Search

Profiling side-channel attacks in which an adversary creates a “profile” of a sensitive device and uses such profile to model a target device with similar implementation has received the lion’s share of attention in the recent years. In particular, template attacks are known to be the most powerful profiling side-channel attacks from an information theoretic point of view. When launching such attack, the adversary first builds a model based on the leakage of the profiling (training) device in his disposal which is then exploited in the second phase of the attack (i.e., matching) to extract the key from the target device. Discrepancies between the device used for modeling and the target device affect the attack success. The effect of process variation and temperature misalignment between the profiling and target devices in the template attack’s success has been studied extensively in literature, while the impact of device aging on the template attack’s success is yet to be investigated thoroughly. This paper moves one step forward and studies the impact of device aging, mainly Bias temperature Instability (BTI) and Hot Carrier Injection (HCI), in the devices that have been protected against power analysis attacks via dual rail logics. In particular, we focus on the Wave Dynamic Differential Logic (WDDL) circuits, and via extensive transistor-level simulations we will show how device aging misalignments between the profiling and target devices can hinder template attacks for both unprotected and WDDL protected counterparts. We mounted several attacks on the PRESENT cipher, with and without WDDL protection, at different temperatures and aging times. Our results shows that the attack is more difficult if there is an aging-duration mismatch between the training and target devices, and the attack-efficiency decrease is especially significant for mismatches of few weeks.

In social interactions between humans and Embodied Conversational Agents (ECAs) conversational interruptions may occur. ECAs should be prepared to detect, manage and react to such interruptions in order to keep the interaction smooth, natural and believable. In this paper, we examined nonverbal reactions exhibited by an interruptee during conversational interruptions and we propose a novel technique driven by an evolutionary algorithm to build a computational model for ECAs to manage user's interruptions. We propose a taxonomy of conversational interruptions adapted from social psychology, an annotation schema for semi-automatic detection of user's interruptions and a corpus-based observational analysis of human nonverbal reactions to interruptions. Then we present a methodology for building an ECA behavioral model including the design and realization of an interactive study driven by an evolutionary algorithm, where participants interactively built the most appropriate set of multimodal reactive behaviours for an ECA to display interpersonal attitudes (friendly/hostile) through nonverbal reactions to a conversational interruption.

International audience; Older adults often need some level of assistance in performing daily living activities. Even though these activities are common to the vast majority of individuals (e.g., eating, bathing, dressing), the way they are performed varies across individuals. Supporting older people in performing their everyday activities is a major avenue of research in smart homes. However, because of its early stage, this line of work has paid little attention on customizing assistive computing support with respect to the specific needs of each older adult towards improving its effectiveness and acceptability. We propose a tool-based approach to allowing caregivers to define services in the area of home daily living, leveraging their knowledge and expertise on the older adult they care for. This approach consists of two stages: 1) a wizard allows caregivers to define an assistive service, which supports aspects of a daily activity that are specific to an older adult; 2) the wizard-generated service is uploaded in an existing smart home platform and interpreted by a dedicated component, carrying out the caregiver-defined service. Our approach has been implemented. Our wizard has been successfully used to define existing manually-programmed, activity-supporting services. The resulting services have been deployed and executed by an existing assisted living platform deployed in the home of community-dwelling individuals. They have been shown to be equivalent to their manually-programmed counterparts. We also conducted an ergonomics study involving five occupational therapists, who tested our wizard with clinical vignettes describing fictitious patients. Participants were able to successfully define services while revealing an ease of use of our wizard.

We formulate the entropy of a quantized artificial neural network as a differentiable function that can be plugged as a regularization term into the cost function minimized by gradient descent. Our formulation scales efficiently beyond the first order and is agnostic of the quantization scheme. The network can then be trained to minimize the entropy of the quantized parameters, so that they can be optimally compressed via entropy coding. We experiment with our entropy formulation at quantizing and compressing well-known network architectures over multiple datasets. Our approach compares favorably over similar methods, enjoying the benefits of higher order entropy estimate, showing flexibility towards non-uniform quantization (we use Lloyd-max quantization), scalability towards any entropy order to be minimized and efficiency in terms of compression. We show that HEMP is able to work in synergy with other approaches aiming at pruning or quantizing the model itself, delivering significant benefits in terms of storage size compressibility without harming the model’s performance.

We demonstrate the detection of speech signals and footsteps through a novel polarization-multiplexed coded differential-phase OTDR (optical time domain reflectometry) interrogation of standard telecom fibers in different configurations. We assess the sensitivity to audio signals by measuring the induced phase variation with respect to the sound pressure level of the perturbation, and manage to detect acoustic signals with high reliability. Further experiments are performed using an installed fiber cable showing the ability to track footsteps in an indoor scenario. These experiments pave the way for numerous applications of Distributed Acoustic Sensing (DAS) where a high sensitivity is required.

Side channel attacks (SCAs) on neural networks (NNs) are particularly efficient for retrieving secret information from NNs. We differentiate multiple types of threat scenarios regarding what kind of information is available before the attack and its purpose: recovering hyperparameters (the architecture) of the targeted NN, its weights (parameters), or its inputs. In this survey article, we consider the most relevant attacks to extract the architecture of CNNs. We also categorize SCAs, depending on access with respect to the victim: physical, local, or remote. Attacks targeting the architecture via local SCAs are most common. As of today, physical access seems necessary to retrieve the weights of an NN. We notably describe cache attacks, which are local SCAs aiming to extract the NN's underlying architecture. Few countermeasures have emerged; these are presented at the end of the survey.

International audience; Cyber physical systems are built upon digital and analog circuits, making it necessary to handle different models of computation during their design and verification (e.g., by simulation). When designing these systems, an important aspect to consider is the causality between the different domains. For this, we introduce a new model-driven framework able to identify causality problems and to suggest a valid schedule between the analog and digital domains. Once a valid schedule has been computed, our framework can generate cycle and bit accurate virtual prototypes (in SystemC/SystemC AMS) from high-level SysML models.

Speckle reduction is a key step in many remote sensing applications. By strongly affecting synthetic aperture radar (SAR) images, it makes them difficult to analyse. Due to the difficulty to model the spatial correlation of speckle, a deep learning algorithm with self-supervision is proposed in this paper: SAR2SAR. Multi-temporal time series are leveraged and the neural network learns to restore SAR images by only looking at noisy acquisitions. To this purpose, the recently proposed noise2noise framework has been employed. The strategy to adapt it to SAR despeckling is presented, based on a compensation of temporal changes and a loss function adapted to the statistics of speckle. A study with synthetic speckle noise is presented to compare the performances of the proposed method with other state-of-the-art filters. Then, results on real images are discussed, to show the potential of the proposed algorithm. The code is made available to allow testing and reproducible research in this field., The manuscript is the accepted version of IEEE STARS. Code is made available at https://gitlab.telecom-paris.fr/RING/SAR2SAR

Optical communications networks include basic components such as EDFAs and WSSs that present a polarization gain imbalance. The gain imbalance at the component output depends on the incident state-of-polarization. This well-known phenomenon called Polarization Dependent Loss (PDL) is revisited in this article and systematically analyzed within the framework of communications theory and practical achievable information rates. The rate impairment induced by PDL is discussed on an optical channel with distributed PDL and on a simplified equivalent lumped PDL channel. We have previously designed the New Spatially Balanced (NSB) signaling, an optimal 4D modulation based on this simplified model. We explore in this article its performance through numerical simulations and experimentally over a transmission in a recirculating loop with in-line PDL. Experimental measurements confirm an accurate channel modeling and significant information rate gains brought by NSB that results in SNR gains of the order of 0.4 dB at a practical transmission outage.

International audience; When implemented in software (or hardware), a cryptographic protocol can leak sensitive information during its execution. Side-channel attacks can use those leakages in order to reveal some information about the secret used by the algorithm. The leaking side-channel information can take place in many time samples. Measurement appliances can cope with the acquisition of multiple samples. From an adversarial point of view, it is therefore beneficial to attempt to make the most of highly multivariate traces. On the one hand, template attacks have been introduced to deal with multivariate leakages, with as few assumptions as possible on the leakage model. On the other hand, many works have underlined the need for dimensionality reduction. In this paper, we clarify the relationship between template attacks in full space and in linear subspaces, in terms of success rate. In particular, we exhibit a clear mathematical expression for template attacks, which enables an efficient computation even on large dimensions such as several hundred of samples. It is noteworthy that both of PoI-based and PCA-based template attacks can straightforwardly benefit from our approach. Furthermore, we extend the approach to the masking-based protected implementations. Our approach is validated both by simulated and real-world traces.

International audience; In a vertex-coloured graph, a set of vertices S is said to be a rainbow set if every colour in the graph appears exactly once in S. We investigate the complexities of various problems dealing with domination in vertex-coloured graphs (existence of rainbow dominating sets, of rainbow locating-dominating sets, of rainbow identifying sets), including when we ask for a unique solution: we show equivalence between these complexities and those of the well-studied Boolean satisfiability problems.

International audience; YAGO is one of the large knowledge bases in the Linked Open Data cloud. In this resource paper, we present its latest version, YAGO 4, which reconciles the rigorous typing and constraints of schema.org with the rich instance data of Wikidata. The resulting resource contains 2 billion type-consistent triples for 64 Million entities, and has a consistent ontology that allows semantic reasoning with OWL 2 description logics.

International audience; We address the estimation of the Ricean K factor when the available complex channel samples are noisy and subject to Nakagami-m shadowing, i.e., the line-of-sight component is modeled as a Nakagami-m random variable. We propose two estimators: one based on the expectation-maximization (EM) procedure, and a second one based on the method of moment (MoM). The MoM estimator can be used to initialize the EM procedure. We show by simulations that the two proposed estimators outperform the existing ones.

In this work, we consider the problem of blind source separation (BSS) by departing from the usual linear model and focusing on the linear-quadratic (LQ) model. We propose two provably robust and computationally tractable algorithms to tackle this problem under separability assumptions which require the sources to appear as samples in the data set. The first algorithm generalizes the successive nonnegative projection algorithm (SNPA), designed for linear BSS, and is referred to as SNPALQ. By explicitly modeling the product terms inherent to the LQ model along the iterations of the SNPA scheme, the nonlinear contributions of the mixing are mitigated, thus improving the separation quality. SNPALQ is shown to be able to recover the ground truth factors that generated the data, even in the presence of noise. The second algorithm is a brute-force (BF) algorithm, which is used as a post-processing step for SNPALQ. It enables to discard the spurious (mixed) samples extracted by SNPALQ, thus broadening its applicability. The BF is in turn shown to be robust to noise under easier-to-check and milder conditions than SNPALQ. We show that SNPALQ with and without the BF postprocessing is relevant in realistic numerical experiments., 23 pages + 24 pages of Appendix containing the proofs

We propose a method to estimate the amplitude of an unexpected power loss which, leveraging on a calibration, enables the real-time monitoring of a network link. It is based on an existing fiber-longitudinal power profile evaluation technique. The reliability of the method is assessed experimentally. When the anomaly is located at 0 km from the beginning of the span, the estimation bias is smaller than 0.2 dB for losses up to 10 dB. When the anomaly is located at 25 km from the beginning of the span, the same estimation bias is observed but for losses up to 5 dB. In both cases, the standard deviation of the estimation is smaller than 0.2 dB.

We introduce dual-polarization probing codes based on two circularly shifted frequency sweep signals enabling perfect channel estimation. This is achieved with a probing length equal to at least twice the fiber round-trip propagation time., Conference paper (27th International Conference on Optical Fiber Sensors), 4 pages, 7 figures

International audience; Affinity diagramming is widely applied to analyze qualitative data such as interview transcripts. It involves multiple analytic processes and is often performed collaboratively. Drawing on interviews with three practitioners and upon our own experience, we show how practitioners combine multiple analytic processes and adopt different artifacts to help them analyze their data. Current tools, however, fail to adequately support mixing analytic processes, devices, and collaboration styles. We present a vision and prototype ADQDA, a cross-device, collaborative affinity diagramming tool for qualitative data analysis, implemented using distributed web technologies. We show how this approach enables analysts to appropriate available pertinent digital devices as they fluidly migrate between analytic phases or adopt different methods and representations, all while preserving consistent analysis artifacts. We validate this approach through a set of application scenarios that explore how it enables new ways of analyzing qualitative data that better align with identified analytic practices. CCS Concepts: • Human-centered computing → Interactive systems and tools; • Information systems → Collaborative and social computing systems and tools.

International audience; Systems engineering, or engineering in general, has long been relying on document-centric approaches. Switching to model-based systems engineering, or MBSE for short, has extensively been discussed over the past three decades. Since about two decades, MBSE has been commonly associated with the modeling language SysML (Systems Modeling Language), which offers a standardized notation, not a methodology of using it. SysML needs therefore to be associated with a methodology supported by tools. In this paper, a methodology supported by the free and open-source software TTool is associated with SysML. This paper focuses discussion on methodological issues, leading the authors to share their experience in real-time systems modeling. Modeling with SysML is more than just drawing the different diagrams. Associated tools offer possibilities to analyze SysML models for specific properties. In this paper, verification addresses both safety and security properties. The TTool model checker inputs the SysML model enriched with safety properties to be verified and outputs an yes/no answer for each property. Security verification checks SysML models against confidentiality, integrity, and authenticity properties. As an illustration of the proposed approach, an aircraft cockpit door control system is modeled in SysML and verified against safety and security properties.

In modern cloud data centers, reconfigurable devices (FPGAs) are used as an alternative to Graphics Processing Units to accelerate data-intensive computations (e.g., machine learning, image and signal processing). Currently, FPGAs are configured to execute fixed workloads, repeatedly over long periods of time. This conflicts with the needs, proper to cloud computing, to flexibly allocate different workloads and to offer the use of physical devices to multiple users. This raises the need for novel, efficient FPGA scheduling algorithms that can decide execution orders close to the optimum in a short time. In this context, we propose a novel scheduling heuristic where groups of tasks that execute together are interposed by hardware reconfigurations. Our contribution is based on gathering tasks around a high-latency task that hides the latency of tasks, within the same group, that run in parallel and have shorter latencies. We evaluated our solution on a benchmark of 37500 random workloads, synthesized from realistic designs (i.e., topology, resource occupancy). For this testbench, on average, our heuristic produces optimum makespan solutions in 47.4% of the cases. It produces acceptable solutions for moderately constrained systems (i.e., the deadline falls within 10% of the optimum makespan) in 90.1% of the cases.

Speckle fluctuations seriously limit the interpretability of synthetic aperture radar (SAR) images. Speckle reduction has thus been the subject of numerous works spanning at least four decades. Techniques based on deep neural networks have recently achieved a new level of performance in terms of SAR image restoration quality. Beyond the design of suitable network architectures or the selection of adequate loss functions, the construction of training sets is of uttermost importance. So far, most approaches have considered a supervised training strategy: the networks are trained to produce outputs as close as possible to speckle-free reference images. Speckle-free images are generally not available, which requires resorting to natural or optical images or the selection of stable areas in long time series to circumvent the lack of ground truth. Self-supervision, on the other hand, avoids the use of speckle-free images. We introduce a self-supervised strategy based on the separation of the real and imaginary parts of single-look complex SAR images, called MERLIN (coMplex sElf-supeRvised despeckLINg), and show that it offers a straightforward way to train all kinds of deep despeckling networks. Networks trained with MERLIN take into account the spatial correlations due to the SAR transfer function specific to a given sensor and imaging mode. By requiring only a single image, and possibly exploiting large archives, MERLIN opens the door to hassle-free as well as large-scale training of despeckling networks. The code of the trained models is made freely available at https://gitlab.telecom-paris.fr/RING/MERLIN., To appear on IEEE Transactions on Geoscience and Remote Sensing

International audience; To compensate for the poor reliability of Physical Unclonable Function (PUF) primitives, some low complexity solutions not requiring error-correcting codes (ECC) have been proposed. One simple method is to discard less reliable bits, which are indicated in the helper data stored inside the PUF. To avoid discarding bits, the Two-metric Helper Data (TMH) method, which particularly applies to oscillation-based PUFs, allows to keep all bits by using different metrics when deriving the PUF response. However, oscillation-based PUFs are sensitive to side-channel analysis (SCA) since the frequencies of the oscillations can be observed by current or electromagnetic measurements. This paper studies the security of PUFs using TMH in order to obtain both reliable and robust PUF responses. We show that PUFs using TMH are sensitive to SCA, but can be greatly improved by using temporal masking and adapted extraction metrics. In case of public helper data, an efficient protection requires the randomization of the measurement order. We study two different solutions, providing interesting insights into trade-offs between security and complexity.

Embedded systems utilize Physically Unclonable Functions (PUFs) for authentication and identification purposes. However, modeling PUFs’ behavior via machine-learning methods has received utmost attention. Current research on modeling PUFs mainly targets a single PUF instance (PUF producing a single-bit response per query). It is admittedly more challenging to attack multi-bit parallel PUFs (with \(M>1\) PUF instances). In this work, we first target a multi-bit (mainly \(M=2\)-bit) parallel arbiter-PUF using its power traces, then introduce a hybrid countermeasure, combining Dual Rail Logic and Randomized Initialization Logic mechanisms, to thwart such attack. In addition, we explore Randomized Arbiter Swapping and Randomized Response Masking mitigation techniques for providing further protection for parallel PUFs against modeling attacks. To mimic the PUFs’ behavior in real silicon, we add noise artificially in our simulations. The results confirm the high success of the launched attack for the unprotected-PUF, and the resiliency of our countermeasures.

International audience; This paper analyses the multiplexing gain (MG) achievable over Wyner's symmetric network with random user activity and random arrival of mixed-delay traffic. The mixeddelay traffic is composed of delay-tolerant traffic and delaysensitive traffic where only the former can benefit from transmitter and receiver cooperation since the latter is subject to stringent decoding delays. The total number of cooperation rounds at transmitter and receiver sides is limited to D rounds. We derive inner and outer bounds on the MG region. In the limit as D → ∞, the bounds coincide and the results show that transmitting delaysensitive messages does not cause any penalty on the sum MG. For finite D our bounds are still close and prove that the penalty caused by delay-sensitive transmissions is small.

International audience; We use linear programming (LP) to derive upper and lower bounds on the "kissing number" A d of any q-ary code C with distance distribution frequencies Ai, in terms of the given parameters (n, M, d). In particular, a polynomial method gives explicit analytic bounds in a certain range of parameters, which are sharp for some low-rate codes like the first-order Reed-Muller codes. The general LP bounds are more suited to numerical estimates. Besides the classical estimation of the probability of decoding error and of undetected error, we outline recent applications in hardware protection against side-channel attacks using code-based masking countermeasures, where the protection is all the more efficient as the kissing number is low.

We study the design of data publishing mechanisms that allow a collection of autonomous distributed data sources to collaborate to support queries. A common mechanism for data publishing is via views : functions that expose derived data to users, usually specified as declarative queries. Our autonomy assumption is that the views must be on individual sources, but with the intention of supporting integrated queries. In deciding what data to expose to users, two considerations must be balanced. The views must be sufficiently expressive to support queries that users want to ask—the utility of the publishing mechanism. But there may also be some expressiveness restrictions. Here, we consider two restrictions, a minimal information requirement, saying that the views should reveal as little as possible while supporting the utility query, and a non-disclosure requirement, formalizing the need to prevent external users from computing information that data owners do not want revealed. We investigate the problem of designing views that satisfy both expressiveness and inexpressiveness requirements, for views in a restricted information systems - query languages (conjunctive queries), and for arbitrary views.

Cohesion is an affective group phenomenon. It has received a lot of attention from scholars both in Social Sciences and in Affective Computing that showed that cohesion and emotion influence each other, highlighting the need to jointly analyze them. This study presents 2 deep neural network architectures grounded on multitask learning to jointly predict cohesion and emotion. Inspired by 2 major Social Sciences approaches on group emotion (i.e., Top-down and Bottom-up), these architectures exploit cohesion and emotion interdependencies intending to improve the prediction of the dynamics (i.e. changes over time) of the Social and Task dimensions of cohesion. Emotion, here, is addressed in terms of its valence. Both architectures are evaluated against the performances of a similar model that only predicts the dynamics of both the Social and Task dimensions of cohesion, without integrating valence. Statistical analysis shows that only the deep model implementing the Bottom-up approach significantly improved the predictions of the Task cohesion’s dynamics. This result confirms the theoretical and practical benefits of multitasking as it takes full advantage of the inherent relationships between group emotion and cohesion to improve Task cohesion’s predictions.

International audience

Distributed optical fiber sensors using Brillouin scattering have already demonstrated their ability to measure changes in strain and temperature over long distance. However, the simultaneous sensitivity of the Brillouin scattering to these two parameters is a real disadvantage during the instrumentation and monitoring of structures, such as for civil engineering, oil industry applications (i.e. instrumentation of pipelines or wells) or electrical power cables monitoring. In this article, we present a new instrument enabling independently measurement of temperature or strain using a unique fiber in a single measurement. This method, which we will call B-OTDR DSTS for “Distributed Strain and Temperature Sensing”, is based on the measurement of frequency (as a standard BOTDR) but also intensity of Brillouin scattering. However, as the measurement of the latter requires an accurate measurement of power levels along the fiber, it is necessary to correct the losses by a Rayleigh OTDR measurement. The architecture of the BOTDR-DSTS allows simultaneous measurement of the Brillouin and Rayleigh backscatters. We will detail the validation results of this new device and show the interest of such a measurement on the instrumentation of a cable into a 2 km deep observation well.

International audience; Supervised source separation requires expensive synthetic datasets containing clean, ground truth-source signals, while unsupervised separation requires only data mixtures. Existing unsupervised methods still use supervision to avoid over-separation and compete with fully supervised methods. We present a new method of completely unsupervised single-channel blind source separation, based on variational auto-encoding, that automatically learns the correct number of sources in data mixtures and quantitatively outperforms the existing methods. A deep inference network disentangles (separates) data mixtures into low-dimensional latent source variables. A deep generative network individually decodes each latent source into its source signal, such that their sum represents the given mixture. Qualitative and quantitative results from separation experiments on pairs of randomly mixed MNIST handwritten digits and mixed audio spectrograms demonstrate that our method outperforms stateof-the-art unsupervised and semi-supervised methods, showing promise as a solution to this long-standing problem in computer vision and audition.

International audience; The analysis of load curves collected from smart meters is a key step for many energy management tasks ranging from consumption forecasting to customers characterization and load monitoring. In this contribution, we propose a model based on a functional formulation of nonnegative tensor factorization and derive updates for the corresponding optimization problem. We show on the concrete example of multi-sites load curves disaggregation how this formulation is helpful for 1) exhibiting smooth intraday consumption patterns and 2) taking into account external variables such as the outside temperature. The benefits are demonstrated on simulated and real data by exhibiting a meaningful clustering of the observed sites based on the obtained decomposition.

Intrusion Detection Systems are widely used to detect cyberattacks, especially on protocols vulnerable to hacking attacks such as SOME/IP. In this paper, we present a deep learning-based sequential model for offline intrusion detection on SOME/IP application layer protocol. To assess our intrusion detection system, we have generated and labeled a dataset 1 with several classes representing realistic intrusions, and a normal class-a significant contribution due to the absence of such publicly available datasets. Furthermore, we also propose a recurrent neural network (RNN), as an instance of deep learningbased sequential model, that we apply to our generated dataset. The numerical results show that RNN excel at predicting invehicle intrusions, with F1 Scores and AUC values greater than 0.8 depending on each intrusion type., Les systèmes de détection d'intrusion sont largement utilisés pour détecter les cyberattaques, en particulier sur les protocoles vulnérables aux attaques de piratage tels que SOME/IP. Dans cet article, nous présentons un modèle séquentiel basé sur l'apprentissage profond pour la détection d'intrusion hors ligne sur le protocole de couche d'application SOME/IP. Pour évaluer notre système de détection d'intrusion, nous avons généré et étiqueté un ensemble de données 1 avec plusieurs classes représentant des intrusions réalistes, et une classe normale - une contribution significative en raison de l'absence de tels ensembles de données accessibles au public. En outre, nous proposons également un réseau de neurones récurrents (RNN), en tant qu'instance de modèle séquentiel basé sur l'apprentissage en profondeur, que nous appliquons à notre ensemble de données généré. Les résultats numériques montrent que RNN excelle dans la prédiction des intrusions dans les véhicules, avec des scores F1 et des valeurs AUC supérieures à 0,8 en fonction de chaque type d'intrusion.

This paper proposes a novel energy-momentum conserving integration method. Adopting Projective Dynamics, the proposed method extends its unconstrained minimization for time integration into the constrained form with the position-based energy-momentum constraints. This resolves the well-known problem of unwanted dissipation of energy and momenta without compromising the real-time performance and simulation stability. The proposed method also enables users to directly control the energy and momenta so as to easily create the vivid deformable and global motions they want, which is a fascinating feature for many real-time applications such as virtual/augmented reality and games.

International audience

In this work, we study music/video cross-modal recommendation, i.e. recommending a music track for a video or vice versa. We rely on a self-supervised learning paradigm to learn from a large amount of unlabelled data. We rely on a self-supervised learning paradigm to learn from a large amount of unlabelled data. More precisely, we jointly learn audio and video embeddings by using their co-occurrence in music-video clips. In this work, we build upon a recent video-music retrieval system (the VM-NET), which originally relies on an audio representation obtained by a set of statistics computed over handcrafted features. We demonstrate here that using audio representation learning such as the audio embeddings provided by the pre-trained MuSimNet, OpenL3, MusicCNN or by AudioSet, largely improves recommendations. We also validate the use of the cross-modal triplet loss originally proposed in the VM-NET compared to the binary cross-entropy loss commonly used in self-supervised learning. We perform all our experiments using the Music Video Dataset (MVD).

International audience; We study challenge codes for physically unclonable functions (PUFs). Starting from the classical Hadamard challenge code, we augment it by one vector. Numerical values suggest that the optimal choice of this vector for maximizing the entropy is to pick a vector the farthest away from the code formed by the challenges and their binary complements. This leads us to study the covering radius of Hadamard codes. A notion of bent sequence that generalizes the classical notion from Hadamard matrices of Sylvester type to general Hadamard matrices is given. Lower bounds for Paley-type Hadamard matrices are given.

International audience; The information leakage of a cryptographic implementation with a given degree of protection is evaluated in a typical situation when the signal-to-noise ratio is small. This is solved by expanding Kullback-Leibler divergence, entropy, and mutual information in terms of moments/cumulants.

International audience; As the Intenert of Things (IoT) continues to evolve in our daily personal lives and in future industrial systems (industry 4.0), one of the most significant issues is security. IoT systems must overcome a number of challenges, including deficiency of resources, low power consumption, and the need to protect devices against cyber-attacks. Regrettably, issues about energy use and the lack of computing resources limit the cryptographic methods that can be implemented on these devices. Moreover, the conventional use of non-volatile memory for storing secret keys are vulnerable to a number of attacks like reverse-engineering, cold-boot, side channel, device tampering, etc. Physical Unclonable Functions (PUFs) are one of the categories for enhancing physical device security and solving issues involving with the use of traditional cryptographic algorithms. PUFs are associated as lightweight one-way functions used to extract a unique identity for each end-device, based on physical factors introduced during manufacturing which are unforeseeable and unclonable. PUFs are promising hardware security primitive and have seen a lot of attention in the past few years. In this paper, we provide a survey of a large range of PUF schemes proposed in the literature. Our comparison analysis of the surveyed schemes ends with a number of observations. Then we propose a new scheme of Arbiter PUF to create an identity for each IoT device. The proposed scheme is resistant to Machine Learning (ML) attacks. It is proven to avoid the shortcoming of several previously proposed related PUF-based authentication protocols.

Deep learning (DL) has gained popularity in network intrusion detection, due to its strong capability of recognizing subtle differences between normal and malicious network activities. Although a variety of methods have been designed to leverage DL models for security protection, whether these systems are vulnerable to adversarial examples (AEs) is unknown. In this article, we design a novel adversarial attack against DL-based network intrusion detection systems (NIDSs) in the Internet-of-Things environment, with only black-box accesses to the DL model in such NIDS. We introduce two techniques: 1) model extraction is adopted to replicate the black-box model with a small amount of training data and 2) a saliency map is then used to disclose the impact of each packet attribute on the detection results, and the most critical features. This enables us to efficiently generate AEs using conventional methods. With these tehniques, we successfully compromise one state-of-the-art NIDS, Kitsune: the adversary only needs to modify less than 0.005% of bytes in the malicious packets to achieve an average 94.31% attack success rate.

This study investigates chaotic and spiking dynamics of mid-infrared quantum cascade lasers operating under external optical feedback and emitting at 5.5 $\mu {\text{m}}$ and 9 $\mu {\text{m}}$ . In order to deepen the understanding, the route to chaos is experimentally studied in the case of continuous-wave and current modulation operation. The non-linear dynamics are analyzed with bifurcation diagrams. While for quasi-continuous wave operation, chaos is found to be more complex, pure continuous wave pumping always leads to the generation of a regular spiking induced low-frequency fluctuations dynamics. In the latter, results show that by combining external optical feedback with periodic forcing and further induced current modulation allows a better control of the chaotic dropouts. This work provides a novel insight into the development of future secure free-space communications based on quantum cascade lasers or unpredictable optical countermeasure systems operating within the two transparency atmospheric windows hence between 3 $\mu {\text{m}}$ –5 $\mu {\text{m}}$ and 8.5 $\mu {\text{m}}$ –11 $\mu {\text{m}}$ .

International audience; We reconsider the problem of containment of monadic datalog (MDL) queries in unions of conjunctive queries (UCQs). Prior work has dealt with special cases of the problem, but has left the precise complexity characterization open. In addition, the complexity of one important special case, that of containment under access patterns, was not known before. We start by revisiting the connection between MDL/UCQ containment and containment problems involving regular tree languages. We then present a general approach for getting tighter bounds on the complexity of query containment, based on analysis of the number of mappings of queries into tree-like instances. We give two applications of the machinery. We first give an important special case of the MDL/UCQ containment problem that is in EXPTIME, and use this bound to show an EXPTIME bound on containment under access patterns. Secondly we show that the same technique can be used to get a new tight upper bound for containment of tree automata in UCQs. We finally show that the new MDL/UCQ upper bounds are tight. We establish a 2EXPTIME lower bound on the MDL/UCQ containment problem, resolving an open problem from the early 1990s. This bound holds for the MDL/CQ containment problem as well. We also show that changes to the conditions given in our special cases can not be eliminated, and that in particular slight variations of the problem of containment under access patterns become 2EXPTIME-complete.

Linear video coding (LVC) has recently emerged as a very promising paradigm for robust video delivery over unreliable channels. The pioneering LVC paper, introducing SoftCast, showed that, a cross-layer approach, based on linear-only processing followed by analog modulation, was able to improve the quality of received video with respect to traditional, layered approaches, when unreliable links to several users are considered. This paper considers the power allocation problem when the transmission channel consists of several sub-channels, each with its own power constraint. This framework models, e.g., multi-antenna, DSL, or powerline transmission systems. Minimizing the mean-square error at a receiver leads to an precoding and decoding matrix design problem. Our contributions consist in (1) formulating the matrix design problem in such a way that a multi-level water-filling solution proposed for MIMO channels can be adapted to LVC; (2) the proposal of three suboptimal power allocations techniques, with different trade-offs between complexity and efficiency; (3) a precoding matrix design for the multiple receivers scenario; (4) extensive simulations of the power allocation methods. The proposed optimal power allocation largely outperforms a simple adaptation of the allocation considered in SoftCast for the per-subchannel power constrained problem; moreover two of the suboptimal solutions perform very close to the optimal one, with a sensibly reduced complexity.

Multi-core architectures pose many challenges in real-time systems, which arise from contention between concurrent accesses to shared memory. Among the available memory arbitration policies, time-division multiplexing (TDM) ensures a predictable behavior by bounding access latencies and guaranteeing bandwidth to tasks independently from the other tasks. To do so, TDM guarantees exclusive access to the shared memory in a fixed time window. TDM, however, provides a low resource utilization as it is non-work-conserving. Besides, it is very inefficient for resources having highly variable latencies, such as sharing the access to a DRAM memory. The constant length of a TDM slot is, hence, highly pessimistic and causes an underutilization of the memory. To address these limitations, we present dynamic arbitration schemes that are based on TDM. However, instead of arbitrating at the level of TDM slots, our approach operates at the granularity of clock cycles by exploiting slack time accumulated from preceding requests. This allows the arbiter to reorder memory requests, exploit the actual access latencies of requests, and thus improve memory utilization. We demonstrate that our policies are analyzable as they preserve the guarantees of TDM in the worst case, while our experiments show an improved memory utilization. We furthermore present and evaluate an efficient hardware implementation for a variant of our arbitration strategy.

International audience; High performance streaming applications require hardware platforms featuring complex, multi-level interconnects. These applications often resemble a task-farm, where many identical tasks listen to the same input channel. Usual embedded system design tools are not well adapted to capture these applications. In particular, the non-uniform memory access (NUMA) nature of the platforms induces latencies that must be carefully examined. The paper proposes a multi-level modeling methodology and tools (TTool, SoCLib) that have been extended to model the characteristics of streaming applications (multiple tasks, non deter-ministic behavior, I/O devices) in UML/SysML, and to automatically generate a virtual prototype that can be simulated with high precision. The paper uses a typical streaming application to show how latencies can be estimated and fed back to diagrams.

Considering the vigorous drive to insert ever more artificial intelligence (AI) enabled technology into modern societies - from household appliances and personal assistants to business planning and guidance systems - it seems urgent to ensure that the kinds of AI we develop and deploy provide useful tools in the service of humanity, rather than constraining frameworks to limit humanity. In other words, we need "AI for humanity", rather than humanity for AI; or, "human centered AI" rather than "function oriented AI". This challenge is hardly confined to the AI domain. It extends to most computer-based systems, and, indeed, to all technology. This article aims to bring to the fore: the implicit values behind current technological developments - mostly efficiency-driven; the potential negative impacts of unquestioned technological developments and usage - e.g., the totalizing supremacy of quantity over quality; alternative ways of developing and adopting technology - e.g., as tools rather than controllers; and the necessity to permanently analyze, evaluate and alter technical systems during development, before adoption, and as their side effects become obvious. It also aims to emphasise that criticizing certain kinds of technologies is not at all equivalent to being technophobe, or against progress. That would be like equating a critique of fast food to an unnatural stance against eating. Finally, technical developments cannot be considered in isolation. They are a key part of a self-promoting system of market-driven production and sociopolitical transformation.

Using information-theoretic tools, this paper establishes a mathematical link between the probability of success of a side-channel attack and the minimum number of queries to reach a given success rate, valid for any possible distinguishing rule and with the best possible knowledge on the attacker’s side. This link is a lower bound on the number of queries highly depends on Shannon’s mutual information between the traces and the secret key. This leads us to derive upper bounds on the mutual information that are as tight as possible and can be easily calculated. It turns out that, in the case of an additive white Gaussian noise, the bound on the probability of success of any attack is directly related to the signal to noise ratio. This leads to very easy computations and predictions of the success rate in any leakage model.

International audience

Learning from ephemeral data streams has garnered the interest of both researchers and practitioners towards adaptive learning techniques. Despite the convincing results obtained thus far, most of the current research still overlooks that the relevance of features may change throughout the learning process. Scenarios where features become - or cease to be - relevant to the learning task are called feature drifting data streams, and the identification of which features are relevant becomes even more challenging when the feature space is high-dimensional. To select relevant features during the progress of data streams, we propose a merit-guided and classifier-independent dynamic feature selection algorithm named DynamIc SymmetriCal Uncertainty Selection for Streams (DISCUSS). We evaluate our proposal on both synthetic and real-world datasets and show that DISCUSS can boost kNN and Naive Bayes classifiers’ accuracy rates on high-dimensional data streams, while at the expense of limited processing time and memory space. Finally, the drawbacks of the proposed method are assessed, and possible future works on the topic are also discussed.

Data fragmentation and dispersion is recognized as a way of providing confidentiality and availability of data stored inside unattended wireless sensor networks. This paper introduces an additively homomorphic encryption and fragmentation scheme (AHEF). AHEF replaces additively homomorphic secret sharing used in state-of-the-art techniques with additively homomorphic fragmentation. This change has a significant impact on the volume of data stored inside the sensors which in turn supports lower transmission costs. Both, storage and transmission costs are reduced by a factor of at least 2. Moreover, by reducing the number of required computations, AHEF allows sensors to save energy.

We semianalytically and numerically investigate the static and dynamical characteristics of quantum-dot (QD) lasers directly grown on silicon by considering the Shockley-Read-Hall (SRH) recombination. The static characteristics are studied through small-signal analysis, including the ${\ensuremath{\alpha}}_{H}$ factor, damping factor, and modulation dynamics. In addition, the feedback dynamics are analyzed through improved corpuscular rate equations based on the classical Lang-Kobayashi (LK) model with time series, bifurcation diagrams, and phase portraits. We find that a smaller ${\ensuremath{\alpha}}_{H}$ factor but larger damping factor are obtained by decreasing the nonradiative recombination lifetime. On top of that, in both the short- and long-external-cavity regimes, any decrease of the SRH recombination lifetime obliterates significantly chaotic regions and shifts the first Hopf bifurcation point to higher feedback values. Overall, this work provides insights into the understanding of QD laser physics, hence highlighting the influence of the SRH lifetime on the reflection sensitivity of epitaxial QD lasers on silicon. These results are qualitatively consistent with recent experiments and are therefore helpful for designing feedback-resistant lasers for future photonic integrated circuits operating without optical isolation.

International audience; We demonstrate the performance of a novel multi-carrier MIMO fiber sensing interrogator based on OFDM approach. Over an installed telecom fiber cable, coherent fading is mitigated by means of a purely digital subcarrier combiner.