standard XS datasets with additional dedicated resonance (R) files for the low-energy region, which is defined...

262. Next-Generation of Calibration Management System for ATLAS Tile Calorimeter

David Avetisyan (A.Alikhanyan National Science Laboratory (AM))

Track 1 - Data and metadata organization, management and access

Poster Presentation

Calibration Operations Manager Bot for ATLAS Tile Calorimeter (COMBAT) is the next-generation of the calibration management system developed for the High-Luminosity LHC era. It combines modern AI techniques with a fully asynchronous, scalable architecture to meet the evolving operational demands of the ATLAS experiment, including the transition in the database from COOL to CREST for the...

65. Next-Generation Triggers for CMS at the HL-LHC: Toward Offline-Quality Real-Time Reconstruction

Jessica Prendi (ETH Zurich (CH))

Track 2 - Online and real-time computing

Poster Presentation

The Next-Generation Trigger (NGT) program for the CMS High Level Trigger (HLT) aims at enabling full-rate recording of all the events accepted by the Level-1 Trigger at 750 kHz via a dedicated NGT Scouting stream, performing complete physics event reconstruction with no additional filtering. Reconstructed objects are stored directly in a lightweight NanoAOD format, delivering analysis-ready...

135. On the feasibility of transferring classical database operations to the quantum domain

Carla Sophie Rieger (Technische Universitat Munchen (DE))

Track 1 - Data and metadata organization, management and access

Poster Presentation

Efficient database operations are crucial for processing inherently structured data. We investigate the transfer of classical database operations to their counterparts on uniform quantum superposition states of quantum data. Such data may originate from future experiments that incorporate quantum sensors and quantum memories, or by using quantum encoded classical data. Since quantum states...

612. Operational Evolution of FTS3: A DevOps Driven Approach to elastic operations.

Eric Vaandering (Fermi National Accelerator Lab. (US))

Track 4 - Distributed computing

Poster Presentation

The File Transfer Service (FTS3) is a distributed data movement service developed at CERN and widely used to transfer data across the Worldwide LHC Computing Grid (WLCG). At Fermilab, FTS3 supports data transfers for multiple experiments, including Intensity Frontier experiments such as DUNE, enabling reliable data movement between WebDAV endpoints in Europe and the Americas.
At CHEP 2021, we...

202. Optimizing the LHCb HLT2 selections framework

Daniel Magdalinski (Nikhef)

Track 2 - Online and real-time computing

Poster Presentation

The LHCb experiment operates a full-software trigger comprising of two stages, labelled HLT1 and HLT2. The two stages are separated by a disk buffer, which not only allows the HLT2 processing to be asynchronous with respect to the data taking, it also allows real-time alignment and calibration to be performed prior to HLT2 processing. HLT2 then performs full offline-level reconstruction and...

474. Optimizing XRootD-Ceph plugin performance for RAL disk storage

Alexander Rogovskiy (Rutherford Appleton Laboratory), Jyothish Thomas (STFC)

Track 1 - Data and metadata organization, management and access

Poster Presentation

XRootD-Ceph is a storage plugin that allows one to access Ceph object store via xroot protocol. At RAL, we use this plugin for our disk SE. Although it has proven to be suitable for production-quality high-throughput storage, a few optimizations were needed to ensure optimal performance. In this talk we discuss the evolution of the plugin at RAL.

Some changes to the plugin were dictated by...

550. Overview of the Adapted German ATLAS and CMS Grid Computing Structure

Shawn Gregory Zaleski (Rheinisch Westfaelische Tech. Hoch. (DE))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

As the Large Hadron Collider (LHC) finishes collecting data in Run 3, the outlook for future data collection and analysis will require even more data storage and more powerful and efficient dedicated computing resources as it looks to collect much more data in future runs.
From the beginning of the LHC operation 15 years ago, the Germany ATLAS and CMS groups provided massive dedicated grid...

200. Parallel Acceleration Algorithms and Platform Development for High-Throughput XPCS Data Analysis

JIANLI LIU (Computing Center, Institute of High Energy Physics, Chinese Academy of Sciences)

Track 9 - Analysis software and workflows

Poster Presentation

The X-ray photon correlation spectroscopy (XPCS) retrieves the nanoscale dynamic behaviors of materials by analyzing the photon intensity fluctuations in synchrotron X-ray scattering signals. The multitau algorithm calculates delay times across different temporal scales through a hierarchical binning approach, which not only covers a wide temporal range but also controls computational...

101. Performance Portable Lattice Gauge Theory Simulation with Kokkos

LI Haibo lihaibo

Track 6 - Software environment and maintainability

Poster Presentation

We present a performance‑portable lattice gauge theory simulation library implemented using the Kokkos parallel programming model. The library supports efficient Monte Carlo simulations of SU(N) gauge theories across diverse hardware architectures—including CPUs (via OpenMP and Serial backends), NVIDIA GPUs (CUDA), AMD GPUs (HIP), and Intel GPUs (SYCL)—all from a single source code base. It...

526. Porting the CMS Tracker Raw Data Unpacker to GPUs

Nazar Bartosik (UPO e INFN Torino (IT))

Track 2 - Online and real-time computing

Poster Presentation

When the HL-LHC starts in a few years from now the CMS experiment will be challenged with way more complex proton-proton collision events as well as an increased data logging rate. Present projections suggest that the CPU demands for reconstruction and processing will grow beyond the capacity expected from usual technology progress. Therefore, an effort has been started to optimize software to...

603. PRAETOR - Provenance generation of Astronomical Workflows

Michael Johnson (University of Manchester)

Track 1 - Data and metadata organization, management and access

Poster Presentation

Data volumes and rates of research infrastructures will continue to increase in the upcoming years and impact how we interact with their final data products. Little of the processed data can be directly investigated and most of it will be automatically processed
with as little user interaction as possible. Capturing all necessary information of such processing ensures reproducibility of the...

220. Quantum Computing Approaches to Tracking Parameter Optimization in High Energy Physics

Mr Prajjwal Saxena (Indian Institute of Science), Prajjwal Saxena

Track 3 - Offline data processing

Poster Presentation

The optimization of tracking parameters in particle track reconstruction is a high-dimensional, non-convex problem with significant impact on tracking efficiency, resolution, and computational performance. As detector complexity and pileup increase, conventional heuristic and local optimization methods face scalability limitations. In this work, we will investigate quantum optimization...

144. Quantum Graph Network for Boosted Jet Identification

Parichehr Kangazian Kangazi (The Iranian Ministry of Science, Research and Technology (IR))

Track 3 - Offline data processing

Poster Presentation

Identifying jets originating from the decay of highly boosted heavy particles in colliders plays a crucial
role in uncovering potential signs of physics beyond the Standard Model. Despite significant progress in
jet-origin classification algorithms—particularly graph neural networks—the rapidly increasing volume
of collider data and the demand for faster and more efficient processing...

303. Quantum machine learning and the journey to its scalability

Supanut Thanasilp (Chulalongkorn University)

Track 3 - Offline data processing

Poster Presentation

Quantum systems are well known to create non-classical patterns. The thought that they could also be used to recognize highly complex patterns hidden in data is beyond excitement, leading to the young interdisciplinary field of quantum machine learning (QML). Nevertheless, while a quantum advantage in data analysis can be in principle achieved thanks to the exponentially large Hilbert space,...

465. Rapid Deployment and Plugin-Based Architecture Design of Interactive Analysis Workbench (INK) at IHEP

Jingyan Shi (IHEP)

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

The Interactive Analysis Workbench (INK) is a web-based, open-source interactive computing platform developed at the Institute of High Energy Physics, Chinese Academy of Sciences (IHEP, CAS), to address the growing demands of high-energy physics users for interactive data processing, efficient data access, and collaborative analysis workflows. INK enables users to interactively access IHEP...

77. Real-Time PF Clustering and Deep Learning for Soft-Tau Reconstruction in the CMS Phase-2 L1 Scouting Stream

Valentina Camagni (Università degli Studi e INFN Milano (IT))

Track 2 - Online and real-time computing

Poster Presentation

The CMS Phase-2 Level-1 Trigger (L1T) Scouting program introduces real-time software reconstruction at the full 40 MHz rate, enabling physics analyses directly at trigger level. One of the most promising applications is the reconstruction of low-transverse-momentum (soft) hadronic tau leptons, which are essential for searches for low-mass scalars ϕ → ττ but are poorly reconstructed by existing...

204. Real-Time Training in Hardware Upstream of the Trigger

Bostjan Macek (Jozef Stefan Institute (SI))

Track 2 - Online and real-time computing

Poster Presentation

Future high-energy physics (HEP) experiments operate under extreme real-time constraints, where online filtering and trigger decisions increasingly define the ultimate physics reach. Although machine learning is now widely used in online systems, current deployments are almost exclusively limited to inference with offline-trained models. In this contribution, we investigate a complementary and...

587. Real-Time Trigger Architectures for the CTAO LST Advanced Camera: From Logic to CNNs

Daniel Nieto (IPARCOS-UCM)

Track 2 - Online and real-time computing

Poster Presentation

The Cherenkov Telescope Array Observatory (CTAO) represents the next generation of ground-based gamma-ray telescopes, designed to probe the very-high-energy (VHE) sky above 20 GeV with unprecedented sensitivity. The northern array (CTAO-North) will be composed of an ensemble of Medium-Sized Telescopes (MSTs) and four Large-Sized Telescopes (LSTs), the latter designed to detect the...

613. Remote ROOT Data Access: Separating Client and Server Effects in Nginx and XRootD

Aashay Arora (Univ. of California San Diego (US))

Track 1 - Data and metadata organization, management and access

Poster Presentation

The increasing adoption of columnar data formats and lightweight event representations, such as CMS NanoAOD, has made remote data access a significant factor in the performance of physics analysis workflows. In this context, understanding the performance characteristics of different data serving technologies under realistic network conditions is critical.

This work presents a comparative...

269. Replacing the Legacy Tier-0 Accounting System with Standard Technologies for WLCG Accounting

Ben Jones (CERN)

Track 4 - Distributed computing

Poster Presentation

The WLCG Tier-0 Accounting service provides accounting information for scientific computing resources at CERN (Batch, HPC, and BOINC). The service delivers essential information on CPU and walltime usage, which plays a key role in decision-making and planning processes for CERN resource managers across experiments and departments. It also supplies monthly usage data to the WLCG Accounting...

428. Reproducible Dask workflows in the REANA reusable analyses platform

Tibor Simko (CERN)

Track 9 - Analysis software and workflows

Poster Presentation

Dask is a Python library for scaling Python analysis code from local computers to large data centre clusters. Dask is becoming more popular in astronomy and particle physics communities for carrying out data analyses. We describe how we extended the REANA reproducible analysis platform to support Dask workloads. A special attention was paid to respect the Dask version requested by the analyst,...

294. Research on Key Technologies for Automated Penetration Testing Oriented to High-Energy Physics Institutes

Shiyuan Fu (IHEP, CAS)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

To address the issues of traditional penetration testing being highly labour-intensive, fragmented tooling, and cumbersome processes, this paper aims to design an automated penetration testing system to reduce the manpower and time expenditure during testing.This paper proposes an overall design framework for an automated penetration testing system. The system comprises two components: an...

560. Riemannian-based quantum optimization for quantum machine learning

Kasidit Srimahajariyapong (Chulalongkorn University)

Track 3 - Offline data processing

Poster Presentation

The rapid proliferation of quantum machine learning (QML) has highlighted critical bottlenecks in conventional Variational Quantum Algorithms (VQAs), particularly regarding trainability, scalability, and the absence of rigorous optimal solution guarantees. These challenges motivate us to search for alternative optimization paradigms. In this work, we introduce the Double-Bracket Quantum...

579. Scalable Tape Data Management at INFN-CNAF: Extending GEMSS for Multi‑Library Support and High‑Performance Buffering

Andrea Rendina

Track 1 - Data and metadata organization, management and access

Poster Presentation

INFN-CNAF is the national computing center of the INFN (National Institute for Nuclear Physics), dedicated to research and development in information technologies for subnuclear, nuclear, and astroparticle physics. CNAF hosts the largest INFN data center and operates a WLCG Tier-1 site.

For more than 15 years, tape data management at CNAF has been handled using the Grid Enabled Mass Storage...

545. Scalar-Tensor Extension of Non-Metricity Gravity

AVIK DE (universiti malaya)

Track 9 - Analysis software and workflows

Poster Presentation

\documentclass[11pt, a4paper, twoside]{article}
\usepackage{geometry}
\geometry{top=2.54cm, bottom=2.54cm, left=2.54cm, right=2.54cm}

\title{Scalar-Tensor Extension of Non-Metricity Gravity}
\author{AVIK DE}
%\date{January 2026}

\begin{document}

\maketitle

\begin{abstract}
We present a computation-first study of scalar--tensor extensions of symmetric teleparallel...

151. Scheduling for Next Generation Triggers

Mateusz Jakub Fila (CERN)

Track 2 - Online and real-time computing

Poster Presentation

The Next Generation Trigger (NGT) project at CERN aims to extract more physics information from the High Luminosity LHC data. To achieve this, GPUs and other accelerators are being increasingly adopted in LHC experiments, running both procedural code and AI/ML inferences.

As a result, formerly CPU-only modules in the event reconstruction frameworks now interleave their computations with...

382. Semi-Supervised and Manifold Learning on Multi-Modal Event Representations in the LZ Dark Matter Experiment

John Winnicki

Track 9 - Analysis software and workflows

Poster Presentation

LUX-ZEPLIN (LZ) is a dark matter direct-detection experiment using a dual-phase xenon time projection chamber. The LZ experiment has set world-leading limits on WIMP-nucleon interactions. At low energies, backgrounds built from the spurious pairing of unrelated charge and light signals, also known as accidentals, pose a significant analysis challenge. In this work, we study modern unsupervised...

621. Simulation Studies of a Brute-Force χ² Alignment Framework for the CBM Silicon Tracking System

Nora Bluhme (Goethe University Frankfurt (DE))

Track 3 - Offline data processing

Poster Presentation

The Compressed Baryonic Matter (CBM) experiment at the upcoming Facility for Antiproton and Ion Research (FAIR) will investigate heavy-ion collisions at interaction rates of up to $10^7\, \text{s}^{-1}$.

To fully exploit the intrinsic precision of the tracking detectors, an accurate alignment of all sensor elements is essential. Track-based software alignment determines small but critical...

338. SparKer: sparsity, locality and competition for statistical anomaly detection at scale

Gaia Grosso (IAIFI, MIT)

Track 9 - Analysis software and workflows

Poster Presentation

Modern machine learning has revolutionized our ability to extract rich and versatile data representations across scientific domains. However, the statistical properties of these representations are often poorly controlled, challenging the design of robust downstream anomaly detection (AD) methods.
We identify three principled desiderata for anomaly detection in latent spaces under minimal...

450. Sustainable Open Science with Advanced Metadata: The Einstein Telescope Analysis Portal (ETAP)

Paul James Laycock (Universite de Geneve (CH))

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

The first direct observation of Gravitational Waves (GWs) in 2015, produced by the collision of black holes, instigated a demand for open access to GW data. The Einstein Telescope will increase the detection rate of GWs by a factor of a thousand compared to current detectors, producing information-rich data containing a wealth of astrophysical signals. This surge in information density,...

25. Sustainable Simulation: Reducing FullSim Dependence with Cross-Experiment Pretrained Neural Networks

Matthias Schott (CERN / University of Mainz)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

Training state-of-the-art neural networks for high-energy physics (HEP) tasks typically requires massive, fully simulated datasets—whose generation is both computationally expensive and experiment-specific. In this work, we demonstrate that this dependence on large-scale full simulations can be drastically reduced by leveraging pretrained models trained on fast-simulation data. These...

332. SysVar: Consistent and scalable treatment of systematic uncertainties in template-based analyses

Ilias Tsaklidis (University of Bonn)

Track 9 - Analysis software and workflows

Poster Presentation

SysVar is a Python package that provides an end-to-end solution for the treatment and propagation of systematic uncertainties in analyses relying on templates generated from simulated data.

Propagating systematic uncertainties from correction weights into templates while preserving correlations in the signal extraction variables becomes increasingly challenging as analyses scale in size....

572. The INFN Open Access Repository: a FAIR challenge

Stefano Dal Pra (INFN)

Track 8 - Analysis infrastructure, outreach and education

Poster Presentation

The Open Access Repository, active since 2020, is the official INFN archive to host its research outputs according to FAIR principles. We describe its architectural and functional evolution, marked by the migration from Invenio v3 to the high-availability deployment based on Invenio RDM. Several technical issues due to the large "version jump" between the source and target platforms have been...

243. The SPD Software&Computing project.

Dr Danila Oleynik (Joint Institute for Nuclear Research (RU))

Track 4 - Distributed computing

Poster Presentation

The Spin Physics Detector (SPD) collaboration is building a versatile detector at the second interaction point of the NICA (Nuclotron-based Ion Collider fAcility) complex. As the detector's development progresses and the physics research program evolves, the demands for advanced data processing capabilities increase.
A defining feature of the facility is its triggerless (free-run) Data...

421. The tracking performance on LHCb UP detector in Upgrade II

Yisheng Fu (Chinese Academy of Sciences (CN))

Track 5 - Event generation and simulation

Poster Presentation

The LHCb experiment is planning a second major upgrade (Upgrade II) in the 2030s, with the goal of increasing the instantaneous luminosity to 1.0x1034 cm−2s−1. This upgrade aims to enhance the study of heavy flavor physics and to search for potential signals of new physics in the beauty and charm quark sectors. To operate under the demanding conditions of Upgrade II—characterized by higher...

260. TileCalibWeb Robot evolution towards HL-LHC: from COOL to CREST

Yuri Smirnov (Northern Illinois University (US))

Track 1 - Data and metadata organization, management and access

Poster Presentation

The ATLAS TileCalibWeb Robot application is the core tool in the Tile Calorimeter and the main interface for preparing and recording conditions and calibration data into the Online and Offline ORACLE Databases used daily by on-duty data quality control specialists and experts.
During LHC Run 3 TileCalibWeb Robot was significantly improved with numerous changes. These enhancements...

536. Topology-Informed Jet Tagging using Persistent Homology

saurav mittal

Track 9 - Analysis software and workflows

Poster Presentation

We present a topology-informed approach for classifying particle jets using persistent homology, a framework that captures the structural properties of point clouds. Particle jets produced in proton-proton collisions consist of cascades of particles originating from a common hard interaction. Each jet constituent is represented as a point in a three-dimensional feature space defined by the...

549. Toward Standardized Benchmarking Practices for Distributed HEP Computing: The ICSC Project

Gianluca Sabella (University Federico II and INFN, Naples (IT))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

The ICSC initiative (Italian National Centre for High-Performance Computing, Big Data, and Quantum Computing) is creating a flexible cloud platform to manage the escalating computational requirements of the High-Luminosity Large Hadron Collider (HL-LHC) and future collider projects. This approach leverages Kubernetes for orchestration and containerized deployments to streamline access to...

454. Towards the Italian national scientific datalake: case studies and insights

Diego Ciangottini (INFN, Perugia (IT))

Track 1 - Data and metadata organization, management and access

Poster Presentation

The Italian National Institute for Nuclear Physics (INFN) has been operating for more than 20 years the largest scientific distributed computing infrastructure: the Tier-1 at Bologna-CNAF and the 9 Tier-2 centres provide computing and storage resources to support more than 100 scientific collaborations.
In the last years this computer infrastructure has been expanded and modernized, also...

514. TPC MPD GAN Fast Simulation

Dr Alexey Boldyrev

Track 5 - Event generation and simulation

Poster Presentation

Detector response simulation is a computationally expensive step in the Monte Carlo production chain for High Energy Physics experiments. For the MPD experiment at NICA (JINR), we developed a method to accelerate the simulation of the Time Projection Chamber (TPC) response using a Generative Adversarial Network (GAN). Trained on data from standard GEANT4-based simulations, the GAN replaces...

537. Track-based Alignment of the BESIII CGEM Inner Tracker Using Cosmic-ray Data

Mr Jiajv Wang, Prof. Linghui Wu

Track 3 - Offline data processing

Poster Presentation

An upgrade of the inner tracker for the BESIII experiment has been completed in 2024. A three-layer Cylindrical GEM (CGEM) detector was installed in the BESIII detector, replacing the original inner drift chamber. For detector commissioning and alignment, cosmic-ray data were taken both with and without a magnetic field. A track reconstruction algorithm combing CGEM inner tracker (CGEM-IT)...

256. Training and Deployment of Transformers in ATLAS

Alessandro Zaio, Alessandro Zaio (INFN e Universita Genova (IT))

Track 3 - Offline data processing

Poster Presentation

Transformer architectures have rapidly become the state-of-the-art approach for machine-learning models across many domains in science, offering unprecedented performance on complex, high-dimensional tasks. Their adoption within the ATLAS experiment, starting with their usage for flavour tagging, has opened new opportunities, but also introduced substantial challenges regarding large-scale...

476. Transformer-based deep learning approach to measurement of the rare top process $t \to sW$

Woojin Jang (University of Seoul, Department of Physics (KR))

Track 9 - Analysis software and workflows

Poster Presentation

This study explores the feasibility of directly determining the CKM matrix element $|V_{ts}|$ through the rare top quark decay $t \to sW$ in the semileptonic final state of $t\bar{t}$ production. To overcome the significant background challenges inherent in this channel, we introduce a Transformer-based multi-domain $t\bar{t} \to sWbW$ signal event classifier that integrates both jet...

578. Transient stagnation in imaginary time evolution: origins and convergence analysis

Mr Suwannachad Suwannajitt (Chulalongkorn University)

Track 3 - Offline data processing

Poster Presentation

Quantum Imaginary Time Evolution (QITE) has recently received increasing attention as a pathway for ground state preparation on quantum hardware. However, the efficiency of this approach is frequently compromised by energy plateau, dynamical regimes characterized by vanishing energy reduction where the system stagnates near some metastable states. In this work, we dissect the anatomy of these...

215. Translating LHCb’s documentation – first experiences and ensuring maintenance

Andy MORRIS, Zhijie Wang (Lanzhou University (CN))

Track 6 - Software environment and maintainability

Poster Presentation

Since 2015, LHCb’s central onboarding resource for new collaborators has been the Starterkit, a set of self-study lessons that also form the basis of an annual in-person workshop in Geneva. Ahead of Run 3 (2022–2026), a new version of the Starterkit was
developed to accompany the Upgrade I software stack, with improved testing and updated exercises now used in the workshop.
However,...

109. Trigger-Level Analysis for Multi-Billion-Event Datasets

Tobias Fitschen (The University of Manchester (GB))

Track 2 - Online and real-time computing

Poster Presentation

Trigger bandwidth limitations constrain physics analyses that target low-mass resonances, where high-rate data collection is essential. To circumvent this limitation Trigger-Level Analysis (TLA) can be applied. A recent publication by the ATLAS experiment demonstrated this approach during LHC Run 2 by processing a massive dataset of over 60 billion events, more than twice the number of fully...

26. Uncovering Hidden Systematics in Neural Network Models for HEP

Matthias Schott (CERN / University of Mainz)

Track 9 - Analysis software and workflows

Poster Presentation

Neural networks (NNs) are inherently multidimensional classifiers that learn complex, non-linear relationships among input observables. While their flexibility enables unprecedented performance in high-energy physics (HEP) analyses, it also makes them sensitive to small variations in their inputs. Consequently, the propagation and estimation of systematic uncertainties in NN-based models...

538. Unifying Data and Metadata Access with PostgreSQL Foreign Data Wrappers

Mwai Karimi

Track 1 - Data and metadata organization, management and access

Poster Presentation

Modern data ecosystems are increasingly heterogeneous, with data and metadata distributed across multiple databases, file systems, and external services. This fragmentation creates challenges for organising data, managing systems, and enabling efficient access. This poster presents an approach for unifying access to distributed data sources using PostgreSQL Foreign Data Wrappers (FDWs)....

197. Update on the Canadian Belle II Tier1 Raw Data Centre

Dr Marcus Ebert (University of Victoria)

Track 1 - Data and metadata organization, management and access

Poster Presentation

We present an update on the usage of the Canadian Belle II raw data storage and computing infrastructure. The raw data storage system is a ZFS based system and the data access is managed by XRootD; without a WLCG accessible tape system. The system is in production for two years now and we will present our experience with such a system and how it was extended beyond the use as a raw data...

510. Upgrade of the KEDR detector DAQ system and beyond

Dmitriy Maximov

Track 2 - Online and real-time computing

Poster Presentation

The KEDR experiment is ongoing at the VEPP-4M $e^{+}e^{-}$ collider at Budker INP in Novosibirsk. The collider center of mass energy range covers a wide spectrum from 2 to 11 GeV. Most of the up-to-date statistics were taken at the lower end of the energy range around the charmonia region. Activities at greater energies up to the bottomonia lead to a significant increase of event recording...

100. Uproot-custom: Extend and customize TTree reading in Uproot

Mingrun Li

Track 9 - Analysis software and workflows

Poster Presentation

Uproot-custom is an extension of the popular Python ROOT-IO library Uproot that offers a mechanism to enhance TTree data reading capabilities without relying on ROOT. It provides native support for reading more complex TTree data formats (such as deeply nested containers and memberwise-stored data), and a registration mechanism that allows users to customize reading logic to meet their...

223. Using HyperQueue to increase utilization of EuroHPC resources by ATLAS Experiment

Michal Svatos (Czech Academy of Sciences (CZ))

Track 7 - Computing infrastructure and sustainability

Poster Presentation

The distributed computing system of the ATLAS experiment at the Large Hadron Collider (LHC) uses resources from several EuroHPC facilities through both allocated and opportunistic access. HyperQueue, a meta-scheduler developed at IT4Innovations, the Czech National Supercomputing Center, enables the experiment's workload to be adapted to the many-core architecture typical of modern HPC systems....

373. WordPress as a service for a diverse CERN community

Francisco Borges Aurindo Barros (CERN)

Track 7 - Computing infrastructure and sustainability

Poster Presentation

For over a decade, content management systems at CERN have been served by the on-premise Drupal service. In response to the high maintenance requirements of Drupal, the growing adoption of WordPress and the need to improve user experience, site management and governance, the WordPress service was established. The WordPress service provides a managed platform designed to empower and support the...