2pm in A301

Abstract

The encryption algorithms most commonly used today (RSA, AES ...) have
been designed to be the most resistant possible against attacks such as differen-
tial analysis. In fact, an algorithm like AES is still reputed today as extremely
resistant since the only known attack to date is only five times more effective
than an exhaustive search that would take billions of years with the best proces-
sors . However, there is another way to attack a device. It is possible to exploit
temporal or electrical leaks to deduce key pieces. These attacks are known as
side-channel attacks.
The aim of our work is to make the link between side-channel and the theory
of information invented by Claude Shannon, and to study whether the usual
encryption algorithms are strong enough to deal with such attacks. We show
that the encrypting algorithms are not as appropriate as would be desired in
the face of such a threat.

 Jeudi 1er juin, 14 H, Télécom ParisTech, Amphi B312, 46 rue Barrault, Paris 13

La première partie de ce séminaire sera consacrée à un retour d’expérience sur le projet «Patent Factory MIMO», un projet en collaboration avec France Brevet. Ce projet a pour objectif l’élaboration d’un portefeuille brevet autour des décodeurs MIMO à faible complexité. C'est un sujet d’actualité étant donnée l’augmentation considérable des dimensions des systèmes d’aujourd’hui tels que le nombre d’antennes dans les systèmes sans fils, le nombre de modes et  de cœurs utilisés dans une fibre otique, et le nombre de nœuds ou capteurs dans un réseaux.

La seconde partie de ce séminaire sera dédiée à un travail réalisé dans le cadre de ce projet. Il s’agit d’un nouveau décodeur récursif, appelé « Semi-Exhaustive block MIMO decoder ». L’idée étant de diviser le système à décoder en blocs afin de réduire la dimension et par conséquent réduire de manière significative la complexité totale de décodage. A travers la dérivation de la probabilité d’erreurs, nous fixons les paramètres de notre décodeur afin d’atteindre la diversité désirée. 

11am in A301

Abstract:

One fundamental problem in decentralized networked systems is to coordinate activities of different nodes so that they reach a state of agreement. This global objective is typically obtained by local operations, for example, by employing gossip algorithms to achieve consensus over a set of agents, where several data exchanges are iteratively carried out between pairs of adjacent nodes. In these works the consensus is often given as a global function of all local observations, as for example an average of the same random process over all observing nodes. In contrast, we are interested in a generalization of this problem where consensus is meant in a broader sense of achieving coordinated actions by the network nodes, and therefore can be seen as an instance of distributed control in networks. This cooperative behavior is useful in a host of applications, for example in multi-agent systems for exploration of an unknown terrain, distributed surveillance applications, automatic vehicle control applications, or load balancing with divisible tasks in a large computer networks or power grids.

In this talk we address the coordination of multiagent systems over  point-to-point channels and line networks. We first investigate the problem of strong coordination over point-to-point noisy communication channels. Note that strong coordination requires that the L1 distance between the induced joint distribution of action sequences at the two nodes (A and B) by a coordination code and a prescribed joint distribution specified by the system designer must vanish exponentially fast with the block length. To this end, we first propose a joint coordination-channel coding scheme that implicitly uses channel randomness to reduce the randomness required in generating the action sequence at node B. We compare this scheme with a separate coordination-channel coding scheme which extracts channel randomness after the channel decoding stage and observe that the joint scheme is able to provide a lower communication rate compared to the separate scheme under the same amount of total injected randomness into the system. After that inner and outer bounds for the coordination capacity region for the coordination of agents along a line and in a broadcast setting are constructed. We show that for a given coordination demand the choice of the communication topology has a direct effect on the achievable rate. Finally, for point-to-point coordination a low complexity construction is presented based on polar codes which achieves a subset of the (strong) coordination capacity region.

Bio:
Dr. Joerg Kliewer received the Dipl.-Ing. (MSEE) degree in Electrical Engineering from the Hamburg University of Technology, Hamburg, Germany, in 1993 and the Dr.-Ing. degree (Ph.D.) in Electrical Engineering from the University of Kiel, Kiel,  Germany, in 1999, respectively. From 1993 to 1998 he was a Research Assistant at the University of Kiel, Germany, and from 1999 to 2004, he was a Senior Researcher and Lecturer with the same institution. In 2004, he visited the University of Southampton, Southampton, U.K., for one year, and from 2005 until 2007, he was with the University of Notre Dame, Notre Dame, IN, as a Visiting Assistant Professor. From August 2007 until December 2013 he was with New Mexico State University, Las Cruces, NM, as an as an Assistant and most recently as an Associate Professor. In January 2014 he joined the New Jersey Institute of Technology, Newark, NJ, as an Associate Professor.  His research interests span coding and information theory, graphical models, and statistical algorithms, which includes application to networked communication and security, data storage, and biology. Dr. Kliewer was the recipient of a Leverhulme Trust Award and a German Research Foundation Fellowship Award in 2003 and 2004, respectively. He was Associate Editor of the IEEE Transactions on Communications from 2008-2014 and now serves as an area editor for the same journal since 2015. He is also member of the editorial board of the IEEE Information Theory Society Newsletter since 2010 and serves as chair of the outreach committee for the same society since 2012. 

Filterless optical networks based on coherent transceivers and passive broadcast-and-select nodes can be considered as a cost-effective and simpler alternative to active optical switching networks based on Reconfigurable Optical Add/Drop Multiplexers (ROADMs). Filterless networks are increasingly seen as very serious candidate architectures for future networks because of their main attributes (cost-effectiveness, energy efficiency), as well as its inherent gridless, elastic, multicast and cognition capabilities. Furthermore, coherent receivers equipped with digital signal processing (DSP) are foundational technologies for flexible and cognitive optical networking owing to their dynamic impairment compensation, performance monitoring and reconfiguration capabilities.

In this talk, we will review the recent progress on the design of elastic and agile filterless optical networks for terrestrial and submarine applications and present an overview of the current research activities at the Network Technology Lab aiming at leveraging the performance monitoring capability of coherent receivers for characterizing the performance dynamics in coherent optical networks and developing performance predictive models.

Biography

Christine Tremblay is a Professor in the Electrical Engineering Department of the École de technologie supérieure (ÉTS) since 2004. She is the founding researcher and head of the Network Technology Lab, an optical layer test bed composed of 19 network nodes fully equipped with 10G-100G transmission systems for research and teaching in optical communications. Her research interests include cognitive optical networks, optical performance monitoring, as well as optical layer characterization and silicon photonics. Her team introduced the filterless network concept, a step towards more flexible cost-effective and energy-efficient WDM network architectures for core and submarine network applications. Before joining ÉTS, she held senior R&D and technology management positions at Nortel, EXFO, Roctest and the National Optics Institute (INO). She has been co-instructor for two OFC hands-on short courses on fiber characterization and polarization measurements from 2009 to 2015. She is a member of two Quebec FRQNT Strategic Clusters (Center for Advanced Systems and Communications, SYTACom; Center for Optics, Photonics and Lasers (COPL), the Optical Society of America (OSA) and the IEEE Photonics Society.

To meet the computational requirements of future 5G networks, the signal-processing functions 

of baseband stations and user equipments will be accelerated onto programmable, configurable 

and hard-wired components (e.g., CPUs, FPGAs, hardware accelerators). This urges the need to 

generate efficient implementations for such mixed architectures. Existing model-based approaches

 generate executable implementations of Systems-on-Chip (SoCs) by translating models into 

multiple SoC-programming languages (e.g., C/C++, OpenCL, Verilog/VHDL). However, these 

translations do not typically consider the optimization of non-functional properties 

(e.g., memory footprint, scheduling). This paper proposes a novel approach where 

system-level models are optimized and compiled into multiple implementations for 

different SoC architectures. We show the effectiveness of our approach with the 

compilation of UML/SysML models of a 5G decoder. Our solution generates, from a single 

Platform-Specific model, both a software implementation for a Digital Signal Processor 

platform and a hardware-software implementation for a platform based on hardware 

Intellectual Property (IP) blocks. Overall, we achieve a memory footprint reduction 

of 62% in the first case and 30% in the second case.

14 H, Télécom ParisTech, Amphi B312, 46 rue Barrault, Paris 13

Physical Layer Security (PHYSEC) provides means to protect the radio layer of wireless networks without pre-distributed keys, by exploiting the source of randomness built by the propagation and the reception noise of radio-communication signals. The Phylaws project (ICT-FP7 2013-16) proved that this physical source offers good randomness quality, while remaining exclusively shared among legitimate communication nodes and terminals. During this project, three PHYSEC schemes were studied, experimentally demonstrated on Wifi 802.11n/ac à 2.4 et 5 GHz, patented and published:

-    Secure Pairing (SP) of nodes and terminals, which Technology Readiness Level estimated for Wifi is 3,
-    Secret Key Generation SKG, which Technology Readiness Level estimated for Wifi is 5,
-    Secrecy Coding SC, which Technology Readiness Level estimated for Wifi is 5.

The specific interest of PHYSEC processing relies in their limited impact at the upper protocol layers of the OSI model: operating at the radio interface only, they are quite easy to implant. Thus, SKG and SC processing demonstrated in the Phylaws project for transec(*) and netsec(**) protection of 4G and WLAN nodes and terminals have been proposed for 5G standardization: they achieve a better protection of early negotiation and access protocol, enhancement of authentication, better security of messages relevant to subscriber’s identity, location, ciphering capability, etc. Nevertheless, when needed and relevant, the secret extracted at the radio interface can be propagated and exploited into higher protocol layers for further comsec(***) protections.

After a short introduction to security lacks of wireless networks and principle+theoretical performances of Physical Layer Security, this synthetic talk exposes the main processing developed and demonstrated in the Phylaws project. We also illustrate some practical application cases and performances at Wifi links.

(*) TRANSEC (Transmission Security): protection of the radio-communication signal.
(**) NETSEC (Network Security): protection of signaling messages
(***) COMSEC (Communication Security): protection of data messages.

in A301 at 1.30pm

Abstract:

The energy and sampling constrained capacity was recently derived for bursty communication when the information is available infrequently at random times at the transmitter. Since the result depends on the asymptotical analysis, it's natural to doubt whether we can get similar conclusions in the finite length regime. In such regime, dispersion plays a key role in assessing the backoff from capacity due to finite length. We investigate the dispersion under different asynchronism settings, and establishes non-asymptotic tradeoffs between detection delay, output sampling rate, and communication rate for bursty communication. These tradeoffs imply regimes where the gap to capacity is captured by the inverse of the sampling rate rather than the usual dispersion.

in A301 at 11am

Abstract:

Inspired by a problem of list decoding with structural constraints, we introduce a new notion of similarity between graphs, termed graph information ratio. We discuss various properties of this measure, including in particular metric structure and partial ordering of graphs, an information ratio power inequality, relations to graph homomorphism, algebraic identities and inequalities, and more. Joint work with Lele Wang (Stanford).

"Ballooning technique enhanced with memory hotplug for QEMU/KVM guests"

During the seminarI will present current developments done within my PhD thesis, that is an improved memoryvirtualization in Linux/KVM, which enables dynamic regulation of diaggregated RAM resources - without guest VM restart.At the beginning the disaggregation subject will be shortlydescribed in the context of the EC H2020 DREDBOX project and then I will focus more on myinvolvement: memory ballooning and memory hotplug. Also, at the end I will probably presentshort demo.

Jeudi 30 mars, 14H, Amphi B312, Télécom parisTech, 46 rue Barrault, paris 13

Mobile cellular networks are becoming increasingly complex to manage while classical deployment/optimization techniques  are cost-ineffective and thus seen as stopgaps. This is all the more difficult considering the extreme constraints of 5G networks  in terms of data rate (more than 10 Gb/s), massive connectivity (more than 1000000 devices per km2), latency (under 1ms) and energy efficiency (a reduction by a factor of 100 with respect to 4G network). Unfortunately, the development of adequate solutions is severely limited by the scarcity of the actual ressrouces (energy, bandwidth and space). Recently, the community has turned to a new ressource known as Artificial  Intelligence at all layers of the network to exploit the increasing computing power afforded by the improvement in Moore's law in combination with the availability of huge data in 5G networks. This is an important paradigm shift which considers the increasing data flood/huge number of nodes as an opportunity rather than a curse. In this talk, we will discuss through various examples how the recent advances in big data algorithms can provide an efficient framework for the design of 5G Intelligent Networks.

1.30pm in A301

Abstract :

Wireless communication systems use retransmission protocols to achieve more reliable communication. The source retransmits the packets that were not successfully decoded at the destination. Another option is to use a relay for packet retransmissions. We introduce "Cooperative HARQ-II" and compare it with existing relay assisted HARQ protocols. The comparison includes performance evaluation by simulation and theoretical analysis.

Jeudi 2 mars, 14 H, Télécom ParisTech, Amphi B312, 46 rue Barrault, Paris 13

RF power transmitters with high power efficiency and linearity: architectures and linearization by digital predistortion

Power Amplifiers (PA) are critical elements of radiocommunication systems because their power efficiency conditions the autonomy and cost of equipments and their linearity influences on performance of the communication. With the emergence of 5G systems, achieving a good trade-off between linearity and efficiency is becoming more crucial. New post-OFDM waveforms are proposed for 5G networks in order to improve spectral occupancy, decrease latency and facilitate asynchronous communications. However, these waveforms exhibit very high crest factors and are very sensitive to the PA non-linearities. Morever, carrier aggregation allows for very high data rate but creates new challenges for transmitters in terms of bandwidth, intermodulation, harmonics and filtering.

In that context, the presentation will be dedicated to architectures and linearization techniques for high efficiency linear transmitters.
After a short discussion on the impact of the evolution of communication systems and standards on transmitter architectures, it will give an overview of the architectures of power transmitters (Doherty, Envelop tracking, sampled architectures ...) in the context of 3G/4G/5G systems. Then it will focuss on linearization by digital predistortion: basic principles and challenges.

at 1.30pm in A301

Abstract

This talk will be in two parts. First, I will introduce two optimization algorithms, the gradient algorithm and the proximal point algorithm from an unified point of view. Then, I will derive their stochastic version.

Jeudi 2 février 2017, 14H, Télécom ParisTech (salle C49)

Hardware security in mobile and embedded systems is drawing much attention in the context of the rapid growth of Internet-of-Things. Due to the higher accessibility, security threats and vulnerabilities for "things" located everywhere are much more critical in comparison with PCs and servers in a room.  In particular, the threats of side-channel attacks are non-trivial because they can be done by relatively low-cost equipment in a non-destructive manner.  In the last few decades, a variety of side-channel attacks have been introduced and defeated, but they are still being one of the hottest topics in the field of hardware security research.  This talk will start with an overview of hardware security research, and then introduce some state-of-the-art studies including on-going ones collaborated with Telecom ParisTech.

Jeudi 12 janvier, 14 H, Télécom ParisTech, Amphi Grenat, 46 rue Barrault, Paris 13

Les méthodes formelles offrent des outils de preuve "sûrs" pour assurer la fiabilité des systèmes, informatiques en particulier. Pour gérer la complexité, des techniques de structuration ont été développées telles que l'abstraction, le raffinement et la décomposition. Après une présentation rapide des principes généraux qui sous-tendent ces approches, le séminaire en présentera une mise en œuvre sur un exemple. La méthode formelle utilisée sera le B événementiel qui offre un processus de raffinement complet et entièrement prouvé. Nous montrerons comment nous avons utilisé cette technique pour établir une propriété de sécurité dans le cadre du projet SecBus. Ce projet propose une solution pour contrer les attaques sur le bus "CPU-Mémoire". Cette solution consiste à exécuter des algorithmes (cryptage, génération d'information redondante) lors de chaque accès à la mémoire par ce bus. Nous présenterons comment on peut garantir cette solution en B événementiel, c'est-à-dire comment la propriété attendue a été modélisée et comment elle a été prouvée. Nous conclurons par une discussion sur ce que cette expérience nous a appris sur les enjeux, difficultés et intérêts d'une telle approche.

11am in A301

Abstract :

Fano coding is a simple data compression scheme stemming from the dawn of information theory. While not optimal, its performance is good in spite of its simplicity. Interestingly, no rigorous performance analysis has been done. We correct this omission and provide an upper bound on the average codeword length of binary and ternary Fano codes for an arbitrary discrete memoryless source. The performance bound is slightly better than Shannon’s well-known bound for Shannon coding.As a by-product a novel general lower bound on Shannon entropy is derived that might be of interest also in a different context. This bound is expressed with the help of variational distance and provides a special case of a reverse Pinsker inequality.

La vague croissante d'attaques visant à communiquer les systèmes embarqués a exposé leurs 

utilisateurs aux risques de vol d'informations, de dommages monétaires et de blessures. 

En améliorant la modélisation et l'analyse de la sécurité, ces lacunes pourraient être 

atténuées. Une des premières phases de la conception des systèmes embarqués est le 

partitionnement HW / SW. Étant donné que le partitionnement HW / SW a une incidence sur 

l'intégration future de la sécurité dans le système, cette phase serait avantageuse en 

soutenant la modélisation des abstractions de sécurité et des propriétés de sécurité, 

fournissant aux concepteurs une rétroaction de partitionnement utile obtenue à partir 

d'un analyseur formel de sécurité. 

TTool a été étendu pour prendre en charge la modélisation de sécurité, l'intégration de 

sécurité automatisée et l'analyse formelle pendant la phase de partitionnement HW / SW 

pour concevoir des systèmes embarqués sécurisés. Nous avons introduit "Cryptographic 

Configurations", une représentation abstraite de la sécurité qui nous permet de vérifier 

formellement la sécurité avec ProVerif. Notre boîte à outils aide également les concepteurs 

à modéliser la sécurité en ajoutant automatiquement ces représentations de sécurité basées 

sur une cartographie et des exigences de sécurité.

Pour comparer la fréquence de deux horloges atomiques installées à grande distance l’une de l’autre, les horlogers font appel à des réseaux de satellites du type GPS. Mais ces méthodes manquent de précision pour comparer les dernières générations d’horloges atomiques, qui fonctionnent avec des lasers pour sonder la transition atomique qui sert de référence de fréquence. On parle alors d’horloges optiques. En effet, la fréquence de ces horloges optiques peut être désormais évaluée avec dix-sept, voire dix-huit chiffres significatifs, tandis que les signaux GPS ne permettent d’obtenir que quinze à seize chiffres significatifs.
Aussi une recherche très active est menée en France et en Europe pour délivrer et comparer des références de fréquence optique via le réseau Internet de fibre optique. En travaillant dans le domaine optique, et grâce aux propriétés remarquables de la fibre optique, cette nouvelle technologie dépasse de plusieurs ordres de grandeurs les capacités offertes par le service GPS dans le domaine radio-fréquence. Une révolution est en marche, avec une finalité : ouvrir la voie à des applications inédites dans des domaines aussi variés que la métrologie, la physique ou la géodésie !

Le SYRTE, laboratoire en charge de la réalisation de la seconde en France, est aujourd’hui connecté par fibre optique à ses homologues anglais et allemands. Nous montrerons au cours de ce séminaire les résultats des premières comparaisons d’horloges optiques que nous avons pu réaliser et les perspectives d’évolutions des techniques et du réseau dans les prochaines années.

Séminaire exceptionnel : vendredi 25 novembre 2016, 10H, salle C47, Télécom paristech (46 rue Barrault, paris 13)

Lors de ce séminaire seront présentés les différents réseaux sans fils pour l’Internet des Objets tels que : LoRa, Sigfox, ZigBee, Z-Wave, etc … ainsi que leurs couches physiques, leurs portées et niveaux de puissance. Suivra dans une deuxième partie une présentation des différentes offres actuelles des opérateurs classiques et nouveaux tels que : Bouygues, Orange, SIGFOX, Proximus, Swisscom, Fastnet, etc ..., et une présentation des solutions hardwares pour IoT se déclinant sous forme de modules et de chipsets. Le séminaire se terminera par poser la problématique de la consommation énergétique des objets connectés particulièrement ceux alimentés par des batteries.

Bilal Manaï est un serial entrepreneur. Durant les 10 dernières années, il a créé et co-créé Yaslamen, Yedess, IntelliBatteries, Recytex et Ilef Assurances. Avant son expérience entrepreneurial, Bilal a obtenu en 2002 un Doctorat en électronique et communications de Telecom Paris Tech, il a travaillé pendant 3 ans en tant que ingénieur de recherche chez Wavecom une start-up française située dans la région parisienne et rachetée par Sierra Telecom le leader mondial des solutions M2M. Il a également travaillé pendant 6 ans en tant que ingénieur R&D, chef de projet et manager chez Atmel. Il occupe actuellement la position de Directeur Général de Yaslamen depuis 6 ans, une start-up française spécialisée dans la gestion intelligente de batteries, et la position de Directeur Technique de Yedess, un opérateur de Beacon Bluetooth Low Energy. Il a enseigné la microélectronique, la gestion d’énergie et de batteries en tant que intervenant extérieur respectivement à Telecom Paris Tech, l’Institut Supérieur d’Électronique de Paris, Polytech Montpellier, Polytech Marseille et en tant que Professeur assistant à l’Ecole des Mines de Saint Etienne. Actif dans les IoT depuis 2015, il a initié le ENISo IoT challenge et animé différents séminaires sur les IoT à Montréal, Ottawa et Paris. Bilal détient 7 brevets américains et français, a publié plusieurs articles scientifiques et siège au sein de plusieurs comités techniques de conférences internationales.

<article>

at 1.30pm in A301

Abstract :

Shannon’s information theory has had a tremendous impact on various scientific fields 

in the 1950s and 1960s, e.g. experimental psychology. In line with this is Fitts' law; 

a relationship resting on the premise that their exists a motor channel which limits 

human performance, and based on Shannon's capacity formula for the AWGN channel. 

More precisely, to reach a target of size W at a distance D, Fitts' law asserts 

that one's movement time (MT) should follow: MT = a + b log(1+D/W). Although 

providing excellent experimental predictions, the theoretical derivation of the 

law is unsatisfactory: the formula is but a gross analogy and does not build on 

an information theoretic analysis. In this talk we will show that we can build 

a simple model for the aiming task that is compliant with information theoretic 

principles. We will show that when the aiming task is modeled by an additive 

uniform noise channel with the appropriate input, we are able to rederive Fitts' law. 

We also provide a novel solution to integrate the fact that sometimes the target may 

be missed. We then propose a second model based on the capacity achieving scheme 

by Elias to transmit a single signal from a Gaussian source through mutliple uses 

of an AWGN channel with perfect feedback, and show that movement time still is 

expressed as a linear function of the aiming task's SNR. 

</article>

2pm in Amphi Grenat

Abstract

This thesis is about clustering of mobile ad hoc networks, which consists in building sets of nodes called clusters in order to introduce hierarchy in the network and thus improve its scalability. The main goal is to design new distributed clustering algorithms suited to i) unstructured networks, where all the nodes are equal, and suited to ii) structured networks that have an inherent hierarchical structure, and in which the nodes are gathered in operational groups. In order to allow the implementation of a radio resource allocation process that is more efficient within clusters than between clusters, the proposed algorithms form clusters satisfying specific topology constraints:  connectivity, maximum size and diameter. In the first part of the thesis, to compare these new solutions to the ones from the literature, independently to the medium access scheme, we introduce network cost functions which take into account the user traffic profile and the intra-cluster and inter-cluster communication costs. Then, we propose a distributed clustering algorithm suited to structured networks, and compare its performance to several clustering schemes from the literature. A salient feature of this algorithm is that it does not need to resort to the notion of cluster-head. In the last part, thanks to coalition game theory we revisit this algorithm. This theoretical framework allows us to formalize the clustering problem in a more general context. This leads us to defining a generic algorithm suitable to any kind of ad hoc network, and enables us to acquire a better knowledge of its properties.

Jeudi 3 novembre, 14 H, Télécom ParisTech, Amphi Saphir, 46 rue Barrault, Paris 13

Decades of research efforts made energy harvesting techniques grow from long-established concepts into devices for powering ubiquitously deployed consumer electronics. Our environment provides a plethora of virtually cost-free sources of energy, e.g., vibrational, photovoltaic, thermoelectric, wind, just to name a few. A crescent attention has been lately devoted to another potential source of energy to harvest, i.e., electromagnetic waves. Modern wireless networks indeed provide an abundance of possible sources of such energy. In this context, a wireless energy transfer (WET) can be performed by means of several technologically different techniques, depending on the nature of the propagation of the electromagnetic wave whose energy is harvested. This talk will start from a general introduction to WET and radio frequency (RF) energy harvesting, to then illustrate the benefits and impact that these approaches.

Disaster Management is a vital application in the field of radar remote sensing. The requirements and challenges of disaster management, particularly in the context of humanitarian missions, posed on the Radar remote sensing systems has now become an important and ever expanding area of remote sensing applications. To date, there is a deficit in the inter-linking between radar remote sensing Engineers and disaster management scientists. This paper will address this issue in the spirit of a tutorial talk. More specifically, the paper will include the following key aspects:
• Main definition and concepts of humanitarian interventions and missions
• Presentation of the typical architecture of a mission scenario
• Data acquisition with focus on the sensor part i.e. the RADAR Instrument in the context of obstacle avoiding and GPR (Ground Penetrating Radars)
• Communications
• Illustration via examples (focused mainly on a new Radar approach)
• Conclusions and perspectives

in Amphi Rubis at 11am

Abstract:

The problem of estimating an arbitrary random vector from its observation corrupted by additive white Gaussian noise, where the cost function is taken to be the Minimum Mean p-th Error (MMPE), is considered. The classical Minimum Mean Square Error (MMSE) is a special case of the MMPE. Several bounds, properties, and applications of the MMPE are derived and discussed. Properties of the MMPE as a function of the input distribution, Signal-to-Noise-Ratio (SNR) and order p are derived. In particular, it is shown that the MMPE is a continuous function of p and SNR. These results are
possible in view of interpolation and change of measure bounds on the MMPE. Some basic properties of the optimal MMPE estimator are pointed out.

The ‘Single-Crossing-Point Property’ (SCPP) that bounds the MMSE for all SNR values above a certain value, at which the MMSE is known, together with the I-MMSE relationship is a powerful tool in deriving converse proofs in
multi-user information theory. By studying the notion of conditional MMPE, a unifying proof (i.e., for any p) of the SCPP is presented. A complementary bound to the SCPP is then shown, which bounds the MMPE for all SNR values below a
certain value, at which the MMPE is known.

The notion of the MMPE is applied to derive bounds on the conditional differential entropy as well as to generalize the Ozarow-Wyner mutual information lower bound for a discrete input on AWGN channel. It is also used to bound the MMSE of finite length codes improving on previous characterizations of the phase transition phenomenon for capacity-achieving
codes of infinite length. A short outlook of future applications concludes the presentation. 
----------------------------------------------------------------------------
Joint work With A. Dytso, R. Bustin, D. Tuninetti, N. Devroye and H.V. Poor

10.30 am in A301

Abstract : We address the problem of power and bandwidth allocation for Hybrid Automatic Repeat reQuest (HARQ) in
Orthogonal Frequency Division Multiple Access clustered ad hoc networks with statistical Channel State Information (CSI)
and practical Modulation and Coding Schemes (MCS). The sum of the energy efficiency (EE) of the different users is
maximized with Quality of Service (QoS) requirement. Using a tight approximation of the paquet error rate, we propose an
algorithm to find the optimal resource allocation. To reduce the complexity, we also propose two sub optimal algorithms, one based on alternating optimization and the other on a high Signal to Noise Ratio approximation. It is shown that the proposed algorithms allow a substantial gain in term of EE compared with conventional algorithms.

Jeudi 13 octobre,Télécom ParisTech, Amphi Thévenin, 46 rue Barrault, Paris 13, 14 H

In parallel with the onset of the standardization of so-called 5th generation cellular communication systems, the telecommunication industry is undergoing a historical change in the way systems will be designed, implemented and deployed. This change is a consequence of the confluence of the more general information technology industries and traditional telecom operators and vendors. In addition, the rapid emergence of open-source models in the design of telecom and network infrastructure poses a formidable challenge to equipment vendors and standardization bodies. This challenge could potentially drastically change the industry value chain and the underlying competitive dynamics and fundamentally alter the way these traditional industries do business.

This seminar will discuss open-source technologies and their impact on emerging 5th generation telecommunication systems. In particular, we will expose some of the choices made by EURECOM concerning the evolution of the software-radio platform activity which led to the creation of the OpenAirInterface Software Alliance. This alliance aims to soften the blow for traditional standards-based telecommunication industries in the transition to open-source methodologies. Moreover it aims to ensure better collaboration with academia and public-sector research organizations in the area of telecommunication systems.

in A301 at 11am

Abstract A class of diamond networks is studied where the broadcast component is orthogonal and modeled by two independent bit-pipes. New upper and lower bounds on the capacity are derived. The proof technique for the upper bound generalizes bounding techniques of Ozarow for the Gaussian multiple description problem (1981) and Kang and Liu for the Gaussian diamond network (2011). The lower bound is based on Marton's coding technique and superposition coding. The bounds are evaluated for Gaussian and binary adder multiple access channels (MACs). For Gaussian MACs, both the lower and upper bounds strengthen the Kang-Liu bounds and establish capacity for interesting ranges of bit-pipe capacities. For binary adder MACs, the capacity is established for all ranges of bit-pipe capacities.

Bio  Shirin Saeedi Bidokhti is a postdoctoral researcher jointly at the Technical University of Munich, where she works with Gerhard Kramer, and at Stanford University where she works with Tsachy Weissman. She got her B.Sc (2005) in Electrical Engineering from University of Tehran, and her M.Sc (2007) and PhD (2012) both in Communication Systems from Ecole Polytechnique Fédérale de Lausanne (EPFL). Her PhD advisors were Suhas Diggavi and Christina Fragouli. Dr. Saeedi was awarded a Prospective Researcher Fellowship (2012) and an Advanced Postdoc Mobility Fellowship (2014) both from the Swiss National Science Foundation. Her research interests include information theory and applications, multi-user communication systems and network coding.

15H30, Amphi B310, Télécom ParisTech, 46 rue Barrault, Paris 13

Android malware unfortunately have little difficulty to sneak in marketplaces. While known malware and their variants are nowadays quite well detected by anti-virus scanners, new unknown malware, which are fundamentally different from others (e.g. ”0-day”), remain an issue. To discover such new malware, the SherlockDroid framework filters masses of applications and only keeps the most likely to be malicious for future inspection by anti-virus teams. Apart from crawling applications from marketplaces, SherlockDroid extracts code-level features, and then classifies unknown applications with Alligator. Alligator is a classification tool that efficiently and automatically combines several classification algorithms.

To demonstrate the efficiency of our approach, we have extracted properties and classified over 600,000 applications during two crawling campaigns in July 2014 and October 2014. This lead us to discover several unknown families of Android malware, for a total of around 20 malicious samples (in 2016: 7). Slides on the overall approach will be first presented. A demonstration of Android malware classification will follow.