Postdoctoral position: Banking data analysis for fraud detection in online payments, IRIMAS/UHA, Mulhouse France



  • Postdoctoral position: Banking data analysis for fraud detection in online payments


  • We offer a postdoc position on Deep Learning for bank frauds detection.


  • A 12 months contract, in collaboration with Oméga and MSD research groups from IRIMAS Institute, Mulhouse, France.


  • More information below or at https://maxime-devanne.com/jobs/Post-DocSATTEnygmaEng.pdf


  • Context: The volume of online sales in France is continuously increasing. Remote payments on the Internet concentrate most of the fraud on French credit cards (70% of the amount of fraud (255 M €) while it represents only 14% of the value of national transactions). Fraud is a complex phenomenon to detect. Indeed, fraudsters constantly adapt their techniques in order to outsmart the system, which poses a financial and reputational risk to e-commerce sites and banks.


  • Goal of the project: The project aims to develop machine learning methods for the detection of credit card frauds. This project is a collaboration between the company Enygma, specializing in the detection of bank frauds, and the IRIMAS research institute specializing in machine learning methods.


  • Prerequisites: Applicants must have a PhD in Computer Science and demonstrate experience in the field of machine learning. We are looking for a profile with good experience in deep learning methods as well as programming in Python. Experience in optimization as well as CUDA programming would be a plus.


  • Working conditions:


  • Duration: 12 months


  • Gross salary : 3250 €/month


  • Location: Université de Haute-Alsace, Mulhouse


  • Working environment: The recruited person will be integrated within the IRIMAS research institute and will work more specifically in collaboration with Prof. Lhassane Idoumghar and Dr. Julien Lepagnot from the OMéGA team and Dr. Jonathan Weber and Dr. Maxime Devanne from the MSD team. It will benefit from the laboratory's computing servers and a dynamic and stimulating scientific environment.


  • Contact: Please contact Prof. Lhassane Idoumghar ([email protected]), Dr. Jonathan Weber ([email protected]), Dr. Julien Lepagnot ([email protected]) or Dr. Maxime Devanne ([email protected]) for more information about the position.


  • Maxime Devanne Maitre de Conférences en Informatique ENSISA/IRIMAS


  • Université de Haute-Alsace https://maxime-devanne.com


Two PhD positions at Sorbonne University: Deep Learning for modeling complex dynamics in the physical world



  • Two PhD positions are available at Sorbonne University on the following topic: Deep Learning for modeling complex dynamics in the physical world.


  • Both positions focus on the challenge of combining prior physical knowledge together with machine learning for modeling complex spatio-temporal dynamics.


  • One position is available at Sorbonne-University, Abou Dhabi, on the Modeling of Dynamical Systems with Geoscience Application: https://mlia.lip6.fr/wp-content/uploads/2021/09/PhD-position_SUAD-Deep-Learning-for-Modeling-Physical-Dynamics.pdf


  • One position is available at Sorbonne University, Paris, on Deep Learning in Computational Fluid Dynamics : https://mlia.lip6.fr/wp-content/uploads/2021/10/Machine-Learning-Computational-Fluid-Dynamics-PhD-3-1.pdf


  • Prof. Patrick Gallinari Sorbonne Universite - LIP6 Boite 169


  • 4 place Jussieu, 75252 Paris Cedex 05, France


  • Tel: 33144277370, fax: 33144277000


  • http://www-connex.lip6.fr/~gallinar/


Offre de stage M2 - Clustering imperfect data for personalized medicine: multimodal (kinetic, clinical and biological) data in spondylarthritis



  • Hello everyone, Please find attached an M2 internship offer at EuroMov digital Health in Motion (IMT Mines Alès, University of Montpellier).


  • Regards, Abdelhak Imoussaten


  • Clustering imperfect data for personalized medicine: multimodal (kinetic, clinical and biological) data in spondylarthritis


  • Spondylarthritis affects 1/1000 individuals, mainly young men, with a delay in diagnosis resulting in high medical costs.


  • Spondylarthritis is characterized by disabling back pain and 2/3 of patients show biological signs of chronic inflammation.


  • In 90% of these people, there is a particular gene in the histocompatibility system (HLA B27), which makes it very likely that they have an inappropriate immune response that causes inflammation.


  • Logically, most patients respond well to biotherapy (anti-TNFα, anti-IL-17), but this is not the case for 30% of them.


  • Especially for these non-responders, we need novel discriminating features to tailor therapy.


  • To do so, we record clinical, biological and movement data over time and we analyze this data within an interdisciplinary team at the interface of Health, Movement and Data sciences.


  • We recruit 3 master students (medical sciences + movement sciences + data sciences) working together under the co-supervision of 3 senior scientists: C. Jorgensen (medicine), D. Mottet (movement) and A. Imoussaten (data sciences).


  • In the data science project, the goal is to build cautious algorithms to classify patients by analyzing the imperfect multimodal data (clinical, biological and movement).


  • We address the classification problem by focusing on a careful dealing with the uncertainties in the recorded data.


  • Uncertainty is related to the reliability of the sensors, but it is also related to the variability of human movement/physiology/biology and its evolution over time.


  • Here, it is essential to have the most faithful representation of the data, including its imperfection, to feed the medical strategy.


  • Candidate Profile and missions


  • Skills:


  • - Machine learning


  • - Decision theory


  • - Time Series


  • - R, Python


  • The missions of the data science intern are:


  • - To operationalize the access to medical records following ethics regulations.


  • In close collaboration with the medical intern, the data management plan that follows the ethical approval is implemented digitally.


  • - To organize data of various types/nature into a coherent database


  • Data to process include textual reports (clinical data), textual and numerical spreadsheet information (biological data) and long/short timeseries of motion capture (movement data).


  • - To design algorithms that cautiously process imperfect input data This is the core mission.


  • What is challenging is the management of imprecise data (necessary to cautious classification) and the tractability and interpretability of the algorithms (necessary to make sense for the medical people).


  • Mail (cv + letter of motivation) to [email protected] avant le 31 octobre 2021.


Thesis subject at the Grenoble Informatics Laboratory: Operational population of knowledge bases and Neural Networks



  • Thesis subject within the framework of the Popcorn project (collaborative project with two companies) supervised by Benjamin Lecouteux, Gilles Sérasset and Didier Schwab (Grenoble Computer Science Laboratory, Study Group in Machine Translation / Automated Language and Speech Processing)


  • Title: Operational population of knowledge bases and Neural Networks


  • The project addresses the problem of the semi-automated enrichment of a knowledge base through the automatic analysis of texts. In order to obtain a breakthrough innovation in the field of Automatic Natural Language Processing (NLP) for security and defense clients, the project focuses on the processing of French (even if the approaches adopted will subsequently be generalizable to other languages). The thesis work will address different aspects:


  • ● Automatic annotation of textual documents by the detection of mentions of entities present in the knowledge base and their semantic disambiguation (polysemy, disambiguation);


  • ● The discovery of new entities (people, organizations, equipment, events, places), their attributes (age of a person, reference number of an equipment, etc.), and relationships between entities (a person works for an organization, people involved in an event, ...). Particular attention will be given to being able to adapt flexibly to changes in ontology, taking into account the role of the user and the analyst for the validation / capitalization of the extractions carried out.


  • The project focuses on the following three areas of research:


  • ● Generation of synthetic textual data from reference texts;


  • ● Recognition of entities of interest, associated attributes and relationships between entities.


  • ● The semantic disambiguation of entities (in case of homonymy for example)


  • Required profile: - Solid experience in programming & machine learning for Automatic Language Processing (NLP), especially deep learning


  • - Master in Machine Learning or computer science, a NLP or computational linguistics component will be a plus


  • - Good knowledge of French


  • Practical details: - Beginning of the thesis on January 1, 2022


  • - Full-time doctoral contract at LIG (Getalp team) for 3 years (salary: min 1768 € gross monthly)


  • Scientific environment: The thesis will be carried out within the Getalp team of the LIG laboratory ( https://lig-getalp.imag.fr/ ).


  • The recruited person will be welcomed into the team which offers a stimulating, multinational and pleasant working environment.


  • The means to carry out the doctorate will be provided both with regard to missions in France and abroad and with regard to equipment (personal computer, access to GPU servers of the LIG, Jean Zay computing grid of the CNRS ).


  • How to apply Candidates must hold a Master's degree in computer science in Machine Learning or in automatic natural language processing (obtained before the start of the doctoral contract).


  • They should have a good knowledge of machine learning methods and ideally experience in collecting and managing corpora.


  • They must also have a good knowledge of the French language.


  • Applications must contain: CV + cover letter / message + master's notes + letter (s) of recommendations; and be addressed to Benjamin Lecouteux ( [email protected] ), Gilles Sérasset ( [email protected] ) and Didier Schwab ( [email protected]) )


Thesis offer: Operational population of knowledge bases and Neural Networks, LIG (Grenoble)



  • From: [email protected] Date: Fri, 8 Oct 2021 17:53:50 +0200


  • Thesis subject within the framework of the Popcorn project (collaborative project with two companies) supervised by Benjamin Lecouteux, Gilles Sérasset and Didier Schwab (Grenoble IT Laboratory, Group Study in Machine Translation / Automated Language Processing and of the Word)


  • Title : Operational population of knowledge bases and neural networks


  • The project addresses the problem of semi-automated enrichment of a knowledge base through automatic text analysis.


  • In order to obtain a breakthrough innovation in the field of Treatment Automatic Natural Language (NLP) for security clients and defense, the project focuses on the treatment of French (even if the approaches adopted will subsequently be generalizable to other languages). The thesis work will address different aspects:


  • - Automatic annotation of text documents by the detection of mentions of entities present in the knowledge base and their semantic disambiguation (polysemy, disambiguation);


  • - The discovery of new entities (people, organizations, equipment, events, places), their attributes (age of a person, equipment reference number, etc.), and relationships between entities (a person works for a organization, people involved in an event, ...).


  • Particular attention will be given to being able to adapt flexibly to changes in ontology, taking into account the place of the user and the analyst for the validation / capitalization of the extractions carried out.


  • The project focuses on the following three areas of research: - Generation of synthetic textual data from texts of reference;


  • - Recognition of entities of interest, associated attributes and relationships between entities.


  • - The semantic disambiguation of entities (in case of homonymy by example)


  • Required profile: - Solid experience in programming & machine learning for Automatic Language Processing (NLP), in particular deep learning


  • - Master in Machine Learning or IT, a TAL component or computational linguistics will be a plus


  • - Good knowledge of French


  • Practical details: - Beginning of the thesis on January 1, 2022


  • - Full-time doctoral contract at LIG (Getalp team) for 3 years (salary: min 1768 € gross monthly)


  • Scientific environment: The thesis will be carried out within the Getalp team of the LIG laboratory ( https://lig-getalp.imag.fr/ ).


  • The recruited person will be welcomed into the team which offers a stimulating, multinational and pleasant working environment.


  • The means to carry out the doctorate will be ensured both in terms of concerns missions in France and abroad as regards the equipment (personal computer, access to the GPU servers of the LIG, CNRS Jean Zay calculation grid).


  • How to apply Applicants must have a Masters in Computer Science in Machine Learning or automatic natural language processing (obtained before the start of the doctoral contract).


  • They must have a good knowledge of learning methods automatic and ideally experience in collecting and managing corpus.


  • They must also have a good knowledge of the language French.


  • Applications must contain: CV + letter / motivation message + master's notes + letter (s) of recommendations; and be addressed to


  • Benjamin Lecouteux ( [email protected] ), Gilles Sérasset ( [email protected] ) and Didier Schwab ( [email protected] )


Offer of these: Extraction of information from scientific articles, University of Franche-Comte



  • Dear colleagues, The CRIT laboratory of the University of Franche-Comté (Besançon) offers a doctoral contract in NLP on the extraction of information from scientific articles.


  • The thesis will be funded by the ANR InSciM project "Modelling uncertainty in science". Thank you for disseminating this offer very widely to potentially interested students.


  • Bien cordialement, Iana Atanassova


  • A 3-year PhD fellowship in Natural Language Processing is available at the CRIT - Centre Tesnière laboratory, University of Bourgogne Franche-Comté, France.


  • Start date : November 2021


  • Location : Besançon, France


  • Contact : Iana Atanassova - [email protected]


  • Research Team and Doctoral School The CRIT - Centre Tesnière laboratory (http://tesniere.univ-fcomte.fr/), at the University of Bourgogne Franche-Comté, Besançon, France, develops linguistic approaches in NLP for information extraction, IR and text mining.


  • The PhD thesis will be part of the French ANR project « InSciM – Modelling Uncertainty in Science » starting in October 2021 with scientific coordinator Dr Iana Atanassova.


  • Doctoral school: LECLA (https://www.adum.fr/as/ed/actu.pl?site=lecla)


  • The PhD thesis can be written in French or in English. The PhD candidate will have the possibility to obtain the «European Doctorate» label.


  • Project Context and Motivation As the world undergoes profound transformations, science is highly solicited, such as in the context of health crises (Covid-19), the reflection and dialogue on climate change, ecological and energy transformations, monetary transformation, humanitarian issues, or geopolitical crises. The perception of uncertainty in scientific discourse is therefore an important issue for all scientific activities.


  • In science, the production of new knowledge uses rigorous methodological approaches based on the object of study and its disciplinary field.


  • However, the use of tools or observations that produce a margin of error, as well as the use of abductive and inductive reasoning imply the presence of uncertainty, which can be specific to each discipline, linked to the object of the study and the methodologies that are used. Uncertainty in science is an integral part of the research process.


  • This project aims to study uncertainty in science through ontological and linguistic modelling of this notion from datasets of articles in different disciplines.


  • The objectives are to propose a linguistic model of the expression of uncertainty in scientific articles, in order to propose a tool to identify and classify these phenomena present in papers in different disciplines in Social Sciences and Humanities (SSH) and in Science, Technology, and Medicine (STM).


  • Objectives and Scope of the PhD Thesis The major objectives of the research during the PhD fellowship include: - on the theoretical level, designing a disciplinary linguistic ontology of uncertainty and the ways in which it is expressed in publications, taking into account disciplinary differences: in experimental sciences, in exact sciences, and in humanities and social sciences.


  • - on the application level, constructing reference corpora - Gold Standards - that will be annotated to identify the textual segments carrying uncertainty, as well as the objects on which the uncertainty applies (results, values, models, . . . ).


  • These reference corpora will be made available to the community to allow, among other things, the training of machine learning or deep learning models.


  • The first task will be to prepare the datasets of full-text scientific papers in English and in different disciplines. We will need to identify APIs and datasets in full text that are available for academic use, and to harvest and prepare these data for further processing.


  • As the full text data can be available in different formats, e.g. XML, LaTeX, HTML, we will need to identify the tools and develop the processing pipelines to obtain a unified representation of all documents.


  • The production of a Gold Standard, an annotated dataset with concepts of uncertainty, will be a major result from this project. The Gold Standard is a large-scale dataset that is annotated using the ontology of uncertainty and where all measures have been taken to guarantee a high quality and consistency of the annotation.


  • Its creation requires the design of linguistic rules and knowledge-based methods for the initial annotation, and repeated manual evaluations of annotated samples in order to control the quality of the output in terms of precision and recall. This dataset will then be used to train language models.


  • http://ontologia.fr/TOTh/Conference/TOTh2018/TOTh_2018.pdf


  • Requirements - Master’s degree in one of the following disciplines: Natural Language Processing, Linguistics, Computer Science, Information Science, Data Mining or related disciplines.


  • - Experience in working with research articles and understanding of the research process.


  • - Strong motivation, willingness to learn new skills and to work in academy, good communication skills.


  • - English : spoken and written.


  • - Computer programming skills and experience in NLP are a plus.


  • Application Candidates must submit their applications by e-mail to Iana Atanassova ( [email protected] ). The applications must include the following documents:


  • - CV and cover letter


  • - Diplomas (including exam scores)


  • - Names and contact information for possible reference persons


  • - Optionally, an example of your previous work (if possible related to NLP), e.g. a master’s thesis, a research paper, etc.


  • Applications will be considered on the fly. It is advisable to apply as soon as possible. Selected candidates will be invited for interviews in October 2021.


Job : Opening for Chairs in AI at Inria



  • Dear Sir/Madame, Inria, the illustrious French national research institute for digital science and technology, is looking to recruit up to 3 high-level senior or young senior scientists (post-tenure or equivalent at the least, internationally known, influential and productive) in Artificial Intelligence (in the broadest sense).


  • This recruitment is part of the "Choose France" program of the national AI strategy, "AI for Humanity", launched by President Macron in 2018.


  • Inria is managing the Choose France recruitment on behalf of all French research and higher education institutions, in its role as coordinator of the research component of the national AI strategy.


  • The description and application requirements for the Chair in AI are available at the following addresses:


  • English: https://www.inria.fr/en/choose-france-inria-recruiting-3-high-level-scientists-ai https://www.inria.fr/sites/default/files/2021-10/AI%20chairs%20for%20Choose%20France.pdf


  • French: https://www.inria.fr/fr/trois-chaires-en-ia-chez-inria-pour-choose-france


  • We would very much appreciate your help in forwarding this announcement to relevant contacts within your network.


  • Best regards, Frédérique Vidal On behalf of Justine Cassell, Head of the Search Committee


  • Mission Intelligence Artificielle Coordination du Plan National de Recherche en Intelligence Artificielle


  • https://www.inria.fr/fr/la-mission-ia-un-programme-national-de-recherche-en-intelligence-artificielle


  • http://inria.fr/mission-ia Centre d’expertise de Paris | PMIA – GPAI (https://gpai.ai/fr/)


  • Domaine de Voluceau - Rocquencourt B.P. 105 - 78153 Le Chesnay France