Publications

2022

• The Dark Side of Perceptual Manipulations in Virtual Reality
  
  • Tseng Wen-Jie
  • Bonnail Elise
  • Mcgill Mark
  • Khamis Mohamed
  • Lecolinet Eric
  • Huron Samuel
  • Gugenheimer Jan
  , 2022, pp.1-15. Virtual-Physical Perceptual Manipulations" (VPPMs) such as redirected walking and haptics expand the user's capacity to interact with Virtual Reality (VR) beyond what would ordinarily physically be possible. VPPMs leverage knowledge of the limits of human perception to efect changes in the user's physical movements, becoming able to (perceptibly and imperceptibly) nudge their physical actions to enhance interactivity in VR. We explore the risks posed by the malicious use of VPPMs. First, we defne, conceptualize and demonstrate the existence of VPPMs. Next, using speculative design workshops, we explore and characterize the threats/risks posed, proposing mitigations and preventative recommendations against the malicious use of VPPMs. Finally, we implement two sample applications to demonstrate how existing VPPMs could be trivially subverted to create the potential for physical harm. This paper aims to raise awareness that the current way we apply and publish VPPMs can lead to malicious exploits of our perceptual vulnerabilities. (10.1145/3491102.3517728)
  DOI : 10.1145/3491102.3517728
• Co-producing industrial public goods on GitHub: Selective firm cooperation, volunteer-employee labour and participation inequality
  
  • O'Neil Mathieu
  • Cai Xiaolan
  • Muselli Laure
  • Zacchiroli Stefano
  New Media and Society, SAGE Publications, 2022. The global economy’s digital infrastructure is based on free and open source software. To analyse how firms indirectly collaborate via employee contributions to developer-run projects, we propose a formal definition of ‘industrial public goods’ – inter-firm cooperation, volunteer and paid labour overlap, and participation inequality. We verify its empirical robustness by collecting networks of commits made by firm employees to active GitHub software repositories. Despite paid workers making more contributions, volunteers play a significant role. We find which firms contribute most, which projects benefit from firm investments, and identify distinct ‘contribution territories’ since the two central firms never co-contribute to top-20 repositories. We highlight the challenge posed by ‘Big Tech’ to the non-rival status of industrial public goods, thanks to cloud-based systems which resist sharing, and suggest there may be ‘contribution deserts’ neglected by large information technology firms, despite their importance for the open source ecosystem’s sustainability and diversity. (10.1177/14614448221090474)
  DOI : 10.1177/14614448221090474
• Cooperation as a signal of time preferences
  
  • Lie-Panis Julien
  • André Jean-Baptiste
  Proceedings of the Royal Society B: Biological Sciences, Royal Society, The, 2022, 289 (1973). Many evolutionary models explain why we cooperate with non-kin, but few explain why cooperative behaviour and trust vary. Here, we introduce a model of cooperation as a signal of time preferences, which addresses this variability. At equilibrium in our model (i) future-oriented individuals are more motivated to cooperate, (ii) future-oriented populations have access to a wider range of cooperative opportunities, and (iii) spontaneous and inconspicuous cooperation reveal stronger preference for the future, and therefore inspire more trust. Our theory sheds light on the variability of cooperative behaviour and trust. Since affluence tends to align with time preferences, results (i) and (ii) explain why cooperation is often associated with affluence, in surveys and field studies. Time preferences also explain why we trust others based on proxies for impulsivity, and, following result (iii), why uncalculating, subtle and one-shot cooperators are deemed particularly trustworthy. Time preferences provide a powerful and parsimonious explanatory lens, through which we can better understand the variability of trust and cooperation. (10.1098/rspb.2021.2266)
  DOI : 10.1098/rspb.2021.2266
• Towards a web-of-things approach for OPC UA field device discovery in the industrial IoT
  
  • Nguyen Quang-Duy
  • Dhouib Saadia
  • Chanet Jean-Pierre
  • Bellot Patrick
  , 2022, pp.1-4. In the OPC UA standard, the OPC UA information model is the key to semantic interoperability. It refers to an organized structure representing resources in the form of OPC UA nodes residing in an OPC UA address space. Before runtime, an OPC UA server requires an OPC UA information model already filled with OPC UA nodes corresponding to the available resources of its field devices. It is tricky to the Industrial Internet of Things, a scenario where random field devices can join a system at any time. Indeed, the OPC UA server needs to dynamically discover such devices and update its OPC UA information model with new relevant OPC UA nodes. Regarding the above challenge, this paper introduces a Web-of-Things (WoT) approach that enables field devices to register with an OPC UA server to join the system without manual configuration. The approach relies on WoT Discovery and WoT Thing Description. (10.1109/WFCS53837.2022.9779181)
  DOI : 10.1109/WFCS53837.2022.9779181
• Unrolling PALM for sparse semi-blind source separation
  
  • Fahes Mohammad
  • Kervazo Christophe
  • Bobin Jérôme
  • Tupin Florence
  , 2022. Sparse Blind Source Separation (BSS) has become a well established tool for a wide range of applications - for instance, in astrophysics and remote sensing. Classical sparse BSS methods, such as the Proximal Alternating Linearized Minimization (PALM) algorithm, nevertheless often suffer from a difficult hyperparameter choice, which undermines their results. To bypass this pitfall, we propose in this work to build on the thriving field of algorithm unfolding/unrolling. Unrolling PALM enables to leverage the data-driven knowledge stemming from realistic simulations or ground-truth data by learning both PALM hyperparameters and variables. In contrast to most existing unrolled algorithms, which assume a fixed known dictionary during the training and testing phases, this article further emphasizes on the ability to deal with variable mixing matrices (a.k.a. dictionaries). The proposed Learned PALM (LPALM) algorithm thus enables to perform semi-blind source separation, which is key to increase the generalization of the learnt model in real-world applications. We illustrate the relevance of LPALM in astrophysical multispectral imaging: the algorithm not only needs up to $10^4-10^5$ times fewer iterations than PALM, but also improves the separation quality, while avoiding the cumbersome hyperparameter and initialization choice of PALM. We further show that LPALM outperforms other unrolled source separation methods in the semi-blind setting.
• Shape Reconstruction in 2D: From Theory to Practice
  
  • Ohrhallinger S
  • Parakkat Amal Dev
  • Peethambaran J
  , 2022. Shape reconstruction from unstructured points in a plane is a fundamental problem with many applications that has generated research interest for decades. Involved aspects like handling open, sharp, multiple and non-manifold outlines, run-time and provability as well as potential extension to 3D for surface reconstruction have led to many different algorithms. This multitude of reconstruction methods with quite different strengths and focus makes it a difficult task for users to choose a suitable algorithm for their specific problem. In this tutorial, we present the development history of algorithms, together with their related proximity graphs, all in detail. Then, we show algorithms targeted at specific problem classes, such as reconstructing from noise, outliers, or sharp corners. We will also include the latest developments in the field, namely based on Voronoi balls and the sphere-of-influence graph. Examples of the evaluation will show how its results can guide users to select an appropriate algorithm for their input data. We will also explain how to integrate new algorithms into our benchmark framework. Region reconstruction will be shown as an additional field closely related to boundary reconstruction.
• Flexible distribution and quantum state tomography of frequency entangled photon pairs from a 21GHz SOI frequency comb using frequency quantum gates
  
  • Henry Antoine
  • Fioretto Dario
  • Procopio Lorenzo
  • Monfray Stephane
  • Boeuf Frédéric
  • Vivien Laurent
  • Cassan Eric
  • Ramos Carlos Alonso
  • Bencheikh K.
  • Zaquine Isabelle
  • Belabas Nadia
  , 2022.
• Four-wave mixing in 1.3 μm epitaxial quantum dot lasers directly grown on silicon
  
  • Duan Jianan
  • Dong Bozhang
  • Chow Weng W
  • Huang Heming
  • Ding Shihao
  • Liu Songtao
  • Norman Justin C
  • Bowers John E
  • Grillot Frédéric
  Photonics research, Optical Society of America, 2022, 10 (5), pp.1264-1270. This work compares the four-wave mixing (FWM) effect in epitaxial quantum dot (QD) lasers grown on silicon with quantum well (QW) lasers. A comparison of theory and experiment results shows that the measured FWM coefficient is in good agreement with theoretical predictions. The gain in signal power is higher for p-doped QD lasers than for undoped lasers, despite the same FWM coefficient. Owing to the near-zero linewidth enhancement factor, QD lasers exhibit FWM coefficients and conversion efficiency that are more than one order of magnitude higher than those of QW lasers. Thus, this leads to self-mode locking in QD lasers. These findings are useful for developing on-chip sources for photonic integrated circuits on silicon. (10.1364/prj.448082)
  DOI : 10.1364/prj.448082
• High speed mid-infrared Stark modulator for optical data transmission up to 10 Gbit.s−1
  
  • Bonazzi Thomas
  • Dely Hamza
  • Spitz Olivier
  • Rodriguez Etienne
  • Gacemi Djamal
  • Todorov Yanko
  • Pantzas Konstantinos
  • Beaudoin Grégoire
  • Sagnes Isabelle
  • Grillot Frédéric
  • Vasanelli Angela
  • Sirtori Carlo
  , 2022, pp.ATu4O.2. Using the Stark effect in coupled InGas/AlInAs quantum wells, we demonstrate a mid-infrared broadband optoelectronic external modulator enabling 10 Gbit/s free space optical data-transmission in the second atmospheric window (9 μm) at room temperature. (10.1364/CLEO_AT.2022.ATu4O.2)
  DOI : 10.1364/CLEO_AT.2022.ATu4O.2
• Modeling Rowhammer memory corruption in the gem5 simulator
  
  • France Loïc
  • Bruguier Florent
  • Mushtaq Maria
  • Novo David
  • Benoit Pascal
  , 2022. In modern computers, the main memory is the target of a security threat called Rowhammer, which causes bit flips in adjacent victim cells of aggressor rows. Numerous countermeasures have been proposed, some of the most efficient ones relying on memory controller modifications, which make them non-integrable in existing systems. These solutions have to be effective against attacks on current and future architectures and technology nodes. In order to prove the efficiency of such mitigation techniques, we have to use simulation platforms. Unfortunately, existing architecture simulators do not provide any implementation of unintended memory modifications like bitflips. Integrating memory corruption into architecture simulators would allow the construction of attacks and mitigations for current and future computers, using feedback from the simulator. In this paper, we propose an implementation of the Rowhammer effect in the gem5 architecture simulator, demonstrate its capabilities and state its limitations.
• Functional time series modeling and application to representation and analysis of multi-site electric load curves for energy management
  
  • Durand Amaury
  , 2022. The analysis of electrical load curves collected by smart meters is a key step for many energy management tasks ranging from consumption forecasting and load monitoring to customers characterization and segmentation. In this context, researchers from EDF R&D are interested in extracting significant information from the daily electrical load curves in order to compare the consumption behaviors of different buildings. The strategy followed by the group which hosted my doctorate is to use physical and deterministic models based on information such as the room size, the insulating materials or weather data, or to extract hand-designed patterns from the electrical load curves based on the knowledge of experts. Given the growing amount of data collected, the interest of the group in statistical or data-driven methods has increased significantly in recent years. These approaches should provide new solutions capable of exploiting massive data without relying on expensive processing and expert knowledge. My work fits directly into this trend by proposing two modeling approaches: the first approach is based on functional time series and the second one is based on non-negative tensor factorization. This thesis is split into three main parts. In the first part, we present the industrial context and the practical objective of the thesis, as well as an exploratory analysis of the data and a discussion on the two modeling approaches proposed. In the second part, we follow the first modeling approach and provide a thorough study of the spectral theory for functional time series. Finally, the second modeling approach based on non-negative tensor factorization is presented in the third part.
• Functional anomaly detection and robust estimation
  
  • Staerman Guillaume
  , 2022. Enthusiasm for Machine Learning is spreading to nearly all fields such as transportation, energy, medicine, banking or insurance as the ubiquity of sensors through IoT makes more and more data at disposal with an ever finer granularity. The abundance of new applications for monitoring of complex infrastructures (e.g. aircrafts, energy networks) together with the availability of massive data samples has put pressure on the scientific community to develop new reliable Machine-Learning methods and algorithms. The work presented in this thesis focuses around two axes: unsupervised functional anomaly detection and robust learning, both from practical and theoretical perspectives.The first part of this dissertation is dedicated to the development of efficient functional anomaly detection approaches. More precisely, we introduce Functional Isolation Forest (FIF), an algorithm based on randomly splitting the functional space in a flexible manner in order to progressively isolate specific function types. Also, we propose the novel notion of functional depth based on the area of the convex hull of sampled curves, capturing gradual departures from centrality, even beyond the envelope of the data, in a natural fashion. Estimation and computational issues are addressed and various numerical experiments provide empirical evidence of the relevance of the approaches proposed. In order to provide recommendation guidance for practitioners, the performance of recent functional anomaly detection techniques is evaluated using two real-world data sets related to the monitoring of helicopters in flight and to the spectrometry of construction materials.The second part describes the design and analysis of several robust statistical approaches relying on robust mean estimation and statistical data depth. The Wasserstein distance is a popular metric between probability distributions based on optimal transport. Although the latter has shown promising results in many Machine Learning applications, it suffers from a high sensitivity to outliers. To that end, we investigate how to leverage Medians-of-Means (MoM) estimators to robustify the estimation of Wasserstein distance with provable guarantees. Thereafter, a new statistical depth function, the Affine-Invariant Integrated Rank-Weighted (AI-IRW) depth is introduced. Beyond the theoretical analysis carried out, numerical results are presented, providing strong empirical confirmation of the relevance of the depth function proposed. The upper-level sets of statistical depths—the depth-trimmed regions—give rise to a definition of multivariate quantiles. We propose a new discrepancy measure between probability distributions that relies on the average of the Hausdorff distance between the depth-based quantile regions w.r.t. each distribution and demonstrate that it benefits from attractive properties of data depths such as robustness or interpretability. All algorithms developed in this thesis are open-sourced and available online.
• Riemannian space tessellation with polyhedral room images
  
  • Polack Jean-Dominique
  • Meacham Aidan
  • Badeau Roland
  • Valière Jean-Christophe
  , 2022. Counting the images sources of rectangular rooms is a well known technique, based on mirroring the original rooms on all its walls in order to tesselate the Euclidian space, leading to a quadratic increase with layer order. We show that a similar mirroring technique can be applied to polygonal and polyhedral rooms of arbitrary shapes, leading to the tessellation of a Riemannian space with negative curvature. From this tessellation we derive a close formulation for counting the numbers of image sources, which increases exponentially with layer order. Thus, a bridge between rooms with flat walls and generic mixing rooms with partially curved walls is obtained.
• Confirming dimensional reduction assumptions for the energy-stress tensor through comparison with high-frequency wave-based pressure simulations
  
  • Meacham Aidan
  • Badeau Roland
  • Polack Jean-Dominique
  , 2022. In room acoustics, the energy-stress tensor represents the conservative relationships between the acoustic energy density, sound intensity, and the symmetric wave-stress tensor. In real rooms, the off-diagonal components of the wave-stress tensor are non-zero, implying the existence of shear stresses acting upon the energetic quantities. Assumptions regarding these terms in 1- and 2-dimensional spaces [Dujourdy et al. 2017, 2019] were used to reduce the energy-stress tensor relationships to a tractable system capable of predicting frequency- dependent stochastic reverberation decays in those spaces [Meacham et al. 2019]. Direct verification of those assumptions at a single location in a real space would require more measurements at varying positions than can be reliably captured without robotization, let alone in acoustically distinct regions of a room. Therefore, in this work, we aim to verify the 1-dimensional reduction assumptions by examining a high- frequency wave-based pressure simulation, allowing averaging over a wide number of sampling positions at multiple locations throughout a space, providing insight into the relationship between room geometry and the terms of the energy-stress tensor.
• Algorithmes rapides pour la modélisation d'une réponse de salle dont l'atténuation dépend de la fréquence
  
  • Aknin Achille
  • Badeau Roland
  , 2022. En traitement du signal audio, la modélisation mathématique de la réponse de salle permet d'améliorer la qualité de l'estimation de signaux sources à partir de signaux réverbérés, afin par exemple d’effectuer une déréverbération ou de séparer un mélange de sources sonores. Dans un article précédent, nous avons travaillé sur l'implémentation d’un modèle stochastique de réponse impulsionnelle de salle, dans lequel l’atténuation exponentielle de la puissance au cours du temps dépend de la fréquence. En effet cette caractéristique est particulièrement importante si l’on veut prendre en compte la dépendance fréquentielle de l'absorption des murs, qui est généralement supérieure en hautes fréquences par rapport aux basses fréquences. Nous avons présenté une nouvelle structure de matrice, paramétrée par un unique filtre ppp, qui réalise cette atténuation exponentielle dépendant de la fréquence, et nous avons montré qu’elle pouvait être utilisée pour estimer les paramètres d'une réponse de salle. Cependant, cette matrice PPP étant de taille T×TT×TT \times T, où TTT est la longueur de la réponse impulsionnelle de la salle (généralement de l’ordre du millier ou de la dizaine de milliers d’échantillons en pratique), nous ne pouvons pas calculer directement des produits matriciels impliquant cette matrice PPP dans des conditions réelles pour des raisons de coût de calcul. Dans cet article, nous allons donc présenter plusieurs algorithmes rapides de produit matrice-vecteur que nous avons développés, qui exploitent la structure particulière de cette matrice, et dont la complexité est seulement de O(Tlog(T))O(Tlog⁡(T))O(T \log(T)) ou O(Tlog2(T))O(Tlog2⁡(T))O(T \log^2(T)) au lieu de O(T2)O(T2)O(T^2). Grâce à ces algorithmes, il devient possible d’exploiter la matrice PPP pour estimer les paramètres de vraies réponses de salle, sans être limité par la complexité de calcul.
• Riemannian space tessellation with polyhedral room images
  
  • Polack Jean-Dominique
  • Meacham Aidan
  • Badeau Roland
  • Valière Jean-Christophe
  , 2022.
• Spatio-Temporal Wireless D2D Network With Imperfect Beam Alignment
  
  • Quan Yibo
  • Coupechoux Marceau
  • Kelif Jean-Marc
  , 2022, pp.2346-2351. In this paper, we investigate the beam misalignment impacts of a dynamic device-to-device (D2D) communication model, where both transmitters and receivers adopt beamforming (BF) by using uniform linear array (ULA). A time continuous dynamic model is adopted for this network. We use tools of stochastic geometry and the Miyazawa rate conversation law to analyse the stability condition of such a network. An analytical expression of the critical arrival rate is given under a uniform or truncated Gaussian alignment error assumption. In contrast to our previous result, where the beam alignment is perfect, our analytical and numerical results show that, if the beam alignment is not perfect, the critical arrival rate can no longer increase without limit as a function of the number of antenna elements. Closed-form expressions of the upper bounds for critical arrival rates are given for both the uniform and the truncated Gaussian misalignment models. (10.1109/WCNC51071.2022.9771938)
  DOI : 10.1109/WCNC51071.2022.9771938
• Neural rendering for improved cosmetics virtual try-on
  
  • Kips Robin
  , 2022. Augmented reality applications have rapidly spread across online retail platforms and social media, allowing consumers to virtually try on a large variety of cosmetics products. However, even though appreciated by consumers, such applications currently offer limited realism compared to real product images. On the other hand, the rapidly emerging field of generative models and neural rendering offers new perspectives that we will study in this work for realistic image synthesis and novel virtual try-on experiences. First, we introduce a novel makeup synthesis method based on generative networks in which the makeup color can be explicitly controlled, similar to a physically-based renderer. Our model obtains photorealistic results on lips and eyes makeup in high resolution. Furthermore, we relax the need for labeled data by introducing a weakly-supervised learning approach for generative-based controllable synthesis.However, GANs methods suffer from limitations for real-time applications. Thus, we propose a neural rendering approach for virtual try-on of cosmetics in real-time on mobile devices. Our approach is based on a novel inverse graphics encoder network that learns to map a single example image into the space of parameters of a computer graphics rendering engine. This model is trained using a self-supervised approach which does not require labeled training data. This method enables new applications where consumers can virtually try-on a novel, unknown cosmetic product from an inspirational reference image on social media. Finally, we propose a novel method for accelerating the digitization of new cosmetics products in virtual try-on applications. Inspired by the field of material capture, we introduced a controlled application and imaging system for cosmetics products. Furthermore, we illustrate how this novel type of cosmetics image can be used to estimate the final appearance of cosmetics on the face using a neural rendering approach. Overall, the novel methods introduced in this thesis improve cosmetics virtual try-on technologies both directly, by introducing more realistic rendering method, and indirectly, allowing novel experiences for consumers, and accelerating the creation of virtual try-on for new cosmetics products.
• Latent representations for facial images and video editing
  
  • Yao Xu
  , 2022. Learning to edit facial images and videos is one of the most popular tasks in both academia and industrial research. This thesis addresses the problem of face editing for the special case of high-resolution images and videos.In this thesis, we develop deep learning-based methods to perform facial image editing. Specifically, we explore the task using the latent representations obtained from two types of deep neural networks: autoencoder-based models and generative adversarial networks. For each type of method, we consider a specific image editing problem and propose an effective solution that outperforms the state-of-the-art.The thesis contains two parts. In part I, we explore image editing tasks via the latent space of autoencoders. We first consider the style transfer task between photos and propose an effective algorithm that is built on a pair of autoencoder-based networks. Second, we study the face age editing task for high-resolution images, using an encoder-decoder architecture. The proposed network encodes a face image to age-invariant feature representations and learns a modulation vector corresponding to a target age. Our approach allows for fine-grained age editing on high-resolution images in a single unified model.In part II, we explore the editing task via the latent space of generative adversarial models (GANs). First, we consider the problem of facial attribute disentangled editing on synthetic and real images, by proposing a latent transformation network that acts in the latent space of a pre-trained GAN model. We also proposed a video manipulation pipeline, to generalize the editing result to videos. Second, we investigate the problem of GAN inversion -- the projection of a real image to the latent space of a pretrained GAN. In particular, we propose a feed-forward encoder, which encodes a given image to a feature code and a latent code in one pass. The proposed encoder is shown to be more accurate and stable for image and video inversion, meanwhile, maintaining good editing capacities.
• Improving mediated touch interaction with multimodality
  
  • Zhang Zhuoming
  , 2022. As one of the most important non-verbal communication channels, touch is widely used for different purposes. It is a powerful force in human physical and psychological development, shaping social structures as well as communicating emotions. However, even though current information and communication technology (ICT) systems enable the use of various non-verbal languages, the support of communicating through the sense of touch is still insufficient. Inspired by the cross-modal interaction of human perception, the approach I present in this dissertation is to use multimodality to improve mediated touch interaction. Following this approach, I present three devices that provide empirical contributions to multimodal touch interaction: VisualTouch, SansTouch, and In-Flat. To understand if multimodal stimuli can improve the emotional perception of touch, I present the VisualTouch device, and quantitatively evaluate the cross-modal interaction between the visual and tactile modality. To investigate the use of different modalities in real touch communication, I present the SansTouch device, which provides empirical insights on multimodal interaction and skin-like touch generation in the context of face-to-face communication. Going one step forward in the use of multimodal stimuli in touch interaction, I present the In-Flat device, an input/output touch overlay for smartphones. In-Flat not only provides further insights on the skin-like touch generation, but also a better understanding of the role that mediated touch plays in more general contexts. In summary, this dissertation strives to bridge the gap between touch communication and HCI, by contributing to the design and understanding of multimodal stimuli in mediated touch interaction.
• Towards Informed Decision-making: Triggering Curiosity in Explanations to Non-expert Users
  
  • Bertrand Astrid
  , 2022. As AI systems gain performance, their adoption expands to areas considered critical. In finance, increasingly sophisticated recommender systems known as "robo-advisors" are democratizing online access to life insurance. Their users are confronted with a choice that will affect their quality of life, even though they may not have any financial knowledge. To address this asymmetry of information between advisor and client, the French and European legislation [4] require insurance providers to produce "clear, precise and nonmisleading" explanations to guide potential customers towards an "informed" decision. The research area of explainability thus has an important role to play in producing automatic explanations of robo-advisors that are relevant to the client. In this study conducted with the ACPR 1 , we develop a life insurance recommendation interface. We build two explanatory interfaces: a "static" one with local and global explanations, and a "stimulating" one with explanations adapted to the user's knowledge and conveyed through questions: "do you know why X" [2]? We draw on the expertise of experts in customer protection in insurance as well as on the explainability literature on over-reliance issues [1, 3]. In the coming months, we plan to conduct a large-scale study on potential clients to investigate the impact of these different forms of explanations on "informed decision-making".
• Influence of the cavity design on the differential gain and linewidth enhancement factor of a QD comb laser
  
  • Renaud Thibaut
  • Huang Heming
  • Kurczveil Géza
  • Beausoleil Raymond
  • Liang Di
  • Grillot Frédéric
  , 2022, 12141, pp.19. This work investigates the effects of the confinement factor on the linewidth enhancement factor in hybrid silicon quantum dot comb lasers, which is a key parameter involved in frequency comb generation. Experiments are performed on two laser devices sharing the same gain material with slightly different cavity designs resulting in different confinement factors. The results highlight that a lower confinement factor leads to a smaller carrier-induced refractive index variation and a larger differential gain, together resulting in a smaller linewidth enhancement factor, which in turn translates into different sets of performance regarding the feedback applications. This paper brings novel insights on the fundamental aspects of quantum dot comb lasers and provides new guidelines of future on-chip light sources for integrated wavelength-division multiplexing applications. (10.1117/12.2624500)
  DOI : 10.1117/12.2624500
• Une multitude irrationnelle
  
  • Zayana Karim
  • Rabiet Victor
  CultureMath, ENS, 2022.
• Power-aware feature selection for optimized Analog-to-Feature converter
  
  • Back Antoine
  • Chollet Paul
  • Fercoq Olivier
  • Desgreys Patricia
  Microelectronics Journal, Elsevier, 2022. Analog-to-feature (A2F) conversion is an acquisition method thought for IoT devices in order to increase wireless sensor’s battery life. The operating principle of A2F is to perform classification tasks at sub-Nyquist rate, by extracting relevant features in the analog domain and then performing the classification step in the digital domain. We propose to use Non-Uniform Wavelet Sampling (NUWS) combined with feature selection to find and extract from the signal, a small set of relevant features for electrocardiogram (ECG) anomalies detection. A power consumption model for the A2F converter, using NUWS for features extraction, is proposed based on a CMOS 0.18 μm mixed architecture. This model, by evaluating the energy cost of each feature, allows to perform a power-aware feature selection, selecting wavelets in order to maximize classification accuracy while minimizing the energy needed for extraction. We finally demonstrate the benefits of A2F conversion showing that the energy needed can be divided by 15 compared to a classical approach performing a uniform acquisition at Nyquist rate.
• Commodity Bit-Cell Sponsored MRAM Interaction Design for Binary Neural Network
  
  • Cai Hao
  • Bian Zhongjian
  • Fan Zhonghua
  • Liu Bo
  • Naviner Lirida
  IEEE Transactions on Electron Devices, Institute of Electrical and Electronics Engineers, 2022, 69 (4), pp.1721-1726. Binary neural networks (BNNs) can transform multiply-and-accumulate (MAC) operations into XNOR and accumulation (XAC), which has been proven to greatly reduce the hardware resource consumption and improve the computing speed with small accuracy loss. Among various emerging nonvolatile memories (NVMs), spin transfer torque magnetic random access memory (STT-MRAM) shows great prospect for in-memory computing framework. In this work, a device-circuit interaction design approach is investigated with commodity MRAM bit-cell. The bit-cell array (BCA) and related peripheral circuits are minimally modified to realize in-magnetic random access memory (MRAM) one-step and unlimited-width convolution operations. The in-memory computing framework is implemented with 16-nm FinFET process and magnetic tunnel junction (MTJ) compact model. Mixed National Institute of Standards and Technology database (MNIST) handwritten digit recognition is demonstrated using this in-memory computing proposal. The recognition latency of one-step convolution is 21% improved than that of XAC convolution, whereas energy consumption of XAC is 30% lower than the one-step operations. The recognition accuracy of one-step/XAC convolution achieves 93.3%/96.5%, respectively. (10.1109/TED.2021.3134588)
  DOI : 10.1109/TED.2021.3134588