En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)

PhD Thesis - Interactive quality assessment model for uncertain data about maritime trade during XVI

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

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

Faites connaître cette offre !

General information

Reference : UMR7266-CHRPLU-002
Workplace : LA ROCHELLE
Date of publication : Tuesday, January 08, 2019
Scientific Responsible name : Christine PLUMEJEAUD-PERREAU
Type of Contract : PhD Student contract / Thesis offer
Contract Period : 36 months
Start date of the thesis : 1 October 2019
Proportion of work : Full time
Remuneration : 1 768,55 € gross monthly

Description of the thesis topic

This subject takes place inside a program named PORTIC financed by French National Agency for Research (ANR). PORTIC intends to study the spatial and economic dynamics at work which resulted in increasingly integrated markets, a process which paved the way to, and sustained the Industrial Revolution. By crossing data on the shipping activities of French ports and those of the 18th-century French balance of trade, PORTIC aims at understanding the articulation between regional, national and international trade. PORTIC builds on two massive corpora produced by two previously achieved ANR programs, Navigocorpus and Toflit18. The combination of information contained in the two corpora will make it possible to estimate the respective part of national and international trade, to better perceive the ports which articulated markets and their interrelations, to analyze regional specialization mechanisms at work among different ports, to assess the importance of warfare on a port economy, to gauge the relevance of smuggling across the Channel, to quantify the part of French international transport services which is absent from contemporary statistics, to compute the ratio between trade value and tonnages or crew size in maritime transport flows.
PORTIC, a project jointly conceived by historians, economists, GIS specialists, computer scientists, and specialists in information and interaction design on the Web, will develop tools for a clear and scientifically sound visualization of these historical data, and their diverse reliability degree. Data's imperfection derives from missing information, contradictory data provided by different historical sources, and their imprecise content. The imperfect character of part of the information constitutes an essential element for a correct historical understanding, but it is presently insufficiently taken into account. Digital Humanities are fundamental to all the phases of the project. Data mining and semi-automatized procedures will make it possible to detect outliers and contradictory information and to let historians qualify its reliability.

This PhD project aims to tackle the qualification issues of those data with an approach combining both symbolic and numeric methods through an iterative process integrating experts' feedbacks for the corpus' data curation.
Different aspects will be dealt with during this thesis project:
- A semantic model for trajectories derived from a generic spatiotemporal model (Tran et al. 2016) will be used to deduce inconsistencies in the database (contradictory information, incoherent itineraries).
- This model will be connected with an engine executing some non-parametric and non-supervised statistical methods for patterns and outliers detection (data-mining field).
- A semantic quality model will extend the current existing semantic model for trajectories in order to handle the qualitative annotations.
- The findings will be displayed in the geo- and data-visualization interfaces (developed elsewhere inside the project), thus allowing for the expert's feedbacks to be integrated into the semantic model for a further iterative exploration of various hypotheses. This will induce a support for non-monotonous reasoning.
The approach will be assessed firstly by comparing ancient raw datasets with the same already manually corrected, and then on newly collected data in the project by making the software interact with historians having the human expertise role.

References:
Moreau, C., Devogele T. and Etienne L., 2018 Extraction de motifs de trajectoires sémantiques similaires. In : Proceedings of Spatial Analysis and GEOmatics (SAGEO'2018), Eds : Mathieu Roche, Maguelonne Teisseire, Montpellier, 6-9 nov. 2018, France
Etienne L., Devogele T., Bouju A., 2009. « Analyse de similarité de trajectoires d'objets mobiles suivant le même itinéraire : Application aux trajectoires de navires », Revue Ingénierie des Systèmes d'Information (ISI), Hermès, vol. 14:5, p. 85-106.
Grosso E., Bouju A., Mustière S., 2009. "Data Integration GeoService: A First Proposed Approach Using Historical Geographic Data" In: Proceedings of 9th International Workshop on Web and Wireless Geographical Information Systems, Eds: J.D. Carswell, A. S. Fotheringham & G.McArdle, W2GIS 2009 7-8 December, Maynooth, Ireland Lecture Notes in Computer Science, pp 103-119.
Plumejeaud C., Villanova-Oliver M., 2012. “QualESTIM: Interactive Quality Assessment of Socioeconomic Data Using Outlier Detection”, in: Gensel J., Josselin D., Vandenbroucke D. (Eds.), Bridging the Geographic Information Sciences: International AGILE'2012 Conference, Avignon (France), April, 24-27, 2012., Heidelberg, pp. 143–160. https://doi.org/10.1007/978-3-642-29063-3_8
Plumejeaud C., Grosso E., Parent B., 2014. 'Dissemination and geovisualisation of territorial entities' history', Journal of Spatial Information Science. doi:10.5311/JOSIS.2014.8.119 ; http://josis.org/index.php/josis/article/view/119
Plumejeaud C., Cristofoli P., Motte C., 2015. “De l'étude des nomenclatures territoriales à la modélisation des dynamiques des territoires administratifs en France”, Revue internationale de géomatique 25, p. 355–392. https://doi.org/10.3166/rig.25.355-392
Plumejeaud-Perreau, C., 2018. La qualité des donnéées. In : Action Nationale de Formation 'Sciences des données' [en ligne]. Sète. 6 novembre 2018. [Consulté le 29 novembre 2018]. Disponible à l'adresse : http://rbdd.cnrs.fr/IMG/pdf/qualite_des_donnees_plumejeaud_2018_04112018.pdf?517/365a13edab604bd0700b045bfac29a3607acb649
Ray C., Napoli A., Bouju A., Martin P.-Y., 2015. Detection of faked AIS messages and Resulting Risks, in: IF&GIS 2015 - 7th International Workshop on Information Fusion and Geographic Information Systems. Grenoble, France.
Tran B.-H., Bouju A., Plumejeaud-Perreau C., Bretagnolle V., 2016. “Towards a semantic framework for exploiting heterogeneous environmental data”. International Journal of Metadata, Semantics and Ontologies, vol. 11, no 3, p. 191-205, https://doi.org/10.1504/IJMSO.2016.081586
Tran B.-H., 2017. Une approche sémantique pour l'exploitation de données environnementales - Application aux données d'un observatoire. Thèse de doctorat, Université de la Rochelle.
Wannous R., Malki J., Bouju A., Vincent C., 2017. “Trajectory ontology inference considering domain and temporal dimensions—Application to marine mammals”. Future Generation Computer Systems 68, 491–499. https://doi.org/10.1016/j.future.2016.01.012

Work Context

The student will be recruted by the CNRS, at the Rochelle, at the Center for the Littoral, Environment, and Societies (LIENSs, Lab UMR 7266). CNRS whose motto is « pushing back the frontier of science », and covers all scientific fields from the study of matter and the living world to that of human societies. The EUCLIDE doctoral school will follow the management of the thesis.

The LIENSs laboratory gathers expertise in different scientific fields (environmental, engineering, social sciences -geography and history- and humanities) to address issues related to the sustainable management of the coastal zone (https://lienss.univ-larochelle.fr/).
PORTIC's interactive data curation model coupled with the geo-visualization framework for imperfect data is valuable for many of the fields and projects in this research center. For instance, LIENSs is part of the regional project Atlas historique de la Nouvelle-Aquitaine, which features a case study focusing on the Marennes district: a small administrative capital (admiralty, election, bailiwick and then arrondissement) located on the estuary of the Seudre, which was linked to La Rochelle's maritime economy, in an environment that underwent major changes since the 15th-16th centuries. LIENSs is also a key partner of the DYPOMAR project, as the multi-scalar study of port dynamics is essential to understand present-day coastal transformations.
The PORTIC LIENSs team is coordinated by Christine Plumejeaud-Perreau, who has been working for the DYPOMAR project with Alain Bouju of the Laboratory for Computing, Images, and Interaction (L3i) at University of La Rochelle since 2014. This team will be in charge of the implementation of a progressive, user-friendly interface for the geo-visualization of maritime flows, and of the integration of a semantic data model dedicated to trajectories for detecting logical inconsistencies. The close collaboration that already exists in La Rochelle between LIENSs and L3I (200 meters away from) is a guarantee for a good management of the thesis.
Supervisors :
• A. Bouju (HDR) is an expert of semantic works on trajectories, processing for instance marine mammals' telemetric data (Wannous et al., 2017), ships' mobility (Etienne et al., 2009; Ray et al., 2015). His research centre, L3i, has an internationally recognized competence for handling cultural heritage in digital systems, and A. Bouju has worked on Alpage project for handling metadata and services for integration of old imperfect maps (Grosso et al., 2009).
• C. Plumejeaud has worked on interactive outliers' detection and long-term database building during her thesis (Plumejeaud et al., 2012), and she has often been the interface between historians and computer scientists, such for instance for the ANR GeoPeuple program, that produced geo-visualisation interfaces for understanding municipalities' history (Plumejeaud et al., 2014, Plumejeaud et al., 2015).
PORTIC's interactive data curation model coupled with the geo-visualization framework for imperfect data is valuable for many of the fields and projects in this research center. For instance, LIENSs is part of the regional project Atlas historique de la Nouvelle-Aquitaine, which features a case study focusing on the Marennes district: a small administrative capital (admiralty, election, bailiwick and then arrondissement) located on the estuary of the Seudre, which was linked to La Rochelle's maritime economy, in an environment that underwent major changes since the 15th-16th centuries. LIENSs is also a key partner of the DYPOMAR project, as the multi-scalar study of port dynamics is essential to understand present-day coastal transformations.

Constraints and risks

The corpus of data to be analyzed is already present and structured and this is a guarantee for the doctoral student.
The close collaboration between the L3i and the LIENSs (located 200 m apart) is also a guarantee for a fertile dialogue and, ultimately, the good management of the thesis. Alain Bouju and Plumejaud-Perreau already co-supervised a PhD. thesis using semantic technologies (Tran, 2017), with a particular spatiotemporal ontology having effective capabilities for reasoning and logical inference, in order to offer a generic framework for a cross-analysis and visualization of long-term biodiversity and crop data.
Regarding the academic future of the doctoral student, although the subject is carried out in an interdisciplinary environment, the thesis will be clearly supported in computer science (section 6 of the CNRS, 27 of the CNU, qualification area of the two supervisors) and we aim publications in international journals. The doctoral student will also have funds to meet his research community, at both nationally and internationally levels.

Additional Information

Education: Master specialty Computer Sciences / Knowledge engineering
Desired Experience in Statistics and Data Mining (outliers detections, patterns recognition), Semantic Web (Protégé, ontologies) and Linked Data (RDF, LOD).
The application must include a CV, a letter of motivation, a transcript of Master 2 or engineering school last year, as well as the latest internship report.

We talk about it on Twitter!