Publications

2025

• Memory Distortions Induced by Reality-Altering Media
  
  • Bonnail Elise
  , 2025. Memory, particularly autobiographical and episodic memory, is a central cognitive function, essential for constructing a sense of identity, supporting social bonds, and making decisions. However, our memory is not always reliable, as we are prone to forgetting and distortions. Memory distortions can arise from source confusion, when memories are attributed to an incorrect source, for example when imagined events are confused for real ones. This phenomenon makes us vulnerable to external suggestions that can lead to the creation of false memories.While research has focused on the impact of fabricated narratives and edited photos on memory, newer forms of media now expose us to an altered perception of reality. In this thesis I introduce the term “reality-altering media” to describe media that add virtual elements to our perception of reality, either retrospectively through altered re-representations, such as edited photos, or in real-time through co-temporal media, such as Extended Reality (XR), including Augmented Reality (AR) and Virtual Reality (VR). Co-temporal media such as XR are becoming increasingly popular and are used in various contexts. However, there is limited understanding of the phenomenon of source confusion between events experienced through XR and those from reality, as well as how this technology might be exploited to manipulate users' memories. Regarding altered re-representations, the evolution of photo-editing, driven by the advent of generative Artificial Intelligence (AI) tools, allows individuals to effortlessly edit their own photos with just a few clicks. Yet, the effects of self-photo editing on memory remain unexplored.This thesis explores the impacts of new forms of reality-altering media on memory. Specifically, I investigate how XR could be leveraged to manipulate memories in a future where such technology would be ubiquitous (RQ1), whether new forms of reality-altering media (VR and AI-edited photos) can lead to memory distortions (RQ2), and how users distinguish between real and virtual events (RQ3).To address these questions, I use a combination of speculative design and empirical lab studies. For RQ1, I conducted three speculative design workshops (n=12) involving XR and memory experts who created and discussed scenarios of memory manipulations. Through qualitative analysis, I define XR Memory Manipulations (XRMMs) and categorize them into three classes (at encoding, pre-retrieval, at retrieval), which differ in terms of technology used (AR, VR) and impact on memory (influencing quality of memories, inducing forgetting, distorting memories). These scenarios allow to anticipate potential benefits but also future risks associated with the development of new forms of immersive reality-altering media. For RQ2 and RQ3, I conducted two mixed-method lab studies. The first study (n=29) confirms the occurrence of source confusion between real and VR experiences with current technology, providing quantitative and qualitative insights into the factors influencing this confusion. The second study (n=34) evaluates the impact of self-editing photos using generative AI (adding and removing elements) compared to photo editing by an external person and no editing, showing that editing one's own photo increases the chances of forgetting elements from the original event. In both studies, I collected qualitative data to identify the cues and mechanisms participants used to distinguish between memories from virtual and real events. This analysis builds on the Source Monitoring Framework, providing insights for designing technologies that minimize risks of memory distortions.From these results, I discuss the need for ethical considerations around the loss of perceptual integrity (the trust that what we perceive and remember accurately reflects reality), to ensure a responsible use of reality-altering media.
• Analog Neural Networks Trained by Equilibrium Propagation
  
  • Kiraz Zulal
  , 2025. Contribution to “Visions of Research, PhD-Driven Breakthroughs - Harnessing AI: PhD Research to Advance and Adapt Artificial Intelligence” Title: Analog Neural Networks Trained by Equilibrium Propagation This contribution is published in the Hi! PARIS white book: Hi! PARIS (2025). Visions of Research: PhD-Driven Breakthroughs. ● Full white book: https://www.hi-paris.fr/visions-of-research-phd-driven-breakthroughs/ ● Direct link to my contribution (page 39): https://heyzine.com/flip-book/0c61e8e85b.html#page/38
• Piecewise Constant Spectral Graph Neural Network
  
  • Martirosyan Vahan
  • Giraldo Jhony H
  • Malliaros Fragkiskos D
  Transactions on Machine Learning Research Journal, [Amherst Massachusetts]: OpenReview.net, 2022, 2025. Graph Neural Networks (GNNs) have achieved significant success across various domains by leveraging graph structures in data. Existing spectral GNNs, which use low-degree polynomial filters to capture graph spectral properties, may not fully identify the graph's spectral characteristics because of the polynomial's small degree. However, increasing the polynomial degree is computationally expensive and, beyond certain thresholds, leads to performance plateaus or degradation. In this paper, we introduce the Piecewise Constant Spectral Graph Neural Network (PieCoN) to address these challenges. PieCoN combines constant spectral filters with polynomial filters to provide a more flexible way to leverage the graph structure. By adaptively partitioning the spectrum into intervals, our approach increases the range of spectral properties that can be effectively learned. Experiments on nine benchmark datasets, including both homophilic and heterophilic graphs, demonstrate that PieCoN is particularly effective on heterophilic datasets, highlighting its potential for a wide range of applications. The implementation of PieCoN is available at https://github.com/vmart20/PieCoN.
• F-StrIPE: Fast Structure-Informed Positional Encoding for Symbolic Music Generation
  
  • Agarwal Manvi
  • Wang Changhong
  • Richard Gael
  , 2025. While music remains a challenging domain for generative models like Transformers, recent progress has been made by exploiting suitable musically-informed priors. One technique to leverage information about musical structure in Transformers is inserting such knowledge into the positional encoding (PE) module. However, Transformers carry a quadratic cost in sequence length. In this paper, we propose F-StrIPE, a structure-informed PE scheme that works in linear complexity. Using existing kernel approximation techniques based on random features, we show that F-StrIPE is a generalization of Stochastic Positional Encoding (SPE). We illustrate the empirical merits of F-StrIPE using melody harmonization for symbolic music.
• Sécurité des canaux cachés : de la théorie à la pratique
  
  • Guilley Sylvain
  • Rioul Olivier
  , 2025. La sécurité des implémentations cryptographiques s'appuie sur la confidentialité des clés secrètes et privées, ainsi que sur l'intégrité des clés publiques. Ainsi, des protocoles de confiance peuvent être construits, pour établir des applications de sécurité, tels que l'authentification (en terme d'identité) ou l'attestation (vis-à-vis des firmware). Maintenant, les attaques sur les canaux cachés permettent de remonter aux clés sans cryptanalyse, simplement en espionnant une émanation fortuite de l'équipement considéré. Or de nombreux travaux académiques ont été réalisés pour caractériser de telles attaques. Par exemple, le livre "Mathematical Foundations for Side-Channel Analysis of Cryptographic Systems", publié en 2024 par les auteurs, démontre différents chemins d'attaque. La communauté scientifique a aussi proposé des bornes de sécurité, c'est-à-dire une limite inférieure sur le nombre de traces espionnées pour retirer suffisamment d'information permettant de remonter à la clé. Maintenant, ces bornes s'appuient sur des notions et des métriques de théorie de l'information. Or l'estimation de ces grandeurs n'est pas triviale. Dans cette exposé de conférence, nous démontrons comment relier la complexité des attaques à un nombre de mesure, i.e., à un effort tel que décrit par les Critères Communs (CC). Cette mise en pratique est utile par les laboratoires réalisant les certifications. Notre présentation vise à rendre didactique et application des travaux publiés à des conférences de cryptographie de haut niveau. Une traduction dans le langage de la normalisation (notamment le standard ISO/IEC 17825) sera proposée.
• Multi-Sensor Data Fusion for Enhanced Detection of Laser Fault Injection Attacks in Cryptographic Hardware: Practical Results
  
  • Ebrahimabadi Mohammad
  • Viera Raphael
  • Guilley Sylvain
  • Danger Jean-Luc
  • Dutertre Jean-Max
  • Karimi Naghmeh
  , 2025, pp.1 - 2. Though considered secure the cryptographic hardware can be compromised by fault injection attack, especially laser illumination due to its precision in targeting specific areas and its fine temporal control. To address this threat, this paper presents a low-cost detection scheme that utilizes Time-to-Digital Converters (TDCs) to sense the IR drops induced by laser illumination. To achieve a high detection rate while minimizing false alarms, the proposed approach incorporates multiple sensors, with as few as two sensors demonstrated in the study. The effectiveness of the scheme is validated using a real laser setup to illuminate a targeted AES module implemented on an AMD/Xilinx Artix-7 FPGA. (10.23919/date64628.2025.10992902)
  DOI : 10.23919/date64628.2025.10992902
• A VERSATILE FRAMEWORK FOR EVALUATING SINGLE NEURON TRACKING IN BEHAVING ANIMALS
  
  • Reme Raphael
  • Newson Alasdair
  • Angelini Elsa
  • Olivo-Marin Jean-Christophe
  • Lagache Thibault
  , 2025, pp.1-5. Accurately tracking neuronal activity in behaving animals, such as C. elegans, Drosophila, Zebrafish and Hydra vulgaris, presents significant challenges due to complex motions and background noise. While recent advancements in genetic engineering and fluorescence microscopy have improved imaging capabilities, existing tracking algorithms have struggled to perform effectively in these dynamic environments, often relying on simpler motion models that do not replicate behavioral conditions. In particular, the lack of annotated datasets for these motions limits the evaluation and improvement of such tracking algorithms. To address this, we developed a novel simulator that generates synthetic tracking data for particles on a deformable background, closely mimicking live animal recordings. This simulator produces annotated videos that reflects the intricate movements seen in Hydra Vulgaris, allowing for a robust evaluation of four tracking algorithms. The findings highlight the current limitations of these methods in challenging scenarios, paving the way for improved cell tracking techniques in dynamic biological systems. (10.1109/ISBI60581.2025.10981111)
  DOI : 10.1109/ISBI60581.2025.10981111
• ASML-REG: Automated Machine Learning for Data Stream Regression
  
  • Verma Nilesh
  • Bifet Albert
  • Pfahringer Bernhard
  • Bahri Maroua
  , 2025, pp.440-447. Online learning scenarios present a significant challenge for Au-toML techniques due to the dynamic nature of data distributions, where the optimal model and configuration may change over time. While most research in machine learning for data streams has primarily focused on classification algorithms, regression methods have received significantly less attention. To address this gap, we propose ASML-REG, an Automated Streaming Machine Learning framework designed specifically for regression tasks on data streams. ASML-REG continuously explores a vast and diverse space of pipeline configurations, adapting to evolving data by focusing on the current best design, performing adaptive random searches in promising areas, and maintaining an ensemble of top-performing pipelines. Our experiments with real and synthetic datasets demonstrate that ASML-REG significantly outperforms current state-ofthe-art data stream regression algorithms. (10.1145/3672608.3707742)
  DOI : 10.1145/3672608.3707742
• Assessment of the representativeness of numerical subjects for the WBAN indoor channel modeling
  
  • Youssef Badre
  • Roblin Christophe
  , 2025, pp.1-5. <div><p>Wireless Body Area Networks peculiarities can be summed up in two points: firstly, the high number of sources of variability, i.e. the subject, the antennas, the frequency, the environment, etc.; secondly the strong electromagnetic disturbance of the human body (due to its composition, around 70% of water). Concerning the subject, there are two aspects that can be considered: the morphology and the dynamic. Here, we are interested only on the first one. When this source of variability is considered in the channel modeling, the problem is the difficulty of obtaining a representative statistical sample. One solution would be to use numerical subject models, which would give us a much wider range of subjects than is currently proposed in the literature. Then the purpose of this article is to assess the representativeness of this approach in the context of WBAN channel modeling, by comparing key channel parameters (the mean Path Loss of the on-body cluster and the surrounded antenna pattern) obtained from simulations results for different numerical subjects and experimentations performed with a whole body phantom. In this study, the scenario based approach is preferred and simulations and experimentations are performed in the 1 st UWB band ([3.1, 4.8] GHz). The results obtained are satisfactory, and validate the use of this approach in view of the assumptions exposed for modeling the WBAN channel.</p></div> (10.23919/EuCAP63536.2025.10999677)
  DOI : 10.23919/EuCAP63536.2025.10999677
• Impact of Rough Building Material Surfaces on Reflection Coefficients from 5 GHz to 260 GHz
  
  • Conrat Jean-Marc
  • Cousin Jean-Christophe
  • Begaud Xavier
  , 2025. When an electromagnetic wave is reflected by a smooth and planar surface, its amplitude is decreased by the Fresnel reflection coefficients depending on the material permittivity, incidence angle and polarization. Recent work showed that the permittivity does not depend on the frequency from 2 to 260 GHz implying that the reflection gain is constant from 2 to 260 GHz for a given incidence angle and polarization. But Fresnel coefficients are only valid for materials with a smooth surface which is not always the case for building materials. Materials such as concrete or patterned glass may be rough or have a relief surface. This paper analyzes the reflection gain deviation from the Fresnel coefficients when material surfaces are not smooth and proposed a simplified model based on the Rayleigh-Rice or Beckmann-Kirchhoff theoretical approaches. The investigated frequencies ranged from 5 to 260 GHz.
• An Efficient Compact Dual-Band Metasurface RF Energy Harvester
  
  • Sharifi Raziyeh
  • Lepage Anne Claire
  • Niotaki Kyriaki
  • Begaud Xavier
  , 2025. This paper presents a dual-band metasurface-based RF energy harvester to efficiently collect the ambient radio frequency energy at 2.45 GHz and 5.2 GHz (Wi-Fi bands). The methodology to design the dual-band metasurface by employing the EM simulation software CST Studio Suite is presented. The dual-band metasurface is designed based on a proposed single-band metasurface harvester, with both unit cell designs achieving high absorption characteristics at their operating frequencies. An array of 4×5 unit cells for both the single-band and dual-band designs were fabricated to validate the concept. The first results for the single band design demonstrate a good agreement between the simulated and the measured results.
• Millimeter Wave Rectifiers: A Review of Technology and Performance
  
  • Wang Yibo
  • Niotaki Kyriaki
  • Lepage Anne Claire
  • Begaud Xavier
  , 2025. Rectifiers operating at the millimeter wave (mmWave) frequencies have recently attracted interest for a wide range of applications, including wireless power transmission and energy harvesting. This work presents an overview of the rectifiers operating in the mmWave band, and more specifically from 24 GHz to 94 GHz. It also highlights the main mmWave rectifier's design challenges along with some novel techniques used to optimize their performance.
• Wavelength and Code Orthogonality Based Distributed Acoustic Sensing over a Passive Optical Network
  
  • Kumar Choudhury Pallab
  • Awwad Élie
  , 2025. We present a wavelength and code orthogonality based DFOS enabling simultaneous sensing of all paths of a splitter-based passive optical network. Strain sensitivity of 80nmpp is measured with no penalty on coexisting 10Gb/s downstream transmission.
• Reconfigurable intelligent surfaces (RIS) using NOMA with thermal energy harvesting
  
  • Boujemaa Hatem
  • Alhussein Musaed
  • Rekaya Ghaya
  Signal, Image and Video Processing, Springer Verlag, 2025, 19 (5), pp.419:1-419:8. The integration of Reconfigurable Intelligent Surfaces (RIS) with Non-Orthogonal Multiple Access (NOMA) and thermal energy harvesting presents a novel approach to enhancing wireless communication networks. RIS technology optimizes signal propagation and improves network efficiency through programmable surface elements, while NOMA increases spectral efficiency by allowing multiple users to share the same frequency resource. When combined with thermal energy harvesting, which captures ambient heat and converts it into electrical power, this integration offers a sustainable solution to power the RIS infrastructure. This paper explores the synergistic benefits of RIS using NOMA with thermal energy harvesting, examining its impact on network performance, energy efficiency, and sustainability. Through a review of recent advancements and research, we discuss how this combined approach can address key challenges in modern wireless communications and contribute to the development of greener, more efficient networks. (10.1007/s11760-025-03996-x)
  DOI : 10.1007/s11760-025-03996-x
• Software Identification for Cybersecurity: Survey and Recommendations for Regulators
  
  • Barais Olivier
  • Cosmo Roberto Di
  • Mé Ludovic
  • Zacchiroli Stefano
  • Zendra Olivier
  , 2025.
• Statistical wave field theory: Special polyhedra
  
  • Badeau Roland
  Journal of the Acoustical Society of America, Acoustical Society of America, 2025, 157 (3), pp.2263-2278. The statistical wave field theory establishes mathematically the statistical laws of the solutions to the wave equation in a bounded volume. It provides the closed-form expression of the power distribution and the correlations of the wave field jointly over time, frequency, and space, in terms of the geometry and the specific admittance of the boundary surface. In a recent paper, we presented a mathematical approach to this theory based on the Sturm-Liouville theory and the theory of dynamical billiards. We focused on mixing billiards that generate an isotropic wave field, and we retrieved the well-known statistical properties of reverberation in room acoustics. In the present paper, we introduce a simpler geometric approach, dedicated to a particular class of non-ergodic billiards. Though limited to only a few polyhedra, this approach offers a precious insight into various aspects of the theory, including the first examples of anisotropic wave fields, whose statistical properties are related to mathematical crystallography. We also show that the formulas that we obtain in this anisotropic case are closely related to those of the mixing case, albeit based on a different mathematical approach. (10.1121/10.0036254)
  DOI : 10.1121/10.0036254
• Multispectral Texture Synthesis using RGB Convolutional Neural Networks
  
  • Ollivier Sélim
  • Gousseau Yann
  • Lefebvre Sidonie
  IEEE Transactions on Geoscience and Remote Sensing, Institute of Electrical and Electronics Engineers, 2025, 63, pp.5402914. State-of-the-art RGB texture synthesis algorithms rely on style distances that are computed through statistics of deep features. These deep features are extracted by classification neural networks that have been trained on large datasets of RGB images. Extending such synthesis methods to multispectral images is not straightforward, since the pre-trained networks are designed for and have been trained on RGB images. In this work, we propose two solutions to extend these methods to multispectral imaging. Neither of them require additional training of the neural network from which the second order neural statistics are extracted. The first one consists in optimizing over batches of random triplets of spectral bands throughout training. The second one projects multispectral pixels onto a 3 dimensional space. We further explore the benefit of a color transfer operation upstream of the projection to avoid the potentially abnormal color distributions induced by the projection. Our experiments compare the performances of the various methods through different metrics. We demonstrate that they can be used to perform exemplar-based texture synthesis, achieve good visual quality and comes close to state-of-the art methods on RGB bands. Code is available at \url{https://github.com/selim2483/multispectral-textureCNN} (10.1109/TGRS.2025.3554931)
  DOI : 10.1109/TGRS.2025.3554931
• Using a Probabilistic Database in an Image Retrieval Application
  
  • Yunus Fajrian
  • Karmakar Pratik
  • Senellart Pierre
  • Abdessalem Talel
  • Bressan Stéphane
  , 2025. ProvSQL is a PostgreSQL extension implementing provenance management and probabilistic database features. ProvSQL seamlessly extends relational database functionality to support the storage, tracking through derivations and transformations, and querying of metadata that explain and qualify the data and query results. In this demonstration, ProvSQL is used to implement a content-based image retrieval system. A deep learning object detection model identifies objects of selected classes located within the images of a large-scale image data set. The uncertainty associated with object detection is recorded. ProvSQL's provenance model incorporates this uncertainty into the retrieval process, thus facilitating the generation of accurate and reliable results and allowing for decision-making in scenarios with incomplete or uncertain information. The demonstration illustrates how ProvSQL handles query processing, uncertainty tracking, and probability computation. It highlights the utility of a probabilistic database for applications dealing with uncertain data, compared to traditional threshold-based approaches.
• Role of deep learning into three-dimensional reconstruction for breast tomosynthesis
  
  • Quillent Arnaud
  , 2025. Digital breast tomosynthesis (DBT) is an X-ray imaging technique introduced in the 2010s that provides a three-dimensional reconstruction of tissues. Commercial systems rely on acquiring images from a limited number of views sampled every 1 to 3 degrees, evenly distributed over a scanning cone of 15 to 40°. The acquired projections are then processed by a reconstruction algorithm that generates the volume reviewed by the radiologist. Numerous analytical and neural network reconstruction algorithms exist. However, the narrow opening of the acquisition cone limits the quality of the reconstructed volume, particularly its resolution in the thickness of the breast. Orthogonal slices thus become unreadable with current techniques :the gradual extinction of artefacts from an object outside its focal planes disrupts the reading of objects located in neighbouring planes. Thus, there is currently no commercial solution effectively addressing limited-angle artefacts in breast tomosynthesis.The use of deep learning faces several challenges. Indeed, paired data in real conditions that would allow supervised learning are absent,and volumes from other breast imaging modalities cannot be used directly . Questions also arise regarding the properties of the obtained methods. On the one hand, the reconstructed volumes are not always consistent to the measurements, a property well mastered by conventional methods. Moreover, the uncertainty of the image reconstructed by deep learning is almost never explicit. In this thesis, we first propose a post-processing approach for conventional reconstruction. To develop our method, we create a synthetic database consisting of digital phantoms. We simulate their X-ray projections and reconstruct 3D volumes using conventional methods,thus creating a dataset that we use to train a convolutional neural network in a supervised manner. We demonstrate that the proposed strategy significantly improves the quality of orthogonal planes and is therefore promising to address the problem of reconstruction in breast tomosynthesis. Secondly, we seek to obtain an estimate of the reliability of the volumes predicted by the neural network. We adopt a Bayesian perspective, differentiating between aleatory and epistemic uncertainties. We model the first term using a Laplace distribution and the second by approximating the posterior predictive distribution, then compare the results obtained with a Monte Carlo Dropout method and a deep ensemble. We show that the computed uncertainty is a good approximation of the actual error and reuse to minimise a data consistency term. Finally, we improve the realism of the synthetic volumes by adapting images from another imaging modality that does not present the same artefacts as tomosynthesis. After appropriate denoising, we segment these images and simulate their compression to replicate the conditions of the medical examination. We redesign the neural network to allow 3D learning that considers information from the three anatomical directions, and solve the convergence problems related to the calculation of the uncertainty terms. We then impose the fidelity constraint to the projections by limiting the contribution of the neural model to the kernel space of the projection operator. Finally, we evaluate our methods on volumes with specific geometric properties as well as on clinical images, and highlight the benefits and limitations of our deep learning reconstruction approaches for breast tomosynthesis.
• Adaptive Passive Beamforming in RIS-Aided Communications with Q-Learning
  
  • Chêne Thomas
  • Bounhar Oumaïma
  • Othman Ghaya Rekaya-Ben
  • Damen Oussama
  , 2025, pp.1-6. Reconfigurable Intelligent Surfaces (RIS) appear as a promising solution to combat wireless channel fading and interferences. However, the elements of the RIS need to be properly oriented to boost the data transmission rate. In this work, we propose a new strategy to adaptively configure the RIS without Channel State Information (CSI). Our goal is to minimize the number of RIS configurations to be tested to find the optimal one. We formulate the problem as a stochastic shortest path problem, and use Q-Learning to solve it. (10.1109/WCNC61545.2025.10978715)
  DOI : 10.1109/WCNC61545.2025.10978715
• Stochastic Geometry-Based MCS Adaption Analysis for Uplink Cellular Networks
  
  • Guo Xinyi
  • Liu Qiong
  • Wang Shanshan
  • You Li
  , 2025, pp.1-6. The link adaptation plays a crucial role in the fifth generation (5G) and future wireless networks, where adaptive modulation and coding (AMC) is vital for significantly increasing the data transmission rate and quality of service (QoS) by adjusting the modulation and coding scheme (MCS). In this work, we investigate the stochastic geometry-based MCS adaption for the uplink cellular networks with Poisson distributed base stations (BS) and user equipments (UE). We first define the conditional received rate by quantizing the channel quality, i.e., signal to interference ratio (SIR), using the sets of thresholds. Basically, higher SIR indicates better channel condition and applys higher order of modulation scheme, which leads to higher received rate. We then derive the framework of meta distribution on the conditional received rate, the spatially-average spectral efficiency (SE), and the variance of the SE. In addition, beta approximation and several bounds are presented to simplify the calculation of meta distribution. We validate the proposed framework by numerical simulations under different system parameters. (10.1109/WCNC61545.2025.10978118)
  DOI : 10.1109/WCNC61545.2025.10978118
• Transformer-Based Packet Scheduling under Strict Delay and Buffer Constraints
  
  • Nérondat Sylvain
  • Leturc Xavier
  • Le Martret Christophe
  • Ciblat Philippe
  , 2025, pp.1-6. <div><p>This paper presents a packet scheduler for managing multiple links with varying channel capacities, where each link carries multiple data flows with finite buffers and strict delay constraints. Packet loss can result from buffer overflow or delay violations. We propose a deep reinforcement learning scheduler based on an encoder-only transformer (EOT) architecture, capable of handling a variable number of links without dedicated training. Using deep Q-learning, the scheduler minimizes the packet loss rate (PLR). Simulations show that our approach outperforms a state-of-the-art fully connected (FC) scheduler, delivering better performance under diverse configurations of links, packet arrival rates, and channel capacities.</p></div> (10.1109/WCNC61545.2025.10978237)
  DOI : 10.1109/WCNC61545.2025.10978237
• Towards semantically enriched embeddings for knowledge graph completion
  
  • Alam Mehwish
  • van Harmelen Frank
  • Acosta Maribel
  Neurosymbolic Artificial Intelligence, 2025, 1, pp.1-16. Embedding based Knowledge Graph (KG) completion has gained much attention over the past few years. Most of the current algorithms consider a KG as a multidirectional labeled graph and lack the ability to capture the semantics underlying the schematic information. This position paper revises the state of the art and discusses several variations of the existing algorithms for KG completion, which are discussed progressively based on the level of expressivity of the semantics utilized. The paper begins with analysing various KG completion algorithms considering only factual information such as transductive and inductive link prediction and entity type prediction algorithms. It then revises the algorithms utilizing Large Language Models as background knowledge. Afterwards, it discusses the algorithms progressively utilizing semantic information such as class hierarchy information within the KGs and semantics represented in different description logic axioms. The paper concludes with a critical reflection on the current state of work in the community, where we argue that the aspects of semantics, rigorous evaluation protocols, and bias against external sources have not been sufficiently addressed in the literature, which hampers a more thorough understanding of advantages and limitations of existing approaches. Lastly, we provide recommendations for future directions. (10.3233/NAI-240731)
  DOI : 10.3233/NAI-240731
• Noise and dynamics of hybrid plasmonic semiconductor lasers for InP-based quantum integrated optical communications
  
  • Cui Di
  , 2025. Integrating plasmonics into photonic integrated circuits (PICs) enables the development of ultra-compact devices while maintaining high performance. Plasmonic structures enhance light-matter interactions, leading to advanced functionalities such as high-sensitivity sensing and nonlinear optics. By offering subwavelength control and faster modulation speeds, they overcome the limitations of silicon photonics in light confinement and modulation efficiency.This thesis presents a comprehensive experimental study of hybrid plasmonic semiconductor lasers, emphasizing their potential applications in plasmonic PICs. Key characteristics such as linewidth enhancement, relaxation oscillations, and relative intensity noise (RIN) are examined, along with nonlinear dynamics induced by external perturbations like optical feedback and optical injection. Experimental results reveal that hybrid plasmonic lasers exhibit greater resistance to optical feedback compared to conventional quantum well (QW) lasers, demonstrating reduced dynamic instabilities and superior feedback tolerance. This suggests they could be integrated into photonic circuits without requiring optical isolators.Optical injection experiments show that hybrid plasmonic lasers deviate from the conventional chaotic behavior observed in semiconductor lasers, instead exhibiting sustained feedback-induced oscillations. Sensitivity tests under optical feedback further confirm their reluctance to transition into chaotic states, even under destabilizing conditions. These findings highlight the significant role of surface plasmon polariton (SPP) interactions in enhancing nonlinear effects. The resonance properties of the metal coating and underdamped relaxation oscillations in the surface plasmon waveguide contribute to the laser's unique nonlinear behavior. With their strong resistance to optical feedback, absence of chaotic oscillations, and distinct dynamic properties, hybrid plasmonic lasers emerge as promising candidates for large-scale CMOS-compatible photonic integration, particularly in eliminating the need for bulky optical isolators in PICs.
• Self-Supervised Learning of Audio Representations for Musical Applications
  
  • Riou Alain
  , 2025. The goal of this PhD is to propose new paradigms for training deep neural networks to produce audio representations that are suited for diverse musical applications, with a special focus on self-supervised learning (SSL) approaches.In our first work, we focus on the task of monophonic pitch estimation. We solve the task by training a Siamese architecture on pairs of pitch-shifted Constant-Q Transform (CQT) frames. The network is trained with a novel class-based equivariance criterion, using the (known) shift between the frames as a supervision signal. This strategy enables our model to directly predict pitch distributions in a fully self-supervised way, without access to any annotated data. Despite being extremely lightweight (30k parameters), our model significantly outperforms previous self-supervised baselines and is on par with supervised ones.Then, we investigate the paradigm of Joint-Embedding Predictive Architectures (JEPA) for music-related applications. From given context/target pairs, an encoder and a predictor are jointly trained to produce latent representations of the pair and predict the target representations from the context ones. Opposite to contrastive learning, JEPAs do not require any negative samples and can learn a richer latent space thanks to the predictor. By using the different sources of music tracks as context/target pairs, we show that these architectures can capture both local and global musical features, making them useful for a variety of tasks such as compatibility estimation, musical stem retrieval, track alignment, genre classification, auto-tagging and beat-tracking.