Faites connaître cette offre !
Reference : UMR5060-PHILAN-001
Workplace : PESSAC
Date of publication : Wednesday, June 17, 2020
Scientific Responsible name : Philippe Lanos
Type of Contract : PhD Student contract / Thesis offer
Contract Period : 36 months
Start date of the thesis : 1 October 2020
Proportion of work : Full time
Remuneration : 2 135,00 € gross monthly
Description of the thesis topic
The ChronoCurve project follows on from the Chronomodel ANR-11-MONU-007 (November 2011 - January 2015), which produced time series modeling software based on Bayesian statistics. It is free, open-source and multi-platform (https://chronomodel.com/). This statistical processing and modelling software is the basis for the construction of chronological scenarios by archaeologists, historians, paleo-ecologists and geologists in possession of a large number of dating data linked according to more or less complex models. The stratigraphic constraints, phasing and correlations that link the data together, as well as the different sources of uncertainty involved at the different levels of dating methods are thus taken into account within a single tool.
The ChronoCurve thesis project aims to address the issue of the construction of reference frames (called "calibration curves"). These reference frames are essential to convert any laboratory (dating) measurement into a calendar date: this is for example the case for radiocarbon and archaeomagnetism.
The project is clearly interdisciplinary and exploratory. Within the framework of a Bayesian approach, it proposes to provide new tools for the construction of reference curves (geomagnetic secular variation curves, age-depth curves) and for the calculation of dates in archaeomagnetism (vectorial calculation) and dendrochronology (Bayesian synchronization of dendrological sequences).
We suggest a change of paradigm in the field of date computation: Bayesian statistical modeling allows to consider simultaneously the calculation of dates and the construction of reference curves. This is a novelty that requires statistical modeling work with simulations.
On the other hand, the development of Plugins for dendrochronology and for archaeomagnetism will require new numerical calculation modes associating MCMC (Markov Chain Monte Carlo) and particle filtering.
Methodologies to be set up
The mathematical tools implemented are based on matrix computation, reproducing kernels, spline functions, Bayesian statistics, MCMC computation and particle filtering. The various algorithmic developments will be done in C++ under Qt. The ChronoCurve software will be integrated in ChronoModel and will be cross-platform and open-source on GitHub.
The PhD student's role in the project will be to set up and test the algorithms for calculating reference curves. The aim will be to optimise the numerical calculation procedures (matrix calculation, calculation by reproducing kernels, MCMC calculation). The development of the "dendrochronology" and "archaeomagnetism" plugins to be implemented in ChronoModel will also require specific research on Bayesian modelling and numerical calculation.
This doctoral contract is part of the CNRS 80 Prime 2020 program.
The doctoral school of reference will be Montaigne-Humanité at Bordeaux-Montaigne University.
The PhD student will be assigned to the UMR 5060 IRAMAT (Institut de Recherche sur les Archéomatériaux), in the CRP2A laboratory (Centre de Recherche en Physique Appliquée à l'Archéologie) based at the Bordeaux-Montaigne University.
Location of the thesis: in Rennes, at Géosciences-Rennes, Beaulieu campus, Université Rennes 1, within the framework of a partnership between IRAMAT and Géosciences-Rennes.
The thesis will be carried out in interaction with the "Chronology and Reference Frameworks" team at CRP2A in Bordeaux, with the "Chronological Modelling and Dating by Archaeomagnetism" team at Géosciences-Rennes and with the "Landscapes, Climate Change, Biodiversity" team (in collaboration with D. Marguerie, co-director of the thesis) of UMR 6553 ECOBIO in Rennes.
Constraints and risks
Round trips between Rennes and Bordeaux are to be expected for team meetings and for the follow-up of the doctoral school courses during the thesis to validate the PhD.
The desired candidate will have a profile in applied mathematics specializing in statistics with skills in the field of stochastic numerical simulation. In addition, the candidate should have a strong interest in applications to archaeology and natural sciences involving chronological issues (Paleoenvironments).
Skills and Qualifications :
- Engineer / Master 2 in Applied Mathematics, Statistics and Probability
- Skills in computer programming and numerical computation
- Strong interest in the human sciences, particularly in archaeology and dating methods in archaeometry
- Autonomy, open-mindedness, capacity for teamworking
- Good oral and written communication skills (French and English)
We talk about it on Twitter!