A5036547514- Parallel Computing and Optimization Techniques
- Distributed and Parallel Computing Systems
- Data Mining Algorithms and Applications
- Particle physics theoretical and experimental studies
- Interconnection Networks and Systems
- Particle Detector Development and Performance
- Advanced Data Storage Technologies
- Anomaly Detection Techniques and Applications
- Semantic Web and Ontologies
- Computational Physics and Python Applications
- Color Science and Applications
- Scientific Computing and Data Management
- Data Quality and Management
- Distributed systems and fault tolerance
- Algorithms and Data Compression
- Embedded Systems Design Techniques
- Data Stream Mining Techniques
- Data Management and Algorithms
- COVID-19 diagnosis using AI
- Wikis in Education and Collaboration
- Medical Imaging Techniques and Applications
- Big Data Technologies and Applications
- Biomedical Text Mining and Ontologies
- Software System Performance and Reliability
- Superconducting Materials and Applications
Laboratoire de Recherche en Informatique
2009-2023
Université Paris-Sud
2009-2023
Centre National de la Recherche Scientifique
2004-2023
Institut national de recherche en informatique et en automatique
2007-2023
Université Paris-Saclay
1994-2023
Laboratoire Interdisciplinaire de Physique
2022
Laboratoire Interdisciplinaire des Sciences du Numérique
2022
Ministère de la Culture
2009
Université Paris Cité
1991-2002
Global computing achieves high throughput by harvesting a very large number of unused resources connected to the Internet. This parallel model targets architecture defined nodes, poor communication performance and continuously varying resources. The unprecedented scale global paradigm requires us revisit many basic issues related programming models, class applications or algorithms suitable for this architecture. XtremWeb is an experimental platform dedicated provide tool such studies. paper...
Global Computing platforms, large scale clusters and future TeraGRID systems gather thousands of nodes for computing parallel scientific applications. At this scale, node failures or disconnections are frequent events. This Volatility reduces the MTBF whole system in range hours minutes. We present MPICH-V, an automatic tolerant MPI environment based on uncoordinated checkpoint/roll-back distributed message logging. MPICH-V architecture relies Channel Memories, Checkpoint servers...
Exploiting the rapid advances in probabilistic inference, particular variational Bayes and autoencoders (VAEs), for anomaly detection (AD) tasks remains an open research question. Previous works argued that training VAE models only with inliers is insufficient framework should be significantly modified order to discriminate anomalous instances. In this work, we exploit deep conditional autoencoder (CVAE) define original loss function together a metric targets hierarchically structured data...
Abstract This paper reports on the second “Throughput” phase of Tracking Machine Learning (TrackML) challenge Codalab platform. As in first “Accuracy” phase, participants had to solve a difficult experimental problem linked tracking accurately trajectory particles as e.g. created at Large Hadron Collider (LHC): given $$O(10^5)$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>O</mml:mi> <mml:mo>(</mml:mo> <mml:msup> <mml:mn>10</mml:mn> <mml:mn>5</mml:mn>...
Global Computing platforms, large scale clusters and future TeraGRID systems gather thousands of nodes for computing parallel scientific applications. At this scale, node failures or disconnections are frequent events. This Volatility reduces the MTBF whole system in range hours minutes. We present MPICH-V, an automatic tolerant MPI environment based on uncoordinated checkpoint/roll-back distributed message logging. MPICH-V architecture relies Channel Memories, Checkpoint servers...
Wiki4R will create an innovative virtual research environment (VRE) for Open Science at scale, engaging both professional researchers and citizen data scientists in new potentially transformative forms of collaboration.It is based on the realizations that (1) structured parts Web itself can be regarded as a VRE, (2) such environments depend communities, (3) closed are limited their capacity to nurture thriving communities.Wiki4R therefore integrate Wikidata, multilingual semantic backbone...
This paper discusses a new principle of interconnection network for massively parallel architectures in the field numerical computation. The is motivated by an analysis application features and need to design kind communication networks combining very high bandwidth, low latency, performance independence pattern or load improvement proportional hardware improvement. Our approach associate compiled communications circuit switched network. presents motivations this principle, software issues...
Grid systems are complex heterogeneous systems, and their modeling constitutes a highly challenging goal. This paper is interested in the jobs handled by EGEE grid, mining Logging Bookkeeping files. The goal to discover meaningful job clusters, going beyond coarse categories of "successfully terminated jobs" "other jobs". presented approach three- step process: i) Data slicing used alleviate heterogeneity afford discriminant learning; ii) Constructive induction proceeds learning hypotheses...
Charged particle track reconstruction is a major component of data-processing in high-energy physics experiments such as those at the Large Hadron Collider (LHC), and foreseen to become more challenging with higher collision rates. A simplified two-dimensional version problem set up on collaborative platform, RAMP, order for developers prototype test new ideas. small-scale competition was held during Connecting The Dots / Intelligent Trackers 2017 (CTDWIT 2017) workshop. Despite short time...
The High-Luminosity LHC (HL-LHC) is expected to reach unprecedented collision intensities, which in turn will greatly increase the complexity of tracking within event reconstruction. To out computer science specialists, a machine learning challenge (TrackML) was set up on Kaggle by team ATLAS, CMS, and LHCb physicists experts scientists building experience successful Higgs Machine Learning 2014. A training dataset based simulation generic HL-LHC experiment tracker has been created, listing...
Detecting the changes is common issue in many application fields due to non-stationary distribution of applicative data, e.g., sensor network signals, web logs and grid-running logs. Toward Autonomic Grid Computing, adaptively detecting a grid system can help alarm anomalies, clean noises, report new patterns. In this paper, we proposed an approach self-adaptive change detection based on Page-Hinkley statistic test. It handles without assumption data empirical setting parameters. We validate...
Experimental science often has to cope with systematic errors that coherently bias data. We analyze this issue on the analysis of data produced by experiments Large Hadron Collider at CERN as a case supervised domain adaptation. Systematics-aware learning should create an efficient representation is insensitive perturbations induced effects. present experimental comparison adversarial knowledge-free approach and less data-intensive alternative.