Hello, you will find attached an interdisciplinary thesis offer (Soil Science, Computer Science and Environmental Economics) for which we are looking for a candidate.
Building on previous work (ERDF-Artisol Region project) which has made it possible to establish a soil rating system and a digital mapping of the Occitanie region, the thesis will focus on the development of a participatory approach based on interactive visualization tools. The aim is to enable future local users to get involved in the mapping process of soil quality assessment.
The thesis will be carried out at the UMR LISAH in Montpellier under the supervision of Philippe Lagacherie, in co-direction with Evelyne Lutton, Nadia Boukhelifa (UMR MIA Paris, for the collective AI and visualization part) and Léa Tardieu (UMR TETIS Montpellier).
To apply, send CV, cover letter, contacts of referents and M2 note before September 6, 2021 to [email protected].
Best regards, Evelyne Lutton.
Hi all Here is a PostDoc offer of 12 months to be filled (as soon as possible) within my team in Orange (Châtillon) on an exciting subject: forecasting international roaming traffic with exogenous variables. English-speaking candidates are welcome to apply too.
Applications can be made online via: https://orange.jobs/jobs/offer.do?joid=102232&lang=EN
Excuse me in advance for the possible multiple broadcasts.
Kind regards Nour Eddine Yassine NAIR BENREKIA Data scientist in Orange (Paris)
Your role Your role is to perform post doc work on: "Forecasting methods with exogenous variables for international roaming traffic"
International roaming allows subscribers of a mobile network operator to continue to access their services (voice, sms, data) when travelling abroad, using the resources of a local operator in the country they are visiting. This gives rise to a remuneration of this local operator by the nominal operator of roaming subscribers. This remuneration is framed by commercial agreements between the partner operators according to the total volume of roaming traffic over the year. For a global operator, the costs of inter-operator repayment can reach several billion euros per year. The optimization of these costs is therefore crucial for Orange but two major questions arise:
1) how best to negotiate the cost functions of commercial agreements with each of the partner operators?
2) how to distribute the roaming subscribers abroad of orange subsidiaries among the local operators in the countries visited?
To answer these questions, the wholesale roaming manager needs two types of traffic forecasts. In the 1st phase "negotiation", it is based on traffic forecasts for the coming year and in the 2nd phase "decision of the steering policy" on monthly traffic forecasts. Having good traffic forecasts for subsidiaries is therefore the cornerstone of optimizing traffic orientation decisions.
Within the group, various experiments have shown that: 1) these time series are subject to external events that can have a very strong influence on the volume of traffic and 2) integrating this external information improves the quality of the forecast of this traffic. During the post-doc, we will focus on cross-learning approaches that allow us to take advantage of existing knowledge in other time series and thus improve prediction performance. This knowledge may be effects of events observed in other countries. For example, data on a rare event such as the Olympic Games, which have taken place in the past in some countries, can anticipate the effect they will have in the future in another country. This type of approach became popular after the M4 competition [1] where two approaches stood out from traditional approaches: ES-RNN [3] which learns a neural network from previously pre-processed series using exponential filtering, and FFORMA [2] which learns several models and then combines them with weights learned by an xgboost from the characteristics of the series.
The challenge of this post-doc is to explore the state of the art of these cross-learning approaches and to develop a new forecasting approach that is as efficient as possible. This approach must be able to predict the traffic sent by subsidiaries monthly in each country for the three services and to estimate the uncertainty in each forecast.
[1] Makridakis, S., Spiliotis, E., & Assimakopoulos, V. (2020). The M4 Competition: 100,000 time series and 61 forecasting methods. International Journal of Forecasting
[2] Montero-Manso, P., Athanasopoulos, G., Hyndman, R. J., & Talagala, T. S. (2020). FFORMA: Feature-based forecast model averaging. International Journal of Forecasting
[3] Smyl, S. (2020). A hybrid method of exponential smoothing and recurrent neural networks for time series forecasting. International Journal of Forecasting
Your profile The desired profile is BAC+8: a PhD in machine learning and/or statistics, ideally dedicated to the prediction of time series. Experience in the field of statistics and/or machine learning within an R&D team would be a plus.
The post-doctoral student must have a good knowledge of statistics and mathematics. Knowledge of statistical learning is required.
Programming skills are necessary: mastery of a scripting language (at least) dedicated to data analysis (R, Matlab, Python with Scikit-learn library...). Knowledge of an object-oriented language would be appreciated. Strong motivation, synthesis skills, to write well and present the work (English) and to integrate into a team are also required.
The most of the offer You will work on a use case with high economic stakes (financial costs of the order of several billion euros per year) and you will have the opportunity to quickly integrate your contributions into the Orange ecosystem and to value them in the form of scientific publications.
Entity Orange is a key player in digital innovation. In an information and communication technology sector that is experiencing an upheaval in its value chain, with the multiplication of players and the appearance of new business models, innovation is a major lever of growth for the Orange Group.
Orange's ambition is to make innovation useful and accessible to as many people as possible. By bringing together activities around the creation of strategic innovations, research and the implementation of technical and data policies for the Orange Group, the Orange Innovation (OI) division is the driving force behind this innovation.
By joining Orange within Orange Innovation IT and Services, digital infrastructure & end-to-end Secure Environments, you will join the team:
* Business and Research Artificial INtelligence solutions (BRAIN) within the "Data Intelligence and Algorithms" department of Orange Labs Services. This team of AI experts has real algorithmic and operational know-how. This team's mission is to help our business departments and subsidiaries develop the use of artificial intelligence and big data in their business, to ensure a technological watch around AI and to industrialize internal solutions.
You will work closely with the team: * Modeling and Statistical Analysis (MSA) within the Green Data Modeling (GDM) department. This team is composed of experts in network traffic modeling, both fixed and mobile. The objectives of these activities are: (i) to improve our knowledge of network usage and (ii) to estimate and optimise the infrastructure costs of networks using next-generation technologies (optical fibre, 4G and 5G mobile networks, etc.) by geographical area.
Hello, as part of a project of the Institut Polytechnique de Paris, we offer 2 theses on the following themes: - Modeling and safety evaluation of a complex system:
https://cloudgravity.github.io/files/jobs/ceres-lot1_phd.pdf
- Digital twin for the security of technical building management: https://cloudgravity.github.io/files/jobs/ceres-lot2_phd.pdf
We also offer a position as a platform engineer: https://cloudgravity.github.io/files/jobs/ingenieur_pf-ceres_fr.pdf
The topics and the application procedure are detailed in the links below. The theses will start at the earliest in October 2021.
Hello, as part of a recovery plan project, we offer 2 theses on the following themes:
- Security quantification in a 5G network:
https://cloudgravity.github.io /files/jobs/beyond5g-t3.2_phd. pdf
- 5G Slice Security: https://cloudgravity.github.io/files/jobs/beyond5g-t3.3_phd. pdf
The topics and the application procedure are detailed in the links below. The theses will start at the earliest in October 2021.
Hello, You will find attached a thesis proposal to the CEA LIST as part of an ANR project that aims to offer a personalization of online services more respectful of the privacy of users.
The topic focuses on the use of deep learning to do profiling only on mobile devices so as not to expose users' personal data.
Best regards, Adrian Popescu
Contact Lettre de motivation et CV à envoyer à: [email protected]
Hello, INSA Hauts-de-France (INSA HdF) is recruiting a contract teacher-researcher in computer science, equivalent mcf, on a one-year fixed-term contract fromSeptember 2021.
teaching, the interventions will be provided in the Computer Science department of INSA Hauts-de-France which includes a bachelor's degree in Computer Science, a master's degree in Computer Science and 2 engineering specialties: It and Cybersecurity (FISE) and Informatics (FISA).
The lessons will mainly concern object-oriented analysis and development, requiring a good knowledge of the Java language. The candidate will also be required to participate in undergraduate courses in fundamental algorithmics and programming, with the support of Python or the C language.
In research, the teacher-researcher will be integrated into the LAMIH UMR CNRS 8201 laboratory, in the computer science department, with priority in the field of Operational Research and Aid to Decision.
Knowledge in the models, methods and tools of Operational Research (modeling mathematical, meta-heuristics), in discrete event simulation and decision support would be appreciated.
The research will be carried out within the framework of the International Research Project (IRP) «Research
Operational and It in Transport, Mobility and Logistics»ROI-TML of the CNRS, put in
the LAMIH and the Centre Interuniversitaire de Recherche sur les Réseaux d'Entreprise, the Logistics and Transport (CIRRELT - Université de Montréal, Canada).
The complete job description is available on the university's website: https://www.uphf.fr/sites/default/files/pdf/drh/profil_ 27_mcf_article_19_insa_lamih. pdf
The documents to be provided are indicated after the job description.
Applications should be sent to [email protected] and in copy to [email protected].
Attention: the application deadline is August 08, 2021
Cifre thesis, Télécom Paris : "Anomaly detection for large-scale and heterogeneous data from production lines" (“Anomaly detection for large-scale and heterogenous data of production lines”)
Anomaly detection is a branch of artificial learning that aims to identify abnormal and aberrant events. Although it knows many applications, it is still underemployed in the industry, while it can provide an essential tool for monitoring and improving production lines.
In this context, the main objective of this thesis is to develop a methodology for detecting anomalies for large data measured in large quantities at variable frequencies and also having a hierarchical structure; such data – increasingly common in the industry – is still an open challenge.
Indeed, learning tools must not only detect abnormal and aberrant manufacturing parameters with the utmost reliability but also provide an interpretation of the forecast that can be useful in improving the manufacturing process.
The thesis will explore different paths such as data depth and classification of a class. It will be carried out as part of a close collaboration between the company Valeo (production site l'Isle d'Abeau) and Télécom Paris (Institut Polytechnique de Paris). The methods developed will be applied to databases from a set of production lines of the latest generation.
Supervisors: Pavlo Mozharovskyi – LTCI, Télécom Paris, Institut Polytechnique de Paris Florence d'Alché-Buc – LTCI, Télécom Paris, Institut Polytechnique de Paris
Expected qualifications: - Master in Statistics / Data Science / Machine Learning / Artificial Intelligence / Engineering degree with specialization in these fields.
- Very good level in at least one of the machine learning programming languages: R / Python, C / C++, or similar.
Place: - Télécom Paris (Campus de Institut Polytechnique de Paris, 25 km from Paris), 19 place Marguerite Perey, F-91120 Palaiseau. - Valeo, site L'Isle d'Abeau.
Deadline : - Until recruitment, but no later than August 31, 2021.
To apply: Send the following documents to
pavlo.mozharovskyi@telecom- paris.fr and florence.dalche@telecom-paris. en: - cover letter;
- curriculum vitae;
- copy(s) of diploma(s);
- name(s)/email(s) of less than one referrer. -------------------------------------------------------------------------------- Pavlo Mozharovskyi Teacher-researcher
LTCI, Telecom Paris, Institut Polytechnique de Paris 19 place Marguerite Perey, F-91120, Palaiseau, France
https://perso.telecom-paristech.fr/mozharovskyi/