A5086190596- Particle physics theoretical and experimental studies
- Quantum Chromodynamics and Particle Interactions
- High-Energy Particle Collisions Research
- Particle Detector Development and Performance
- Neutrino Physics Research
- Computational Physics and Python Applications
- Dark Matter and Cosmic Phenomena
- Radiation Detection and Scintillator Technologies
- Atomic and Subatomic Physics Research
- Handwritten Text Recognition Techniques
- Superconducting Materials and Applications
- Nuclear physics research studies
- Black Holes and Theoretical Physics
- Medical Imaging Techniques and Applications
- Text and Document Classification Technologies
- Image Retrieval and Classification Techniques
- Stochastic processes and statistical mechanics
- Topic Modeling
- Image and Object Detection Techniques
- Graph Theory and Algorithms
- Scientific Computing and Data Management
- Computability, Logic, AI Algorithms
- Natural Language Processing Techniques
- Algorithms and Data Compression
- Radioactivity and Radon Measurements
University of Zurich
2024-2025
University of Cambridge
2025
University of Bristol
2025
Imperial College London
2025
Manchester Metropolitan University
2021-2023
European Organization for Nuclear Research
2019-2023
National University of Sciences and Technology
2017-2019
University of the Sciences
2017
Table detection is a crucial step in many document analysis applications as tables are used for presenting essential information to the reader structured manner. It hard problem due varying layouts and encodings of tables. Researchers have proposed numerous techniques table based on layout documents. Most these fail generalize because they rely hand engineered features which not robust variations. In this paper, we presented deep learning method detection. method, images first pre-processed....
Document structure analysis, such as zone segmentation and table recognition, is a complex problem in document processing an active area of research. The recent success deep learning solving various computer vision machine problems has not been reflected analysis since conventional neural networks are well suited to the input problem. In this paper, we propose architecture based on graph better alternative standard for recognition. We argue that more natural choice these problems, explore...
We explore the use of graph networks to deal with irregular-geometry detectors in context particle reconstruction. Thanks their representation-learning capabilities, can exploit full detector granularity, while natively managing event sparsity and arbitrarily complex geometries. introduce two distance-weighted network architectures, dubbed GarNet GravNet layers, apply them a typical reconstruction task. The performance new architectures is evaluated on data set simulated interactions toy...
Graph neural networks have been shown to achieve excellent performance for several crucial tasks in particle physics, such as charged tracking, jet tagging, and clustering. An important domain the application of these is FGPA-based first layer real-time data filtering at CERN Large Hadron Collider, which has strict latency resource constraints. We discuss how design distance-weighted graph that can be executed with a less than 1$\mu\mathrm{s}$ on an FPGA. To do so, we consider representative...
Experimental measurements of 𝑏-hadron decays encounter a broad spectrum backgrounds due to the numerous possible decay channels with similar final states. Additionally, computational constraints limit number simulations most significant backgrounds. Identifying leading requires careful analysis state particles, potential misidentifications and kinematic overlaps. This talk introduces an innovative approach utilising machine learning determine critical impacting decays.
Abstract We present the current stage of research progress towards a one-pass, completely Machine Learning (ML) based imaging calorimeter reconstruction. The model used is on Graph Neural Networks (GNNs) and directly analyzes hits in each HGCAL endcap. ML algorithm trained to predict clusters originating from same incident particle by labeling with cluster index. impose simple criteria assess whether associated as prediction are matched those resulting any particular individual particles....
Abstract We present an end-to-end reconstruction algorithm to build particle candidates from detector hits in next-generation granular calorimeters similar that foreseen for the high-luminosity upgrade of CMS detector. The exploits a distance-weighted graph neural network, trained with object condensation, segmentation technique. Through single-shot approach, task is paired energy regression. describe performance terms efficiency as well resolution. In addition, we show jet our method and...
The high-luminosity upgrade of the LHC will come with unprecedented physics and computing challenges. One these challenges is accurate reconstruction particles in events up to 200 simultaneous protonproton interactions. planned CMS High Granularity Calorimeter offers fine spatial resolution for this purpose, more than 6 million channels, but also poses unique algorithms aiming reconstruct individual particle showers. In contribution, we propose an end-to-end machine-learning method that...
We present a case for the use of Reinforcement Learning (RL) design physics instrument as an alternative to gradient-based instrument-optimization methods. It's applicability is demonstrated using two empirical studies. One longitudinal segmentation calorimeters and second both transverse well placement trackers in spectrometer. Based on these experiments, we propose approach that offers unique advantages over differentiable programming surrogate-based optimization First, algorithms possess...
Document structure analysis, such as zone segmentation and table recognition, is a complex problem in document processing an active area of research. The recent success deep learning solving various computer vision machine problems has not been reflected analysis since conventional neural networks are well suited to the input problem. In this paper, we propose architecture based on graph better alternative standard for recognition. We argue that more natural choice these problems, explore...
The high-luminosity upgrade of the LHC will come with unprecedented physics and computing challenges. One these challenges is accurate reconstruction particles in events up to 200 simultaneous proton-proton interactions. planned CMS High Granularity Calorimeter offers fine spatial resolution for this purpose, more than 6 million channels, but also poses unique algorithms aiming reconstruct individual particle showers. In contribution, we propose an end-to-end machine-learning method that...