By continuing to browse the site, you are agreeing to our use of cookies. (More details)
Portal > Offres > Offre UMR8630-AURHEE-001 - post-doctorat: Apprentissage par renforcement pour optimiser le "Time Delay Interferometry" H/F

post-doctoral position: Reinforcement Learning for Time Delay Interferometry Ranging H/F

This offer is available in the following languages:
Français - Anglais

Ensure that your candidate profile is correct before applying. Your profile information will be added to the details for each application. In order to increase your visibility on our Careers Portal and allow employers to see your candidate profile, you can upload your CV to our CV library in one click!

Faites connaître cette offre !

General information

Reference : UMR8630-AURHEE-001
Workplace : PARIS 14
Date of publication : Monday, July 27, 2020
Type of Contract : FTC Scientist
Contract Period : 12 months
Expected date of employment : 1 January 2021
Proportion of work : Full time
Remuneration : Between 2728.25€ et 4307.86€ bruts per month depending on past experience
Desired level of education : 5-year university degree
Experience required : Indifferent

Missions

LISA is a space mission that will detect low frequency gravitational waves. The LISA mission consists in performing interferometric measurements between laser links exchanged between three spacecraft in heliocentric orbits. The signatures from gravitational waves will be imprinted in those interferometric measurements and will have to be extracted from various instrumental noises. In particular, it will be crucial to reduce the frequency laser noise by ~ 8 orders of magnitude in order to reach the sensitivity required to detect gravitational waves. This will be realized by a technique names “Time Delay Interferometry” (TDI) which combines the various interferometric measurements by applying some temporal delays. Two methods are currently known to perform TDI: 1) to estimate these TDI delays by using additional pseudo-range measurements and 2) to estimate these delays directly using some optimization methods (this is known as “TDI-ranging”). In this project, we will explore the second option.

Activities

In this project, we will develop an innovative method based on reinforcement learning to estimate the TDI delays with no knowledge a priori about the spacecraft trajectory. In this project, we will therefore simulate LISA data and explore various machine learning algorithms that would allow to estimate the TDI delays that minimize the residual laser noise. In a second step, we will explore algorithms that would allow to estimate these delays simultaneously with the parameters characterizing the gravitational wave sources. This will allow to study the correlations between the TDI delays and these scientific parameters and to explore the impact of a misestimation of these delays on LISA scientific results.

Skills

Strong skills in software development, data analysis, numerical calculations are required. In addition, a good knowledge of TDI algorithm will be necessary. Notions of machine learning and of reinforcement learning will be favored.

Work Context

This project will be conducted within the « Theory and metrology” group from SYRTE. This group is largely involved in the development of the LISA preprocessing pipeline. SYRTE is therefore a full member of the LISA consortium and is part of the simulation working group and of the “initial noise reduction pipeline” working group. Currently, 5 researchers are working actively on these topics and the group has developed various collaborations around these topics.

We talk about it on Twitter!