A sparsity-promoting reconstruction algorithm for diffuse optical tomography based on a transport model

Séminaire le 8 Septembre 2015, 11h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
K. Prieto (formerly University of Manchester, since Aug 2015 Hokkaido University) and O. Dorn (University of Manchester) (presenting author)


Non-linear inverse problems with sparsity

Séminaire le 8 Septembre 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
P. Maass (University of Bremen)


Compressive Sensing as a tool for exploiting sparsity and incoherence in computational electromagnetics

Séminaire le 8 Septembre 2015, 09h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
A. Massa (presenting author), G. Oliveri, M. Salucci, N. Anselmi (all at ELEDIA, University of Trento, A. Massa as DIGITEO Chair holder L2S & CEA LIST)


S³: Bayesian Cyclic Networks, Mutual Information and Reduced-Order Bayesian Inference

Séminaire le 17 Juillet 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Robert Niven, University of New South Wales, Canberra, Australia


A branch of Bayesian inference involves the analysis of so-called "Bayesian networks", defined as directed acyclic networks composed of probabilistic connections. We extend this class of networks to consider cyclic Bayesian networks, which incorporate every pair of inverse conditional probabilities or probability density functions, thereby enabling the application of Bayesian updating around the network. The networks are assumed Markovian, although this assumption can be relaxed when necessary. The analysis of probabilistic cycles reveals a deep connection to the mutual information between pairs of variables on the network. Analysis of a four-parameter network - of the form of a commutative diagram - is shown to enable thedevelopment of a new branch of Bayesian inference using a reduced order model (coarse-graining) framework.

Bio: https://research.unsw.edu.au/people/dr-robert-niven

Multicast IP

Séminaire le 5 Juillet 2015, 12h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Antthony Busson


S³: The method of brackets

Séminaire le 3 Juillet 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Victor H. Moll, Department of Mathematics, Tulane University, New Orleans, USA


A new heuristic method for the evaluation of denite integrals is presented. This method of brackets has its origin in methods developed for the evaluation of Feynman diagrams. We describe the operational rules and illustrate the method with several examples. The method of brackets reduces the evaluation of a large class of denite integrals to the solution of a linear system of equations.

Bio: Victor H. Moll studied under Henry McKean at the Courant Institute, graduated in 1984 with a thesis on the Stabilization of the standing wave in a caricature for nerve conduction. This so-called caricature had been proposed by McKean as a simpler model from the classical Nagumo and Hodgkin-Huxley models. After graduation, he spent two years as a Lawton instructor at Temple University. In 1986 he moved to Tulane University, New Orleans, where he is now a Professor of Mathematics.
He is interested in all aspects of the mathematics coming from the evaluation of integrals. The subject is full of interesting problems that he shares with colleagues, graduate and undergraduate students. Among the variety of results that have come out of this work, one should mention the theory of Landen transformations that are the rational version of the well-known transformations of Landen and Gauss for
elliptic integrals. His long term project is to provide proofs, automatic and human of all entries
in the classical table of Integrals by I. S. Gradshteyn and I. M. Ryzhik. Most of his work comes from exploring, via symbolic languages, unexpected relations among classical objects. Some of his work has been written in the book Numbers and Functions published in the Student Mathematical Library series from AMS. He is actively involved with bringing undergraduates into Mathematics. He has guided undergraduate research at Tulane University and also was the research leader at the REU programs SIMU (at the University of Puerto Rico at Humacao 2000 and 2002) and at MSRI-UP, Berkeley (2008 and 2014). A large number of his students have continued to graduate school in Mathematics.

Dynamics and Control of Inverted Ultra-Flexible Pendulum on Cart System

Séminaire le 29 Juin 2015, 16h00 à CentraleSupelec (Gif-sur-Yvette) Amphi Blondel
Prof. Prasanna S. Gandhi


Ultra-flexible beams are useful in several flexible mechanisms having applications, for example, in high-precision positioning systems and flexible link robotic systems. They demonstrate interesting dynamics, especially when these systems are in a vertical plane. This talk focuses on a representative of such system: a vertical, large-deformation cantilever with tip mass and actuated at fixed base by a cart. The system is shown to demonstrate multiple equilibria as a function of tip mass using elastica theory. Dynamics of this system is further captured using assumed modes method and imposing length constraint to model potential energy change of tip mass (constrained Langrange formulation). Evolution of chaotic behaviour under harmonic excitation is presented in comparison with experiments. We further develop a nonlinear nested saturations based controller for stabilising the cantilever in central unstable equilibrium with cart also reaching the desired position. Multiple coordinate transformations are performed to obtain a system in feedforward chain of integrators before applying nested saturation based control. Effectiveness of control is established through simulation cases and experimental implementation.

Biography: P.S. Gandhi received the B.Eng. degree in from the University of Bombay, Mumbai in 1994 and the M.Tech degree from the Indian Institute of Technology, Bombay, Mumbai in 1996, both in mechanical engineering. He received the Ph.D. degree in mechanical engineering from the Rice University, Houston in 2001. Since 2001, he has been faculty member, currently Professor, in the Department of Mechanical Engineering at Indian Institute of Technology, Bombay, Mumbai. His research interests are in the areas of MEMS and Microsystems, Mechatronics, and Nonlinear Dynamical Systems and Control. He has been a recipient of 2006 BOYSCAST fellowship of Govt of India, 2007 Best faculty award (ME department) and Prof J.R.Issac fellowship. He has authored several patents and over 85 peer reviewed conference and journal papers. He has coordinated setup of a new laboratory Suman Mashruwala Microengineering Laboratory for research in Microdomain and has successfully developed technologies of Bulk Lithography and Microstereolithography for 3D MEMS fabrication in this laboratory. He has been a qualified teacher of stress relieving, life enhancing techniques of Art of Living foundation by Sri Sri Ravishankar.

S³: Un modèle stochastique de la transcription d’un gène

Séminaire le 26 Juin 2015, 11h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Marc Roussel, University of Lethbridge, Alberta, Canada


Nous étudions depuis quelques années des modèles stochastiques de la transcription, c’est-à-dire de la synthèse de l’ARN à partir de la séquence de l’ADN par une machine moléculaire, l’ARN polymérase. Pour le cas d’une seule polymérase, il est possible de solutionner exactement nos modèles. Lorsque les interactions entre les polymérases sont importantes, il faut par contre utiliser (pour le moment) des méthodes numériques. En forme d’introduction au sujet, je présenterai un de nos modèles les plus simples, et je démontrerai comment on peut obtenir tous les moments voulus de la distribution du temps de transcription, c’est-à-dire comment on peut solutionner ce modèle. Cette distribution pourra être utilisée dans des modèles d’expression génétique, où elle apparaitra comme distribution de retards de la production de l’ARN.

Bio: Marc R. Roussel is Professor at Alberta RNA Research and Training Institute, Department of Chemistry and Biochemistry, University of Lethbridge. More details can be found at his homepage http://people.uleth.ca/~roussel/

S³: High dimensional minimum risk portfolio optimization

Séminaire le 26 Juin 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Liusha Yang, Department of Electronic and Computer Engineering, Hong Kong University of Science and Technology


The performance of the global minimum risk portfolio (GMVP) relies on the accuracy of the estimated covariance matrix of the portfolio asset returns. For large portfolios, the number of available market returns is often of similar order to the number of assets, making the sample covariance matrix performs poorly. In this talk, we discuss two newly-developed GMVP optimization strategies under high dimensional analysis. The first approach is based on the shrinkage Tyler’s robust M-estimation with a risk-minimizing shrinkage parameter. It not only deals with the problem of sample insufficiency, but also the impulsiveness of financial data. The second approach is built upon a spiked covariance model, by assuming the population covariance matrix follows the spiked covariance model, in which several eigenvalues are significantly larger than all the others, which all equal one. The performances of our strategies will be demonstrated through synthetic and real data simulations.

Bio: Liusha Yang received the B.S. in Communication Engineering from the Beijing University of Posts and Telecommunications in 2012. Currently, she is a Ph.D. student in the Department of Electronic and Computer Engineering at the Hong Kong University of Science and Technology. Her research interests include random matrix theory and signal processing, with applications in financial engineering.

Wireless devices and services for distributed sensing, monitoring, and decision support

Séminaire le 25 Juin 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
F. Viani: "Research Associate" et membre du Centre de Recherche ELEDIA de l'université TRENTO


Low-power and compact wireless devices, like smart sensors, embedded systems, smartphones, tablets are more and more becoming everyday life tools, bringing advantages not limited to the mobile communications but also referred to improved context awareness. The potentialities of such wireless technologies are enriched by the integration of dedicated real-time processing techniques which enable not only the distributed sensing of heterogeneous parameters, but also the improved management, understanding, and forecasting of complex processes. The output of such analysis is also exploited to support operators in decision making. Representative application examples are in the field of smart cities and communities, where distributed wireless sensors and mobile devices are largely applied both in indoor (e.g., in smart buildings, smart museums, etc.) and outdoor (smart lighting, road security, fleet management, etc.) scenarios.

Short CV: Federico Viani received the B.S. and M.S. degrees in Telecommunication Engineering and  the PhD degree in Information and Communication Technology from the University of Trento, Italy, in 2004, 2007, and 2010, respectively. Since 2011, Dr. Viani is a Research Associate (Post-Doc) at the Department of Information Engineering and Computer Science (DISI) of the University of Trento, Italy, and a member of the ELEDIA Research Center.
Since 2007, Dr. Viani has been the co-advisor of 18 M.S/B.S Thesis. Since 2010 he has been the official teacher of the Bachelor degree course "Design Techniques for Wireless Communications", and since 2007 he has been a teaching assistant of Bachelor degree and Master degree courses in Telecommunication Engineering offered by the University of Trento, including "Electromagnetic Propagation", "Project Course on Wireless Technologies", "Antennas for Wireless Communications", "Biomedical Diagnostic Techniques", "Mobile Communications".
Dr. Viani is author/co-author of over 77 peer reviewed papers on international journals and conferences, including 28 contributions on peer-reviewed international journals, 49 in international conferences. Moreover, Dr. Viani has been cited 574 times and his H-Index is equal to 14 in the Scopus Database. He has been invited to submit papers to International Journals and to present contributions to Scientific Sessions in International Conferences. He has organized and/or chaired 3 Special Sessions in International Conferences. Since 2007, he has attended 7 national and international conferences, presenting as a speaker 15 contributions.
Since 2007, Dr. Viani has been a Participant in 17 Research Projects, funded by EU, Industries, and National Agencies.
The research activities of Dr. Viani are oriented to the development of methodological strategies and applications in the framework of Electromagnetic Fields (S.S.D. ING‐INF/02, S.C. 09/F1), with main emphasis on applied electromagnetics. He has been involved in activities concerning the design of multiband, wideband, and ultra-wideband antennas, the study and development of optimization techniques as well as learning-by-example methodologies for the solution of complex electromagnetic problems including inverse problems and active/passive wireless localization. He is also involved in the design and development of distributed and pervasive monitoring by means of wireless sensor networks (WSNs) and robot swarms, and in the application of decision support systems (DSS) to fleet management and emergency-related applications.
Dr. Viani is a Reviewer for international Journals, including IEEE Transactions on Antennas and Propagation, IEEE Antennas and Wireless Propagation Letters, Progress in Electromagnetic Research/Journal of Electromagnetic Waves and Applications, IEEE Transactions on Vehicular Technologies.
Dr. Viani is a Senior Member of the IEEE, member of the IEEE Antennas and Propagation Society, and of the European Microwave Association (EuMA).

S³: Stability of continuous-time quantum filters

Séminaire le 19 Juin 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Nina H. Amini, CNRS, Laboratory of Signals and Systems, France


In this talk, we study quantum filtering and its stability problem. Indeed, we show that the fidelity between the state of a continuously observed quantum system and the state of its associated quantum filter, is always a sub-martingale. The observed system could be governed by a continuous-time Stochastic Master Equation (SME), driven simultaneously by Wiener and Poisson processes which takes into account incompleteness and errors in measurements. This stability result is the continuous-time counterpart of a similar stability result already established for discrete-time quantum systems. This result implies the stability of such filtering process but does not necessarily ensure the asymptotic convergence of such quantum filters.

Bio: Nina H. Amini is a CNRS researcher at Laboratory L2S at CentraleSupelec since October 2014. She did her first postdoc from June 2012 for six months at ANU, College of Engineering and Computer Science and her second postdoc at Edward L. Ginzton Laboratory, Stanford University since December 2012. She received her Ph.D. in Mathematics and Control Engineering from Mines-ParisTech (Ecole des Mines de Paris), in September 2012. Prior to her Ph.D., she earned a Master in Financial Mathematics and Statistics at ENSAE and the Engineering Diploma of l’Ecole Polytechnique, in 2009. Her research interests include stochastic control, quantum control, (quantum) filtering theory, (quantum) probability, and (quantum) information theory.

S³: Modeling and mismodeling in radar applications: parameter estimation and bounds

Séminaire le 9 Juin 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Maria S. Greco, Department of Information Engineering, University of Pisa


The problem of estimating a deterministic parameter vector of acquired data is ubiquitous in signal processing applications. A fundamental assumption underlying most estimation problems is that the true data model and the model assumed to derive an estimation algorithm are the same, that is, the model is correctly specified.
This lecture will focus on the general case in which, for some non-perfect knowledge of the true data model or for operative constraints on the estimation algorithm there is a mismatch between assumed and true data model.
After a short first part dedicated to explain the radar framework of the estimation problem, the lecture will be dedicated to the evaluation of lower bounds on the Mean Square Error of the estimate of a deterministic parameter vector under misspecified model with particular attention to Mismatched Maximum Likelihood estimator and Huber bounds.

Bio: Maria S. Greco graduated in Electronic Engineering in 1993 and received the Ph.D. degree in Telecommunication Engineering in 1998, from University of Pisa, Italy. From December 1997 to May 1998 she joined the Georgia Tech Research Institute, Atlanta, USA as a visiting research scholar where she carried on research activity in the field of radar detection in non-Gaussian background.
    In 1993 she joined the Department of Information Engineering of the University of Pisa, where she is Associate Professor since December 2011. She’s IEEE fellow since January 2011 and she was co-recipient of the 2001 IEEE Aerospace and Electronic Systems Society’s Barry Carlton Award for Best Paper and recipient of the 2008 Fred Nathanson Young Engineer of the Year award for contributions to signal processing, estimation, and detection theory. She has been co-general-chair of the 2007 International Waveform Diversity and Design Conference (WDD07), Pisa, Italy, in the Technical Committee of the 2006 EURASIP Signal and Image Processing Conference (EUSIPCO), Florence, Italy, in the Technical Committee of the 2008 IEEE Radar Conference, Rome, Italy, in the Organizing Committee of CAMSAP09, Technical co-chair of CIP2010 (Elba Island, Italy), General co-Chair of CAMSAP2011 (San Juan, Puerto Rico), Publication Chair of ICASSP2014, Florence, Italy, Technical Co-Chair of the CoSeRa2015, Pisa, Italy and Special Session Chair of CAMSAP2015, Cancun, Mexico. She is lead guest editor of the special issue on "Advanced Signal Processing for Radar Applications" to appear on the IEEE Journal on Special Topics of Signal Processing, December 2015, she was guest co-editor of the special issue of the Journal of the IEEE Signal Processing Society on Special Topics in Signal Processing on "Adaptive Waveform Design for Agile Sensing and Communication," published in June 2007 and lead guest editor of the special issue of International Journal of Navigation and Observation on” Modelling and Processing of Radar Signals for Earth Observation published in August 2008. She’s Associate Editor of IET Proceedings – Sonar, Radar and Navigation, Associate Editor-in-Chief of the IEEE Aerospace and Electronic Systems Magazine, member of the Editorial Board of the Springer Journal of Advances in Signal Processing (JASP), Senior Editorial board member of IEEE Journal on Selected Topics of Signal Processing (J-STSP), member of the IEEE Signal Array Processing (SAM) Technical Committees. She's also member of the IEEE AES and IEEE SP Board of Governors and Chair of the IEEE AESS Radar Panel. She's as well SP Distinguished Lecturer for the years 2014-2015, AESS Distinguished Lecturer for the years 2015-2016 and member of the IEEE Fellow Committee.
     Maria is a coauthor of the tutorials entitled “Radar Clutter Modeling”, presented at the International Radar Conference (May 2005, Arlington, USA), “Sea and Ground Radar Clutter Modeling” presented at 2008 IEEE Radar Conference (May 2008, Rome, Italy) and at 2012 IEEE Radar Conference (May 2012, Atlanta, USA), coauthor of the tutorial "RF and digital components for highly-integrated low-power radar" presented at the same conference, of the tutorial "Recent Advances in Adaptive Radar Detection" presented at the 2014 International Radar Conference (October 2014, Lille, France) and co-author of the tutorial "High Resolution Sea and Land Clutter Modeling and analysis", presented at the 2015 IEEE International Radar Conference (May 2015, Washington DC, USA).
    Her general interests are in the areas of statistical signal processing, estimation and detection theory. In particular, her research interests include clutter models, spectral analysis, coherent and incoherent detection in non-Gaussian clutter, CFAR techniques, radar waveform diversity and bistatic/mustistatic active and passive radars. She co-authored many book chapters and more than 150 journal and conference papers.

Novel Microstereolithography (MSL) and Bulk Lithography (BL) technologies for polymer/ceramic 3D MEMS

Séminaire le 3 Juin 2015, 16h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Prof. Prasanna S. Gandhi


Polymers are increasingly being used for fabrication of MEMS for their advantages of low cost, easy disposability, high flexibility, and so on. With the area of printable organic electronics developing fast there are enormous possibilities with polymer MEMS, especially 3D. This talk will present two indigenously developed technologies for fabrication of polymer 3D MEMS.
The proposed MSL technology has demonstrated resolution of 6micron and a large overall size of fabricated component at the same time: a feature seldom found in other such systems in the literature. The core opto-mechanical scanner (patent pending) and mechatronic system built around it would be presented. This system is demonstrated to have positioning accuracy within 100 nm. Several cases of micro-component fabrication will be demonstrated. Main drawback of MSL is stair-stepping effects on slanted walls. To overcome this limitation, recently "Bulk Lithography" technology has been proposed by our group. The main principle used here is to impose spatial variation of laser energy dose while scanning. The method gives remarkable accuracy in getting free-form surface features desired for microlenses, tapered micro cantilevers, tapered diaphragm micromirrors and so on, which are otherwise not feasible to fabricate. Finally, ceramic microstereolithography and preliminary fabrication of ceramic microcomponents would be demonstrated. The proposed technologies hold potential for direct prototyping and also mass production of MEMS devices along with packaging.

Biography: P.S. Gandhi received the B.Eng. degree in from the University of Bombay, Mumbai in 1994 and the M.Tech degree from the Indian Institute of Technology, Bombay, Mumbai in 1996, both in mechanical engineering. He received the Ph.D. degree in mechanical engineering from the Rice University, Houston in 2001. Since 2001, he has been faculty member, currently Professor, in the Department of Mechanical Engineering at Indian Institute of Technology, Bombay, Mumbai. His research interests are in the areas of MEMS and Microsystems, Mechatronics, and Nonlinear Dynamical Systems and Control. He has been a recipient of 2006 BOYSCAST fellowship of Govt of India, 2007 Best faculty award (ME department) and Prof J.R.Issac fellowship. He has authored several patents and over 85 peer reviewed conference and journal papers. He has coordinated setup of a new laboratory Suman Mashruwala Microengineering Laboratory for research in Microdomain and has successfully developed technologies of Bulk Lithography and Microstereolithography for 3D MEMS fabrication in this laboratory. He has been a qualified teacher of stress relieving, life enhancing techniques of Art of Living foundation by Sri Sri Ravishankar

S³: The appliction of medium grazing angle sea-clutter models -- The NRL multi-aperture SAR: system description and recent results

Séminaire le 26 Mai 2015, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Dr. Luke Rosenberg - DSTO, Australia


Details are given in the attached file. Access information are available on the website http://www.lss.supelec.fr/scube/

Seminaire_s3-sondra-icode May 26th

An introduction to neural fields

Séminaire le 12 Mai 2015, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Georgios Detorakis, post-doc L2S


The main objective of the field of Computational Neuroscience is to understand how the brain works through mathematical models (numerical and analytical works). Some of these models are described by neural fields equations. Neural fields are integro-differential equations that describe the spatiotemporal dynamics of the activity of a piece of neocortical tissue. Neural fields equations emerge when one assimilates the high number of neurons of the selected piece of brain tissue in its continuum limit. Neural fields have been studied both analytically and numerically by many researchers. They can model many different and interesting biological phenomena such as attention, working memory, self-organization, or synaptic depression. The seminar consists of two parts. The first part is a brief introduction to neuroscience and the second part is dedicated to neural fields. We will review how neural fields equations can be derived, how the steady-state solution can be computed, and its stability can be insured. Finally, some cognitive models that are based on neural fields will be presented.

Bio:
Georgios Detorakis has studied Applied Mathematics and Neuroscience. He did his PhD on cortical plasticity, self-organization and neural fields. During his PhD, he studied the formation of topographic maps in area 3b of the primary somatosensory cortex and the multimodal problem of “Touch and the body”. He is now a postdoc fellow at L2S working with Antoine Chaillet on Parkinson’s disease in the ANR project "SynchNeuro". He uses delayed neural fields in order to model some brain areas that play a crucial role in Parkinson’s disease motor symptoms.

Stability theory, as told by its developers

Séminaire le 5 Mai 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Antonio Loria, Elena Panteley (L2S)


We present a brief but detailed historical review on the development of stability theory, from its early starts out of the minds of Lagrangia and Dirichlet. The survey carries on to focus on Lyapunov stability. Through exact citations from the works of the developers of stability, including many Soviet texts from the 20th century, we revise the fundamental definitions and theorems; making emphasis on several misfortunate translations which have led to wrong interpretations and ambiguous statements. We favour depth and sacrifice generality: on technical grounds, we focus on the most elementary (yet not so) well-known forms of Lyapunov stability and common but crucial qualifiers that go with it: uniform, asymptotic, global. We revise the origin of the wrongly known invariance principle ...

S³ Working memory in random neural networks

Séminaire le 3 Avril 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Gilles Wainrib, ENS, Computer Science Department


Numerous experimental studies investigate how neural representations of a signal depend on its past context. Although synaptic plasticity and adaptation may play a crucial role to shape this dependence, we study here the hypothesis that this dependence upon past context may be also explained by dynamical network effects, in particular due to the recurrent nature of neural networks connectivity.

Short Bio: Gilles Wainrib is assistant professor in the Computer Science Department at Ecole Normale Supérieure and his research interests range from theoretical biology to applied mathematics and artificial intelligence.

Talk 1: Level set methods for seismic full waveform inversion. Talk 2: Some inverse problems for cargo container screening.

Séminaire le 30 Mars 2015, 13h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Oliver Dorn, Foreign Guest DIGITEO


Bio: Oliver Dorn is currently Lecturer at the School of Mathematics at The University of Manchester. He has obtained his PhD in Applied Mathematics from The University in Muenster, Germany, in 1997, followed by various postdoctoral research stays in the US and Canada. From 2002 until 2007 he was awarded a Ramon y Cajal fellowship at Universidad Carlos III de Madrid, where he became full professor (Profesor Titular) in 2008. He has visited Supélec and Université Paris Sud frequently for longer periods, and has published more than 50 papers in internationally competitive journals and conference proceedings. He is a professional member of SIAM (Society for Industrial and Applied Mathematics), IEEE (Institute of Electrical and Electronics Engineers) and EAGE (European Association of Geoscientists and Engineers) and is on the advisory board of the Journal 'Inverse Problems'. Right now he is at L2S as Foreign Guest DIGITEO.

S³ - Inverse problems in signal and image processing and S³ - Bayesian inference framework: from basic to advanced Bayesian computation

Séminaire le 27 Mars 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Ali Mohammad Djafari, CNRS, L2S


In signal and image processing community, we can distinguish two categories:
- Those who start from the observed signals and images and do classical processing: filtering for denoising, change detection, contour detection, segmentation, compression, …
- The second category called “model based”, before doing any processing try first to understand from where those signals and images come from and why they are here . So, first defining what quantity has been at the origin of those observations, then modeling their link by “forward modeling” and finally doing inversion. This approach is often called “Inverse problem approach”. Then, noting the “ill-posedness” of the inverse problems, many “Regularization methods” have been proposed and applied successfully. However, deterministic regularization has a few limitations and recently the Bayesian inference approach has become the main approach for proposing unsupervised methods and effective solutions in many real applications. Interestingly, even many classical methods have found better understanding when re-stated as inverse problem. The Bayesian approach with simple prior models such as Gaussian, Generalized Gaussian, Sparsity
enforcing priors or more sophisticated Hierarchical models such as Mixture models, Gaussian Scale Mixture or Gauss-Markov-Potts models have been proposed in different applications of imaging systems with great success. However, Bayesian computation still is too costly and need more practical algorithms than MCMC. Variational Bayesian Approximation (VBA) methods have recently became a standard for computing the posterior means in unsupervized methods.
Interestingly, we show that VBA includes Joint Maximum A Posteriori (JMAP) and Expectation-Maximization (EM) as special cases. VBA is much faster than MCMC methods, but, it gives only access to the posterior means.
This talk gives an overview of these methods with examples in Deconvolution (simple or blind, signal or image) and  in Computed Tomography (CT).

Bio: http://djafari.free.fr/index.htm

----------------------------------------------------------------------
Séminaire S³
http://scube.lss.supelec.fr
seminaire.scube@l2s.centralesupelec.fr
----------------------------------------------------------------------

PDE-­‐based inversion method with no forward solver for inverse medium scattering problems

Séminaire le 20 Mars 2015, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Dr. Yu Zhong


Bio: Dr. Yu Zhong received his B.S. and M.S. degrees in electronic engineering from Zhejiang University, Hangzhou, China, in 2003 and 2006, respectively, and the Ph.D. degree from the National University of Singapore, Singapore, in 2010.
He is currently a Scientist in Institution of High Performance Computing (IHPC), A*STAR, Singapore.
His research interests mainly are inverse-­‐scattering problems and electromagnetic modeling on composite materials. He is a regular visitor at the Laboratoire des Signaux et Systèmes (L2S) in Gif-­sur-­‐Yvette, France as an invited senior scientific expert since 2012.
 
Talk 1: PDE-­‐based inversion method with no forward solver for inverse medium scattering problems
A new partial differential equation (PDE) based inversion method for inverse medium scattering problems is proposed in this talk, which does not need to solve any forward problem. The proposed method is the subspace-­‐based optimization method (SOM) in the differential-­‐equation frame. The finite difference scheme is used to discretized the Helmholtz equation, and the twofold subspace-­‐based regularization scheme, as in the integral equation based SOM, is applied in this PDE-­‐based inversion method to stabilize the solver. By using such a PDE-­‐based inversion method, the Green’s funciton for the domain of interests is no longer needed. Representative numerical tests are presented to verify the efficacy of the proposed method.

Talk 2: New integral equation and new partial differential equation for inverse medium scattering problems with strong scatterers 

In this talk, we propose two new equations, an integral equation (IE) and a partial differential equation (PDE), for solving inverse medium scattering problems (IMSP) with strong scatterers. First, we present a new integral equation, which could effectively reduce the globle wave contribution in estimating the contrast (the difference between permittivities of the scatterers and the known background) compared to the original Lippmann-­‐Schwinger equation. Using such a new IE in the IE-­‐based inversion method one is able to solve the highly nonlinear IMSP with strong scatterers.
Subsequently, the connection between the PDE-­‐based inversion method (in Talk 1), using the Helmholtz equation, and the conventional IE based inversion method, using the Lippmann-­‐Schwinger equation, is discussed. With such a connection and the new IE, we propose a new PDE,
using which the PDE-­‐based inversion method can also solve the highly nonlinear IMSP. At last, we discuss the pros and cons of both PDE-­‐ and IE-­‐based inversion methods.

Pages