IN2P3 Machine Learning workshop
jeudi 29 mars 2018 -
10:00
lundi 26 mars 2018
mardi 27 mars 2018
mercredi 28 mars 2018
jeudi 29 mars 2018
10:30
Introduction
-
David Rousseau
(
LAL-Orsay
)
Balázs Kégl
(
LAL/CNRS
)
Introduction
David Rousseau
(
LAL-Orsay
)
Balázs Kégl
(
LAL/CNRS
)
10:30 - 10:45
Room: Amphi
10:45
Neural Network Tracking for LHCb Vertex Detector
-
Da Yu Tou
(
Centre National de la Recherche Scientifique (FR)
)
Neural Network Tracking for LHCb Vertex Detector
Da Yu Tou
(
Centre National de la Recherche Scientifique (FR)
)
10:45 - 11:05
Room: Amphi
The LHCb experiment is scheduled for an upgrade at the end of 2018. In 2021, it has to collect data at collisions rate of 40MHz and an average of 5-6 PVs per event from 1MHz and 1-2 PVs per event today. We present an approach of using deep learning to reconstruct particle tracks in the vertex subdetector of LHCb, the Vertex Locator (VELO).
11:10
TrackML tracking challenge for LHC tracking
-
David Rousseau
(
LAL-Orsay
)
TrackML tracking challenge for LHC tracking
David Rousseau
(
LAL-Orsay
)
11:10 - 11:30
Room: Amphi
11:35
Active learning for "intelligent" simulation
-
Vladimir Gligorov
(
LPNHE
)
Active learning for "intelligent" simulation
Vladimir Gligorov
(
LPNHE
)
11:35 - 11:55
Room: Amphi
12:00
Generative models for ATLAS calorimetry
-
Aishik Ghosh
(
LAL
)
Generative models for ATLAS calorimetry
Aishik Ghosh
(
LAL
)
12:00 - 12:20
Room: Amphi
12:25
Documentation and tutorials discussion
Documentation and tutorials discussion
12:25 - 12:35
Room: Amphi
12:35
Déjeuner au Domus
Déjeuner au Domus
12:35 - 14:00
Room: Amphi
14:00
CTA Cerenkov telescope reconstruction
-
Mikaël Jacquemont
CTA Cerenkov telescope reconstruction
Mikaël Jacquemont
14:00 - 14:20
Room: Amphi
14:25
Transient photometric classification: an astronomical data challenge
-
Emille Ishida
(
LPC-UBP
)
Transient photometric classification: an astronomical data challenge
Emille Ishida
(
LPC-UBP
)
14:25 - 14:45
Room: Amphi
Among the many challenges imposed by the next generation of large scale astronomical surveys, the classification of transient sources is arguably one of the biggest obstacles to be overcomed before we can exploit the full potential of these new instruments. Although most of the standard astrophysical transient studies rely on high resolution spectroscopic observations, the new surveys will mostly deliver low resolution photometric measurements. Machine learning methods are then expected to overcome this sample selection bias providing reliable photometric classifications. In order to have an up to date picture of how different methods behave in this scenario, a new simulated data set is being developed - which will allow machine learning methods to be tested in a controlled environment. Moreover, PLAsTiCC (Photometric LSST Astronomical Time-series Classification Challenge) also aims to be a fertile ground for the development of new approaches based on LSST requirements. In this talk I will discuss the motivations and goals behind this data challenge and give details on how the broader community can engage in the challenge.
14:50
NN for image reconstruction for medical application
-
Francoise Bouvet-Lefebvre
(
IMNC
)
NN for image reconstruction for medical application
Francoise Bouvet-Lefebvre
(
IMNC
)
14:50 - 15:10
Room: Amphi
15:15
Gestion des logs au CCIN2P3 et passage à l'échelle : le ML à la rescousse ?
-
Fabien Wernli
(
Sysadmin
)
Gestion des logs au CCIN2P3 et passage à l'échelle : le ML à la rescousse ?
Fabien Wernli
(
Sysadmin
)
15:15 - 15:40
Room: Amphi
Près de 100 millions de "logs" et 1 milliard de métriques sont collectées par jour dans les deux datacentres du CCIN2P3. Ils sont traités via une plateforme implémantant le modèle dit "lambda" : les événements traversent deux "pipeline" en parallèle. La première de faible latence permet un traitement synchrone, presque temps-réel. La deuxième permet un traitement "batch" asynchrone sur toute ou partie des événements passés. Les deux "pipeline" permettent de notifier les gestionnaires de service du bon ou mauvais fonctionnement de leurs services grâce à des algorithmes et des règles prédéfinis, at modifiables à souhait. Le problème de cette approche est qu'elle ne passe pas à l'échelle. En effet, la quantité de "logs" ne fait qu'augmenter, et le temps humain nécessaire à maintenir le jeu de règles et d'algorithmes qui permettent de détecter les problèmes également. Le besoin de trouver une autre stratégie se fait sentir, et les techniques de ML voire de DL semblent prometteuses pour assister par exemple en utilisant les techniques de "outlier detection".
15:45
Automated training of RAMP on CC GPU cluster
-
Alexandre Boucaud
(
Paris-Saclay Center for Data Science / LAL
)
Automated training of RAMP on CC GPU cluster
Alexandre Boucaud
(
Paris-Saclay Center for Data Science / LAL
)
15:45 - 16:05
Room: Amphi
I will present a work undergone in collaboration with Bertrand Rigaud (CCIN2P3) about the creation of a pipeline to allow ML submissions from RAMP challenges to be automatically trained on the CC infrastructure. This will be used in the future to organise IN2P3-backed challenges (LSST, Euclid) and face a short-term high demand in computing power.
16:05
Experience with GPU platform, CC and elsewhere
Experience with GPU platform, CC and elsewhere
16:05 - 16:15
Room: Amphi
16:15
Conclusion
-
Balázs Kégl
(
LAL/CNRS
)
David Rousseau
(
LAL-Orsay
)
Conclusion
Balázs Kégl
(
LAL/CNRS
)
David Rousseau
(
LAL-Orsay
)
16:15 - 16:30
Room: Amphi