41 Postdoctoral Fellowship at Inria, France

The present postdoctoral project is part of a collaborative project between the Atlantis project-team from the Inria Research Center at Université Côte d’Azur,  (2) Thales R&T in Palaiseau, France, which is the main multidisciplinary research center of Thales group, and one of the main world players in professional electronics (3) and Nanoe in Ballainvilliers, France, whcih is  a SME (Small and Medium Entreprise) specialized  in development and fabrication of innovative ceramic materials for the industries. Atlantis is  a joint project-team  between Inria and  the Jean-Alexandre Dieudonné Mathematics Laboratory at  Université Côte d’Azur. The team  gathers applied mathematicians and  computational scientists who are collaboratively undertaking  research activities aiming at the design, analysis, development and  application of innovative numerical methods for systems of  partial differential equations (PDEs) modelling nanoscale light-matter interaction problems. In this context, the team is  developing  the   DIOGENeS  [https://diogenes.inria.fr/]  software suite,  which  implements  several Discontinuous  Galerkin  (DG)  type methods tailored to the systems  of time- and frequency-domain Maxwell equations  possibly coupled  to  differential  equations modeling  the behaviour of propagation  media at optical frequencies.  DIOGENeS is a unique  numerical   framework  leveraging   the  capabilities   of  DG techniques  for  the simulation  of  multiscale  problems relevant  to nanophotonics and nanoplasmonics.

This is a collaborative research project involving 2 partners.

Jing-Rebecca Li, Researcher (Applied Mathematics); Inria-Saclay, Équipe Idefix (Medical Imaging, Applied Mathematics); ENSTA Paris, Unité de Mathématiques Appliquées (UMA).  Institut Polytechnique de Paris.

Anne Kerbrat, Medical Doctor (Neurologist) and Emmanuel Caruyer, Researcher (Medical Imaging);  CHU de Rennes, Service de neurologie; Equipe Empenn, Univ Rennes, CNRS, Inria, Inserm, IRISA UMR 6074.

The postdoc will be based at Inria-Saclay, with regular visits to Rennes. 

This project involves Statify and  Grenoble Institute of Neuroscience (GIN). T. Christen (CR Inserm) will be the main contact  at GIN. T. Christen has established national and international collaborations on the topic of MRF and has an ongoing partnership with Philips on clinical MRF developments. The additional expertise of P. Rodrigues (CR Inria)  and  F. Forbes  (DR Inria), in particular in inverse problems in a MRI context (Rodrigues et al 2021, Jallais et al. 2021, Boux et al. 2021), will be provided as MRF requires better theoretical characterizations for practical use.

This project aims to develop concepts and a toolchain based on artificial intelligence to find the ‘best’ preconditioner for a Krylov subspace iterative solver for the solution of the linear system coming from a discretized Stokes problem. We will focus in particular on a geometric multigrid preconditioner for the (1,1)-block, here, the Laplace operator, to ensure the scalability of the solver when passing to large-scale problems. We will extend the work by J. Schmitt et al. (Constructing Efficient Multigrid Solvers with Genetic Programming, Gecco ’20, 2020), in which a grammar-guided approach for the optimization of multigrid solvers for matrices from a finite difference discretization of the Laplacian operator is formulated. The particular difficulty for the extension to saddle point systems is the large number of possible combinations of solution techniques. These include, for example, preconditioned Krylov subspace solvers that make use of the block structure of the system, Schur complement approaches, deflation and augmentation techniques, or also monolithic multigrid solvers on the whole block system.

The Inria centre at Université Côte d’Azur includes 37 research teams and 8 support services. The centre’s staff (about 500 people) is made up of scientists of different nationalities, engineers, technicians and administrative staff. The teams are mainly located on the university campuses of Sophia Antipolis and Nice as well as Montpellier, in close collaboration with research and higher education laboratories and establishments (Université Côte d’Azur, CNRS, INRAE, INSERM …), but also with the regiona economic players. With a presence in the fields of computational neuroscience and biology, data science and modeling, software engineering and certification, as well as collaborative robotics, the Inria Centre at Université Côte d’Azur  is a major player in terms of scientific excellence through its results and collaborations at both European and international levels.

Inria is the French national research institute for digital science and technology. World-class research, technological innovation and entrepreneurial risk are its DNA. In 215 project teams, most of which are shared with major research universities, more than 3,900 researchers and engineers explore new paths, often in an interdisciplinary manner and in collaboration with industrial partners to meet ambitious challenges. As a technological institute, Inria supports the diversity of innovation pathways: from open source software publishing to the creation of technological startups (Deeptech). Strengthening partnerships with the State’s Security-Defense sphere is a strategic priority for Inria. In this context, Inria is in the process of creating a Defense and Security Department whose mission is to federate, in the most readable and operational way possible, the various Inria actions that can meet the digital needs of the Defense and Security sphere.

Inria, l’institut national (public) de recherche en sciences et technologies du numérique, s’assume au travers de son contrat d’objectifs et de performance 2019-2023 comme un outil de la souveraineté et de l’autonomie stratégique numérique de la Nation. Le renforcement des partenariats avec la sphère Sécurité-Défense de l’État est une priorité stratégique. Dans ce contexte, Inria est en train de créer un Département Défense et Sécurité dont la mission est de fédérer, de façon la plus lisible et opérationnelle possible, les différentes actions d’Inria pouvant répondre aux besoins numériques de la sphère Défense et Sécurité.

Wave propagation phenomena are ubiquitous in our modern society and manifest themselves in various contexts such as telecommunications, imaging techniques (medical, seismic), information processing (photonics), energy harvesting and health (biomedical applications). Numerical modeling plays an increasing role in understanding complex wave propagation phenomena or for mastering the wave interactions with our environment. Many numerical methods have been designed for the numerical solution of wave PDE models for unsteady or harmonic propagation regimes, i.e., for time-domain and frequency-domain formulations of these models. In particular high order discontinuous finite element methods, named as Discontinuous Galerkin (DG) methods or Hybridized DG (HDG) have shown significant benefit in terms of accuracy and have become popular. The resulting discretized problem reduces to the solution of a large sparse linear system of equations.

We’re looking for someone with experience in neuroimaging of social cognition who can contribute meaningfully to this new era of hyperscanning research. Particularly, we are seeking someone who can assist in the design of rigorous naturalistic hyperscanning paradigms which investigate social phenomena such as collaboration, rapport, learning, and performance, and the analysis of the resulting data, preferably using newer machine learning analytic tools. In addition to investigating understudied phenomena of social cognition, these findings will eventually be utilized to improve a virtual peer – a conversational agent that builds bonds with children in the service of learning.

The applicant must have strong knowledge of robotics, in particular exoskeletons (dynamics of exoskeletons and human/humanoids, whole-body control), as well as strong C++ skills (the digital human simulation and exoskeleton simulation is programmed in C++ and real-time performance is required). Alternatively, a strong knowledge of biomechanics simulation and control of active exo-skeletons.

Networks such as modern telecommunications networks or distributed embedded systems are permanently monitored to allow identification of failure situations; thousands of new data points reflecting the system state changes are generated every minute. Even if faults are rare events, they can easily propagate driven by local and remote dependencies, which makes it challenging to distinguish causes from effects among the thousands of highly correlated alerts.
A timely automated identification and root cause analysis (RCA) of the origins of performance issues allows executing the most adequate corrective actions and preventing their further propagation. In general, RCA is a hard problem, because it requires a deep knowledge of cause-effect dependencies among many features, physical and logical components the network nodes. In a data driven approach, where most of this knowledge is unavailable a priori, a major difficulty emanates from hidden or unknown variables. Furthermore, even in a fully observable system we are faced with the combinatorial explosion of potential cause-effect dependencies and the difficulty to collect enough information for distinguishing causality from spurious correlations.

 

(18) Postdoctoral Fellowship Position

Postdoc Fellowship Position summary/title: Post-Doctorant F/H Analyse d’images et de vidéos

Deadline : 2023-10-23

View details & Apply

 

We consider the sea ice model neXtSIM, developed at Nansen Environmental and Remote Sensing Center (Bergen, Norway) and Institute of Environmental Geosciences (Grenoble, France), a novel model aimed at modelling complex sea ice dynamic in a large range of scales, from local short-term predictions to global climate prediction simulations. The specificity of \nextsim is the use of a purely Lagrangian advection formalism on fully unstructured meshes coupled with a novel rheologic frameowrk, that has given promising results. Such a Lagrangian approach results in strongly deformed meshes over time, in particular in the vicinity of cracks due to the ice displacement. A remeshing step is thus necessary to locally replace stretched or invalid mesh elements and restore the quality of the mesh. However, unlike the rest of the code that was parallelised recently for distributed memory architectures using MPI, the remeshing stage remains sequential, thus strongly impacting the performance of the code. Besides, the current remeshing does not yet take advantage of modern anisotropic mesh adaptation techniques, that aim at optimising the size and orientation of the mesh elements to minimise a certain numerical error estimate, and makes it possible to reduce the computationnal cost while increasing the accuracy. The goal of the collaboration is to leverage these methods to accelerate simulations to be able to repeat easily large-scale high-resolution simulations and study the ice trajectories statistically.

This post-doctoral position will be supported by the HE Flute project.   While this position will be in the MAGNET team in Lille, we will collaborate with the several European project partners. While AI techniques are becoming ever more powerful, there is a growing concern about potential risks and abuses. As a result, there has been an increasing interest in research directions such as privacy-preserving machine learning, explainable machine learning, fairness and data protection legislation. Privacy-preserving machine learning aims at learning (and publishing or applying) a model from data while the data is not revealed. Notions such as (local) differential privacy and its generalizations allow to bound the amount of information revealed. The goal of the multi-disciplinary FLUTE project is to advance and scale up data-driven healthcare by developing novel methods for privacy-preserving cross-border utilization of data hubs. Advanced research will be performed to push the performance envelope of secure multi-party computation in Federated Learning, including the associated AI models and secure execution environments. The INRIA MAGNET team (and hence the recruited collaborators) will contribute to this project among others by researching machine learning algorithms and multi-party protocols with improved scalability in the context of medical data, e.g., by exploiting data sparsity.   This research will involve both theoretical and more applied components.  As coordinator INRIA will also contribute to the integration of the software developed in the FLUTE project (and the complementary FLUTE project).

This post-doctoral position will be supported by the HE Trumpet project.  While this position will be in the MAGNET team in Lille, we will collaborate with the several European project partners. While AI techniques are becoming ever more powerful, there is a growing concern about potential risks and abuses. As a result, there has been an increasing interest in research directions such as privacy-preserving machine learning, explainable machine learning, fairness and data protection legislation. Privacy-preserving machine learning aims at learning (and publishing or applying) a model from data while the data is not revealed. Notions such as (local) differential privacy and its generalizations allow to bound the amount of information revealed. The overall goal of the TRUMPET project is to research and develop novel privacy enhancement methods for Federated Learning, and to deliver a highly scalable Federated AI service platform for researchers, that will enable AI-powered studies of siloed, multi-site, cross-domain, cross-border European datasets with privacy guarantees that exceed the requirements of GDPR. 

A two-year postdoctoral position is available in team BEAGLE at Inria Lyon. At BEAGLE, we study cellular systems through computational models and bioinformatic data analysis, in order to better understand their organizational principles. In our team, we value working together, being innovative, and providing opportunities for professional and personal growth. The position is within the framework of project CLAIRE, a collaboration between Inria’s team BEAGLE and research groups at the Cancer Research Center of Lyon (CRCL). The overall goal of this project is to better understand the emergence of drug tolerance during treatment of lung cancer. You will be advised by Anton Crombach (Inria) and Sandra Ortiz-Cuaran (CRCL), whilst working together with other members of the project. You will be located at Inria’s La Doua site and regularly visit the CRCL.

The overall objective of the DIANA project-team is to design, implement and evaluate advanced networking architectures. To do so, the team works to provide service transparency and programmable network deployments in the context of both wired and next generation wireless cellular networks. The team’s methodology includes advanced measurement techniques, design and implementation of architectural solutions, and their validation in adequate experimental facilities. Recent contributions of the team include: Enhanced Transport-Layer Mechanisms for Multi-Access Edge Computing-Assisted Cellular Networks, Bencharmking Mobile Networks from the Viewpoint of Video Streaming QoE, Introducing Fidelity in Network Emulation, and Enhanced Ray Tracing Techniques for Accurate Estimation of Signal Power.

The vision of the dAIEDGE Network of Excellence is to strengthen and support the development of the dynamic European edge AI ecosystem under the umbrella of the European AI Lighthouse and to sustain the advanced research and innovation of distributed AI at the edge as essential digital, enabling, and emerging technology in an extensive range of industrial sectors.

Current research in control theory is focused on utilizing the dynamics of large networks, such as traffic and transportation infrastructure, multi-agent systems, epidemic spreading, and electrification of vehicles, among others. Dealing with a high number of state variables that describe the nodes or edges of these networks, as well as significant uncertainties, calls for new reduction methods that are suitable for networks and can effectively navigate through different scales. Traditionally, the control community has tackled the control of large-scale network systems by seeking distributed control algorithms, where each node applies a control loop locally based on its own information. However, this approach often requires access to local information that may not be available for all nodes in the network. As an alternative to this decentralized approach, the objective of this thesis is to explore “the continuation method” proposed in the ERC Scale-FreeBack project [2,3,4].The continuation method represents a novel approach for approximating large-scale networks described by sets of ordinary differential equations (ODEs) using partial differential equations (PDEs). It is particularly applicable to high-dimensional spatially distributed ODE systems. Examples of such systems include urban traffic networks, ring oscillators, electrical networks, multi-agent robots, gas dynamics, fluid density studies, and electro-mobility networks. In essence, the continuation process can be understood as the inverse of the typical space discretization of PDEs, where a large set of ODEs is used. It begins with a set of coupled or uncoupled high-dimensional ODEs distributed in space, which is then approximated by a PDE using finite differences. Once this approximation is established, boundary controllers can be designed for the PDE instead of controlling each individual control system, or the PDE can be used for analysis purposes [5]. The control designed at the PDE level is subsequently applied to the real system by discretizing it back to the original ODE system.

This postdoctoral position is funded in the framework of the PEPR “5G et reseaux du futur” of the national recovery plan, in the collaborative project FUNDS lead by Prof François Baccelli. The student will be supervised by Jean-Marie Gorce, professor at INSA Lyon and affiliated to Inria Lyon, in collaboration with François Baccelli. The key objective of the position is to elaborate on a new theoretical framework to design new communication rules in wireless multi-user context with (almost) deterministic delay and reliability guarantees. This objective is required to sustain the development of machine-to-machine (M2M) communications in the context of robotics, industry4.0 and vehicular communications. The candidate will primarily work on elaborating a theoretical framework relying upon information theory, queuing theory and stochastic network calculus.

The partners of the Concace joint project, namely Airbus C R & T – Cerfacs and Inria,  have been studying and developing Krylov subspace solvers and associated software packages for the past few years for the parallel solution of large systems of linear equations. In this work, we will be interested in the solution of long sequences of symmetric non-Hermitian problems arising for instance from the solution of the Helmholtz equation in heterogeneous media in the context of a sensitivity or parametric study.  A key numerical component to ensure the robustness and the efficiency of these numerical methods is the so-called preconditioner, whose definition requires some a priori knowledge on the properties of the linear system or of the underlying problem. Classical approaches consist in exploiting this algebraic or analytic information. In this work, we intend to investigate how machine learning techniques could be used to improve a prescribed preconditioner or to fully learn one.

Therefore, the goal of this post-doctoral position is twofold. We will first focus on developping fast and adaptive ML methods that can detect and update the model to cope with the variation of attack behaviors. Furthermore, we expect the trained detection model to be interpretable. It can evaluate the informativeness of attributes in security reports. It can reveal the causal relationship between these raw attributes and the detection and classification results.  The post-doc researcher will be hosted at Rennes and may be required to travel to work with RESIST team regularly. Travel expenses will be covered within the limits of the scale in force.

Le centre de recherche Inria Saclay – Île-de-France, créé en 2008, accueille 450 scientifiques et 60 membres des services d’appui à la recherche. Les scientifiques sont organisés en 28 équipes de recherche dont 23 sont communes avec des partenaires du plateau de Saclay. Le centre Inria Saclay – Île-de-France est un acteur essentiel de la recherche en sciences du numérique sur le plateau de Saclay. Il porte les valeurs et les projets qui font l’originalité d’Inria dans le paysage de la recherche : l’excellence scientifique, le transfert technologique, les partenariats pluridisciplinaires avec des établissements aux compétences complémentaires aux nôtres, afin de maximiser l’impact scientifique, économique et sociétal d’Inria.

Over the last 30 years the graphics community has actively worked on the simulation of dense fibrous assemblies such as human hair or animal fur. Whether it is for the realistic animation of characters by the entertainment industry, for its applications in cosmetology, or for the design of mechanical structures, simulating accurately such kinds of systems remains challenging and computationally costly. Furthermore, a rich underlying phenomenology arises from the interaction between fibers, driven by the interplay between geometry, frictional contact, and elasticity. Due to the large number of degrees of freedom and the strong coupling between the length-scales at which fibers interact, innovative approaches are crucial.

Cybersecurity is a major concern everywhere with the growth of connected devices that are beyond common computers. To circumvent these problems, decades of research and development have led to build new techniques and tools to fight back against the attacks on the Internet. Nonetheless, the number of attacks and their magnitude still grow. The attack surface continues to increase along with the number of connected devices but also due to the number of applications, services or software that today make the IT ecosystem far from its origin.Techniques used by both attackers and defenders evolve to complex mechanisms. For example, this leads to the massive use of encryption to avoid data leaks but simultaneously attackers benefit from encryption to hide their own activities. Multiple steps attacks also requires to analyze numerous sources of data. As a result, intrusion detection methods relying on artificial intelligence have been investigated both in research and in industry.

This post-doc will work in the TRIPOP team, which is a joint research team of the INRIA centre at Grenoble-Alpes University and the Jean Kuntzmann Laboratory (LJK). The team is mainly interested in the modelling, mathematical analysis, simulation and control of non-smooth dynamic systems. Non-smooth dynamics concerns the study of the temporal evolution of systems that are not smooth in the mathematical sense, i.e. systems that are characterised by a lack of differentiability, either of the mappings in their formulations, or of their solutions with respect to time. The team is one of the few in the world to have brought together researchers in applied mathematics, control theory, nu- merical mechanics and scientific computing in the field of non-smooth dynamics. In mechanics, the main examples of non-smooth dynamical systems are multi-body systems with one-sided contact, Coulomb friction and impacts, or plasticity. The academic supervision of this post-doc will be carried out by Vincent Acary, INRIA Research Director, in close collaboration with Olivier Bru ̈ls, Professor of Mechanics at the University of Lie`ge, specialising in flexible multi-body systems. Visits to the aerospace and mechanics department of the University of Lie`ge’s Aerospace and Mechanical Multibody Systems Laboratory are also envisaged. Industrial support will be provided by Safran Tech, which is funding this project and will help to define the relevant industrial examples.

Machine learning (ML) explainability and ML verification are typically seen as orthogonal concerns. Research in ML explainability is mainly carried out in the ML community, while research in ML verification is mostly carried out in the formal methods community, without much communication and exchanges between these research areas. We advocate for closing this gap. In particular, on the one hand, we plan to employ ML verification approaches to aid ML explainability: we aim to design and leverage correct-by-construction ML model approximations to derive explanations for the model behavior.On the other hand, we plan to leverage ML explainability techniques to guide ML verification, notably to identify relevant ML model properties to verify. The postdoctoral researcher will work closely with and help supervise two PhD students of the ANTIQUE group working on topics related to this project.