Annonce en lien avec l’Action/le Réseau : BigData4Astro

Laboratoire/Entreprise : LESIA, Observatoire de Paris
Durée : 2 years
Contact : baptiste.cecconi@obspm.fr
Date limite de publication : 2020-06-30

Contexte :
The position is open within the MINERVA project (Machine Learning for Radioastronomy at Observatoire de Paris. This project federates astrophysicists interested in a variety of astrophysical phenomena. https://vm-wordpress-lerma01.obspm.fr/minerva/

Sujet :
Radioastronomy is experiencing an explosion of volumes of observational data with the development of giant interferometers (LOFAR, ALMA, NenuFAR, SKA). These instruments produce huge and numerous two and four-dimensional datasets (among the 2D-spatial, 1D temporal and 1D spectral coordinates, depending on observation mode). Faced to these daily TB-scale data (PB-scale with SKA), the traditional methods of source and event detection and classification reach their limits. In parallel, machine learning methods have undergone algorithmic developments that bring them to a high level of maturity.

The goal of this project is to perform pilot implementation of new methods for (i) transient radio sources classification based on their morphology in Time-Frequency domain (such as Solar bursts or Jovian emissions, Pulsars, Fast Radio Bursts (FRB), Earth and planetary lightnings, etc), (ii) real-time event detection and classification on radio-astronomical observed data streams and (iii) processing of multi-instrument and multi-wavelength aggregated data, including triggering from external event detections for low frequency follow ups (such as FRB, GRB (Gamma Ray Bursts), GW (Gravitational Waves), etc).

The successful candidate will carry out an inventory of existing methods and design new tools that shall be applied to spectro-temporal data streams and to large quantities of such data.

MINERVA will make use of datasets from NenuFAR, NDA and LOFAR. The new algorithms will also be tested against existing data collections and event lists.

Profil du candidat :
Applicants should have at least an engineer diploma in the field of Machine Learning or a PhD in physics, astronomy, or computer science by the time of the appointment. Experience in astronomy is not mandatory.

Formation et compétences requises :
We encourage applications from candidates with a strong expertise in either the manipulation or the development of state-of-the-art Machine Learning methods. Experience with manipulating images and data cubes will also be considered. Skills in one or several programing languages (e.g. Python, Fortran, C++) are necessary.

Adresse d’emploi :
Organization: LESIA, Observatoire de Paris
Street Address : 5, place Jules Janssen
City: Meudon
Zip/Postal Code: 92195
Country: France

Document attaché : Open-Position-2-MINERVA-–-Machine-Learning-for-Radioastronomy-at-Observatoire-de-Paris.pdf