Papers

Large-scale MIMO detection remains challenging because exact or near-maximum-likelihood search is difficult to scale, while available quantum resources are insufficient for directly solving full-size detection instances by QAOA. This paper therefore proposes a Block-QAOA-Aware MIMO Detector (BQA-MD), whose primary purpose is to reorganize the detection chain so that it becomes compatible with limited-qubit local quantum subproblems. Specifically, BQA-MD combines block-QAOA-aware preprocessing in the QR domain, a standards-consistent blockwise 5G NR Gray-HUBO interface, an MMSE-induced dynamic regularized blockwise objective, and K-best candidate propagation. Within this framework, fixed-size block construction gives every local subproblem a uniform circuit width and parameter dimension, which in turn enables parameter-transfer QAOA as a practical realization strategy for structurally matched local subproblems. Experiments are conducted on a 16x16 Rayleigh MIMO system with 16QAM using classical simulation of the quantum subroutine. The results show that the regularized blockwise detector improves upon its unregularized counterpart, validating the adopted blockwise objective and the block-QAOA-aware design rationale. They also show that the parameter-transfer QAOA detector nearly matches the regularized blockwise exhaustive reference and clearly outperforms direct-training QAOA in BER, thereby supporting parameter reuse as the preferred QAOA realization strategy within the proposed framework. In the tested setting, MMSE remains slightly better in the low-SNR region, whereas the parameter-transfer QAOA detector becomes highly competitive from the medium-SNR regime onward.

Audio anti-spoofing systems are typically trained to assign one authenticity label to an entire speech utterance. This formulation becomes under-specified for transformations where the underlying speaker identity and linguistic content remain unchanged. We study this problem using benign, authenticity-preserving speech transformations, including voice quality conversion and speech restoration, applied to both bona fide and spoofed speech. Instead of treating all processed audio as spoofed, we factorise labels into source authenticity and processed status. Across SSL representations and DF-Arena fine-tuning experiments, we find that utterance processing status can transfer more reliably than source attribution: detectors can often identify that speech has been processed, while still confusing processed bona fide and processed spoofed speech. These results suggest that audio deepfake defences must move beyond the binary spoofed/authentic paradigm. Robust detection requires granular reporting on source authenticity, processing status, and precise processing localisation.

Diffusion and flow matching TTS faces a tension between discrete temporal structure and continuous spectral modeling. Two-stage models diffuse on fixed alignments, often collapsing to mean prosody; single-stage models avoid explicit durations but suffer alignment instability. We propose a jump-diffusion framework where discrete jumps model temporal structure and continuous diffusion refines spectral content within one process. Even in its one-shot degenerate form, our framework achieves 3.37% WER vs. 4.38% for Grad-TTS with improved UTMOSv2 on LJSpeech. The full iterative UDD variant further enables adaptive prosody, autonomously inserting natural pauses in out-of-distribution slow speech rather than stretching uniformly. Audio samples are available at https://anonymousinterpseech.github.io/TTS_Demo/.

The increasing penetration of renewable generation and the growing variability of electrified demand introduce substantial operational uncertainty to modern power systems. Topology reconfiguration is widely recognized as an effective and economical means to enhance grid resilience. Due to the coexistence of AC power-flow constraints and discrete switching decisions, topology reconfiguration in large-scale systems leads to a highly nonlinear and nonconvex optimization problem, making traditional methods computationally prohibitive. Consequently, several studies have explored reinforcement learning-based approaches to improve scalability and operational efficiency. However, its practical implementation is challenged by the high-dimensional combinatorial action space and the need to ensure safety during learning-based decision-making. To address these challenges, this paper presents a safe and intelligent topology control framework that integrates Large Language Models (LLMs) with a Safety Soft Actor-Critic (Safety-SAC) architecture. Operational voltage and thermal limits are reformulated into smooth safety-cost signals, enabling risk-aware policy optimization within a constrained Markov decision process. A knowledge-based Safety-LLM module is further introduced to refine unsafe or suboptimal transitions through domain knowledge and state-informed reasoning, thus guiding the learning agent toward safer and more effective switching actions. Experiments on the IEEE 36-bus and 118-bus Grid2Op benchmarks show that the proposed method consistently improves reward, survival time, and safety metrics, achieving higher reward, longer survival, and lower safety cost compared with SAC, ACE, and their safety-enhanced variants. These results demonstrate the potential of combining LLM-based reasoning with safe reinforcement learning to achieve scalable and reliable grid topology control.

Integrated Sensing and Communications (ISAC) has emerged as a key enabler for sixth generation (6G) wireless systems by jointly supporting data transmission and environmental awareness within a unified framework. However, communication and sensing functionalities impose inherently conflicting performance requirements, particularly in heterogeneous networks where users may demand sensing only, communication only, or joint services. Selecting a waveform that satisfies diverse service demands therefore becomes a challenging multi objective decision problem. In this paper, a multi objective learning approach for adaptive waveform selection in ISAC systems is proposed. A simulation driven evaluation framework is developed to assess multiple waveform candidates across communication, sensing, and joint performance metrics. Instead of enforcing scalar utility aggregation, waveform performance is represented in a multi dimensional objective space where Pareto optimal candidates are identified for each scenario. A dataset is generated by varying user demand distributions and channel conditions, and multi-label targets are constructed based on Pareto dominance. Machine learning models are trained to learn the mapping between network conditions and Pareto optimal waveform sets, enabling fast waveform selection under dynamic network states. Simulation results demonstrate that the proposed framework effectively adapts waveform selection to heterogeneous service requirements while preserving sensing communication trade offs, providing a forward-looking perspective for 6G and beyond ISAC deployments.

The widespread adoption of electronic health records (EHRs) enables the acquisition of heterogeneous clinical data, spanning lab tests, vital signs, medications, and procedures, which offer transformative potential for artificial intelligence in healthcare. Although traditional modelling approaches have typically relied on multivariate time series, they often struggle to accommodate the inherent sparsity and irregularity of real-world clinical workflows. Consequently, research has shifted toward event stream representation, which treats patient records as continuous sequences, thereby preserving the precise temporal structure of the patient journey. However, the existing literature remains fragmented, characterised by inconsistent definitions, disparate modelling architectures, and varying training protocols. To address these gaps, this review establishes a unified definition of EHR event streams and introduces a novel taxonomy that categorises models based on their handling of event time, type, and value. We systematically review training strategies, ranging from supervised learning to self-supervised methods, and provide a comprehensive discussion of applications across clinical scenarios. Finally, we identify open critical challenges and future directions, with the aim of clarifying the current landscape and guiding the development of next-generation healthcare models.

Distributed multiple-input multiple-output (MIMO) architectures enable large-scale integrated sensing and communication (ISAC) by providing high spatial resolution and robustness through spatial diversity. However, practical phase-coherent sensing is challenged by phase synchronization errors and modeling mismatch caused by grid discretization. Existing over-the-air (OTA) synchronization methods typically treat synchronization and sensing tasks separately, which may lead to inaccurate phase alignment when multipath components are used for imaging. In this paper, we propose a non-line-of-sight (NLOS)-aided joint OTA synchronization and off-grid imaging framework for distributed MIMO ISAC systems. First, a line-of-sight (LOS)-assisted coarse synchronization is performed to establish initial phase coherence across distributed links. Subsequently, an iterative refinement stage exploits reconstructed NLOS components obtained from imaging results. By modeling off-grid effects via a first-order Taylor expansion, we transform measurements with nonlinear off-grid offset into an augmented linear model with jointly sparse reflectivity and off-set variables. The imaging problem is reformulated as a structured sparse recovery task and solved using a tailored off-grid approximate message passing (OG-AMP) algorithm. The imaging and synchronization modules are coupled within a closed-loop alternative optimization framework, where improved imaging enables more accurate phase refinement, and vice versa. Numerical results show that the proposed framework achieves accurate synchronization and imaging under phase errors. Compared with conventional approaches, it shows superior robustness and accuracy.

We investigate the discrete-time stochastic linear quadratic control problem for a population of cooperative agents under the hard equality constraint on total control inputs, motivated by demand response in renewable energy systems. We establish the optimal solution that respects hard equality constraints for systems with additive noise in the dynamics. The optimal control law is derived using dynamic programming and Karush-Kuhn-Tucker (KKT) conditions, and the resulting control solution depends on a discrete-time Riccati-like recursive equation. Application examples of coordinating the charging of a network of residential batteries to absorb excess solar power generation are demonstrated, and the proposed control is shown to achieve exact power tracking while considering individual State-of-Charge (SoC) objectives

Anatomical landmark segmentation serves as a critical initial step for robust multimodal registration during computer-assisted interventions. Current approaches predominantly rely on deep learning, which often necessitates the extensive manual generation of annotated datasets. In this paper, we present a novel strategy for creating large annotated datasets using a statistical shape model (SSM) based on a mean shape that is manually labeled only once. We demonstrate the method's efficacy through its application to deep-learning-based anatomical landmark segmentation, specifically targeting the detection of the anterior ridge and the falciform ligament in 3D liver shapes. A specialized deep learning network was trained with 8,800 annotated liver shapes generated by the SSM. The network's performance was evaluated on 500 unseen synthetic SSM shapes, yielding a mean Intersection over Union of 91.4% (87.4% for the anterior ridge and 87.6% for the falciform ligament). Subsequently, the network was applied to clinical patient liver shapes, with qualitative evaluation indicating promising results and highlighting the generalizability of the proposed approach. Our findings suggest that the SSM-based data generation approach alleviates the labor-intensive process of manual labeling while enabling the creation of large annotated training datasets for machine learning. Although our study focuses on liver anatomy, the proposed methodology holds potential for a broad range of applications where annotated training datasets play a pivotal role in developing accurate deep-learning models.

Echocardiography is widely used for assessing cardiac function, where clinically meaningful parameters such as left-ventricular ejection fraction (EF) play a central role in diagnosis and management. Generative models capable of synthesising realistic echocardiogram videos with explicit control over such parameters are valuable for data augmentation, counterfactual analysis, and specialist training. However, existing approaches typically rely on computationally expensive multi-step sampling and aggressive temporal normalisation, limiting efficiency and applicability to heterogeneous real-world data. We introduce EchoLVFM, a one-step latent video flow-matching framework for controllable echocardiogram generation. Operating in the latent space, EchoLVFM synthesises temporally coherent videos in a single inference step, achieving a $\mathbf{\sim 50\times}$ improvement in sampling efficiency compared to multi-step flow baselines while maintaining visual fidelity. The model supports global conditioning on clinical variables, demonstrated through precise control of EF, and enables reconstruction and counterfactual generation from partially observed sequences. A masked conditioning strategy further removes fixed-length constraints, allowing shorter sequences to be retained rather than discarded. We evaluate EchoLVFM on the CAMUS dataset under challenging single-frame conditioning. Quantitative and qualitative results demonstrate competitive video quality, strong EF adherence, and 57.9% discrimination accuracy by expert clinicians which is close to chance. These findings indicate that efficient, one-step flow matching can enable practical, controllable echocardiogram video synthesis without sacrificing fidelity. Code available at: https://github.com/EngEmmanuel/EchoLVFM

This paper presents a switched model reference admittance control framework to achieve safe and compliant human-robot collaboration through reference trajectory shaping. The proposed method generates variable admittance parameters according to task compliance and task-space safety requirements. Additionally, a disturbance bound is incorporated to enhance robustness against disturbances. Safety guarantees are explicitly established by integrating invariance control, ensuring that the reference trajectory remains within the admissible region. Stability of the switched system is analyzed using a common quadratic Lyapunov function, which confirms asymptotic convergence of the tracking error. The effectiveness of the approach is demonstrated through simulations on a two link manipulator and comparisons with existing methods are also presented. Furthermore, real time implementation on a single link manipulator validates the practical feasibility of the controller, highlighting its ability to achieve both compliance and safety in physical interaction scenarios.

In existing Audio-Visual Speech Enhancement (AVSE) methods, objectives such as Scale-Invariant Signal-to-Noise Ratio (SI-SNR) and Mean Squared Error (MSE) are widely used; however, their correlation with perceived speech quality is often suboptimal and provides limited interpretability for optimization. This work proposes a reinforcement learning-based AVSE framework with a Large Language Model (LLM)-based interpretable reward model. An audio LLM generates natural language descriptions of enhanced speech, which are converted by a sentiment analysis model into a 1-5 rating score serving as the PPO reward for fine-tuning a pretrained AVSE model. Compared with scalar metrics, LLM-generated feedback is semantically rich and explicitly describes speech quality improvements. Experiments on the AVSEC-4 dataset show that the proposed method outperforms a supervised baseline and a DNSMOS-based RL baseline in PESQ, STOI, neural quality metrics, and subjective listening tests.

Pinching-antenna systems (PASs) have been proposed as a flexible antenna technology to fulfill the stringent requirements of high data rate and large-scale equipment deployment in future wireless networks. The principle of PA involves mapping a signal over dielectric waveguides for transmission. By adjusting the positions of pinching antennas (PAs) over the waveguides, with the aim of gain enhancement for line-of-sight links and the reduction of large-scale path loss. Symbol-level precoding (SLP) is a nonlinear precoding technique, which converts multi-user interference into constructive interference via beamforming design at symbol level. In this paper, we study the combination of SLP and PAS, leveraging the advantages of PAS to further enhance the ability of SLP to convert constructive interference. The transmit power minimization problem is formulated and solved for the multiple waveguides multiple PAs system by jointly beamforming and PAs' positions design under the SLP principle. The alternating optimization (AO) framework is applied to decouple the beamforming vector and the position coefficient of PA. For the given beamforming vectors, a new objective function is formulated with respect to the positions of the PAs. With the characteristics of the formulated objective function, the optimization problem for the position coefficients of PAs can be decomposed into multiple independent subproblems, each corresponding to a PA's position coefficient, and a projected gradient descent (PGD)-based method, constrained by the feasible movable region of each PA, is then developed to obtain the suboptimal position coefficients. The performance improvements achieved by the combination of PAS and SLP, as well as the effectiveness of the proposed algorithm are verified through the simulation results.

This paper presents a decentralized frequency-domain framework to characterize the influence of the operating point on the small-signal stability of converter-dominated power systems. The approach builds on Scaled Relative Graph (SRG) analysis, extended here to address Linear Parameter-Varying (LPV) systems. By exploiting the affine dependence of converter admittances on their steady-state operating points, the centralized small-signal stability assessment of the grid is decomposed into decentralized, frequency-wise geometric tests. Each converter can independently evaluate its feasible stability region, expressed as a set of linear inequalities in its parameter space. The framework provides closed-form geometric characterizations applicable to both grid-following (GFL) and grid-forming (GFM) converters, and validation results confirm its effectiveness.

This research presents a novel approach to solving the economic load dispatch (ELD) problem in smart grid systems by leveraging a multi-agent distributed consensus strategy. The core idea revolves around achieving agreement among generators on their incremental cost values, thereby enabling an optimal allocation of power generation. To enhance convergence and robustness, the study introduces an adaptive coupling weight mechanism within a fully decentralized consensus framework, carefully designed with appropriate initial settings for incremental costs. The proposed distributed control protocol is versatile it functions effectively in both constrained and unconstrained generator capacity scenarios. Importantly, the methodology ensures that total power generation continuously matches dynamic load demands throughout the dispatch process, maintaining system-wide balance. To accommodate fluctuating and time varying load profiles, a dummy node is incorporated into the network architecture, acting as a flexible proxy for real time demand changes. The resilience of the method is further evaluated under communication disruptions, specifically by analyzing generator link failures through a switching network topology. Stability of the system is rigorously established using a Lyapunov-based analysis, assuming an undirected and connected communication graph among agents. To validate the practical efficacy of the proposed technique, comprehensive simulations are conducted on the IEEE 30 bus test system within the MATLAB environment, confirming its accuracy, adaptability, and computational efficiency in realistic smart grid conditions.

The paper examines deep learning models for scribe verification in Chinese manuscripts. That is, to automatically determine whether two manuscript fragments were written by the same scribe using deep metric learning methods. Two datasets were used: the Tsinghua Bamboo Slips Dataset and a selected subset of the Multi-Attribute Chinese Calligraphy Dataset, focusing on the calligraphers with a large number of samples. Siamese and Triplet neural network architectures are implemented, including convolutional and Transformer-based models. The experimental results show that the MobileNetV3+ Custom Siamese model trained with contrastive loss achieves either the best or the second-best overall accuracy and area under the Receiver Operating Characteristic Curve on both datasets.

This study investigates the use of self-supervised learning embeddings, particularly BYOL-A, in conjunction with a deep neural network classifier for Music Genre Classification. Our experiments demonstrate that BYOL-A embeddings outperform other pre-trained models, such as PANNs and VGGish, achieving an accuracy of 81.5% on the GTZAN dataset and 64.3% on FMA-Small. The proposed DNN classifier improved performance by 10-16% over linear classifiers. We explore the effects of contrastive and triplet loss and multitask training with optimized loss weights, achieving the highest accuracy. To address cross dataset challenges, we combined GTZAN and FMA-Small into a unified 18-class label space for joint training, resulting in slight performance drops on GTZAN but comparable results on FMA-Small. The scripts developed in this work are publicly available.

Terahertz (THz) communication can offer terabit-per-second rates in future wireless systems, thanks to the ultra-wide bandwidths, but require large antenna arrays. As antenna apertures expand and we enter the near-field scenarios, the conventional binary classification of communication links as either Line-of-Sight (LoS) or Non-Line-of-Sight (NLoS) becomes insufficient. Instead, quasi-LoS scenarios, where the LoS path is partially obstructed, are increasingly prevalent, posing significant challenges for traditional LoS focusing and steering beams. The Airy beam serves as a promising alternative, utilizing its non-diffracting and curved trajectory properties to mitigate such blockages. However, while existing electromagnetics literature primarily explores their physical patterns without practical generation schemes, recent communication-oriented designs predominantly rely on learning-based frameworks lacking interpretable closed-form solutions. To address this issue, this paper investigates a closed-form Airy beam design to efficiently synthesize Airy beam phase profiles based on the positions of the transceivers and obstacles. Specifically, rigorous analytical derivations of the electric field and trajectory are presented to establish a deterministic closed-form design for ULA Airy beamforming. Leveraging 3D wavefront separability, this framework is extended to uniform planar arrays (UPAs) with two operation modes: the hybrid focusing-Airy mode and the dual Airy mode. Simulation results verify the effectiveness of our derived trajectory equations and demonstrate that the proposed closed-form design significantly outperforms conventional beamforming schemes in quasi-LoS scenarios. Furthermore, the proposed method achieves performance comparable to exhaustive numerical searches with low computational complexity and enhanced physical interpretability.

This work aims to address the design of fully distributed control protocols for stochastic consensus, and, for the first time, establishes the existence and uniqueness of solutions for the path-dependent and highly nonlinear closed-loop systems under both undirected and directed topologies, bridging a critical gap in the literature. For the case of directed graphs, a unified fully distributed control protocol is designed for the first time to guarantee mean square and almost sure consensus of stochastic multi-agent systems under directed graphs. Moreover, an enhanced fully distributed protocol with additional tunable parameters designed for undirected graphs is proposed, which guarantees stochastic consensus while achieving superior convergence speed. Additionally, our work provides explicit exponential estimates for the corresponding convergence rates of stochastic consensus, elucidating the relationship between the exponential convergence rate and the system parameters. Simulations validate the theoretical results.

Beyond-diagonal reconfigurable intelligent surfaces (BD-RISs) are an emerging RIS 2.0 technology for future wireless communication. However, BD-RISs are primarily passive without active amplification, suffering from severe multiplicative path loss. To address the concern of multiplicative path loss, in this work we investigate the active BD-RIS including the modeling, architecture design, and optimization. We first analyze the active BD-RIS using multiport network theory with scattering parameters and derive a physical and electromagnetic compliant active BD-RIS aided communication model. We also design two new active BD-RIS architectures, namely fully- and group-connected active BD-RISs. Based on the proposed model and architecture, we investigate the active BD-RIS aided single-input single-output system and derive the closed-form optimal solution and scaling law of the signal-to-noise ratio. We further investigate the active BD-RIS aided multiple-input multiple-output system and propose an iterative algorithm based on quadratically constrained quadratic programming to maximize the spectral efficiency. Numerical results are provided and show that the active BD-RIS can achieve higher spectral efficiency than the active/passive diagonal RIS and passive BD-RIS. For example, to achieve the same spectral efficiency, the number of elements required by active BD-RIS is less than half of that required by active diagonal RIS, showing the advantages of active BD-RIS.

This paper studies non-trivial consensus–a relatively novel and unexplored convergence behavior–on directed signed matrix-weighted networks subject to both additive and multiplicative measurement noises under time-varying topologies. Building upon grounded matrix-weighted Laplacian properties, a stochastic dynamic model is established that simultaneously captures inter-dimensional cooperative and antagonistic interactions, compound measurement noises and time-varying network structures. Based on stochastic differential equations theory, protocols that guarantee mean square and almost sure non-trivial consensus are proposed. Specifically, for any predetermined non-trivial consensus state, all agents are proven to converge toward this non-zero value in the mean-square and almost-sure senses. The design of control gain function in our protocols highlights a balanced consideration of the cumulative effect over time, the asymptotic decay property and the finite energy corresponding to measurement noises. Notably, the conditions on time-varying topologies in our protocols only require boundedness of elements in edge weight matrices, which facilitate the practicality of concept "time-varying topology" in matrix-weighted network consensus algorithms. Furthermore, the proposed protocols operate under milder connectivity conditions and no requirements on structural (un)balance properties. The work in this paper demonstrates that groups with both cooperative and antagonistic inter-dimensional interactions can achieve consensus even in the presence of compound measurement noises and time-varying topologies, challenging the conventional belief that consensus is attainable only in fully cooperative settings.

Spoofing-robust automatic speaker verification (SASV) aims to integrate automatic speaker verification (ASV) and countermeasure (CM). A popular solution is fusion of independent ASV and CM scores. To better modeling SASV, some frameworks integrate ASV and CM within a single network. However, these solutions are typically bi-encoder based, offer limited interpretability, and cannot be readily adapted to new evaluation parameters without retraining. Based on this, we propose a unified end-to-end framework via a three-class formulation that enables log-likelihood ratio (LLR) inference from class logits for a more interpretable decision pipeline. Experiments show comparable performance to existing methods on ASVSpoof5 and better results on SpoofCeleb. The visualization and analysis also prove that the three-class reformulation provides more interpretability.

This study provides a comprehensive synthesis of Artificial Intelligence (AI), especially Machine Learning (ML) and Deep Learning (DL), in ransomware defense. Using a "review of reviews" methodology based on PRISMA, this paper gathers insights on how AI is transforming ransomware detection, prevention, and mitigation strategies during the past five years (2020-2024). The findings highlight the effectiveness of hybrid models that combine multiple analysis techniques such as code inspection (static analysis) and behavior monitoring during execution (dynamic analysis). The study also explores anomaly detection and early warning mechanisms before encryption to address the increasing complexity of ransomware. In addition, it examines key challenges in ransomware defense, including techniques designed to deceive AI-driven detection systems and the lack of strong and diverse datasets. The results highlight the role of AI in early detection and real-time response systems, improving scalability and resilience. Using a systematic review-of-reviews approach, this study consolidates insights from multiple review articles, identifies effective AI models, and bridges theory with practice to support collaboration among academia, industry, and policymakers. Future research directions and practical recommendations for cybersecurity practitioners are also discussed. Finally, this paper proposes a roadmap for advancing AI-driven countermeasures to protect critical systems and infrastructures against evolving ransomware threats.

This paper investigates joint trajectory and active beamforming design for unmanned aerial vehicle (UAV)-enabled ultra-reliable low-latency communication (URLLC) systems under finite blocklength (FBL) transmission. Unlike conventional Shannon-capacity formulations, the FBL regime introduces a signal-to-interference-plus-noise ratio (SINR)-dependent dispersion penalty that increases the sensitivity of reliability to mobility-induced channel variations. To address this challenge, we develop a propulsion-aware model predictive control (MPC) framework that performs receding-horizon joint trajectory and multi-user beamforming optimization while enforcing FBL-based rate constraints. The resulting long-horizon nonconvex problem is decomposed into beamforming and trajectory subproblems using alternating optimization. Concave surrogate is constructed for the Shannon-capacity term, while convex approximations are derived for the dispersion term and the nonlinear propulsion power model, yielding tractable convex subproblems solved iteratively. Compared with an offline MPC baseline, where the predictive problem is solved once over the entire mission horizon without feedback updates, and a conventional offline trajectory-beamforming optimization, the proposed closed-loop framework achieves disturbance-resilient mission completion under UAV position disturbances. Simulation results show that, compared with maximum ratio transmission (MRT) and equal-power allocation, the proposed interference-aware design significantly improves URLLC reliability under stringent minimum rate constraints. The results also quantify the impact of antenna scaling, transmit power, and transmission time on FBL performance, providing insights for reliability-centric UAV-enabled wireless networks in 5G and beyond.

Physical layer key generation (PLKG) has emerged as a promising solution for achieving highly secured and low-latency key distribution, offering information-theoretic security that is inherently resilient to quantum attacks. However, simultaneously ensuring a high data transmission rate and a high secret key generation rate under eavesdropping attacks remains a major challenge. In time-division duplex (TDD) systems with multiple antennas, we derive closed-form expressions for both rates by modeling the legitimate channel as a time-correlated autoregressive (AR) process. This formulation leads to a highly nonconvex and time-coupled optimization problem, rendering traditional optimization methods ineffective. To address this issue, we propose a multi-agent soft actor-critic (SAC) framework equipped with a long short-term memory (LSTM) adversary prediction module to cope with the partial observability of the eavesdropper's mode. Simulation results demonstrate that the proposed approach achieves superior performance compared with other benchmark algorithms, while effectively balancing the trade-off between secret key generation rate and data transmission rate. The results also confirm the robustness of the proposed framework against intelligent eavesdropping and partial observation uncertainty.

Efficient 3D LiDAR point cloud compression (LPCC) and streaming are critical for edge server-assisted robotic systems, enabling real-time communication with compact data representations. A widely adopted approach represents LiDAR point clouds as range images, enabling the direct use of mature image and video compression codecs. However, because these codecs are designed with human visual perception in mind, they often compromise geometric details, which downgrades the performance of downstream robotic tasks such as mapping and object detection. Furthermore, rate-distortion optimization (RDO)-based rate control remains largely underexplored for range image compression (RIC) under dynamic bandwidth conditions. To address these limitations, we propose D-Compress, a new detail-preserving and fast RIC framework tailored for real-time streaming. D-Compress integrates both intra- and inter-frame prediction with an adaptive discrete wavelet transform approach for precise residual compression. Additionally, we introduce a new RDO-based rate control algorithm for RIC through new rate-distortion modeling. Extensive evaluations on various datasets demonstrate the superiority of D-Compress, which outperforms state-of-the-art (SOTA) compression methods in both geometric accuracy and downstream task performance, particularly at compression ratios exceeding 100x, while maintaining real-time execution on resource-constrained hardware. Moreover, evaluations under dynamic bandwidth conditions validate the robustness of its rate control mechanism.

Energy storage shifts energy from off-peak periods to on-peak periods. Unlike conventional generation, storage is duration-limited: the stored energy capacity constrains the duration over which it can supply power. To understand how these constraints affect optimal pricing and investment decisions, we extend the classic two-period peak-load pricing model to include duration-limited storage. By adopting assumptions typical of solar-dominated systems, we link on- and off-peak prices to storage investment costs, round-trip efficiency, and the duration of the peak period. The bulk of the scarcity premium from on-peak prices is associated with the fixed costs of storage as opposed to variable costs stemming from round-trip efficiency losses. Unlike conventional generators, the binding duration constraints lead storage to recover energy capacity costs on a per-peak-event basis instead of amortizing these costs over total peak hours. A numerical example illustrates the implications for equilibrium prices and capacity investment.

In this work, we propose an unsupervised domain adaptation (UDA) framework for 3D volumetric lesion detection that adapts a detector trained on labeled FDG PET/CT to unlabeled PSMA PET/CT. Beyond covariate shift, cross tracer adaptation also exhibits label shift in both lesion size composition and the number of lesions per subject. We introduce self-training with two mechanisms that explicitly model and compensate for this label shift. First, we adaptively adjust the detection anchor shapes by re-estimating target domain box scales from selected pseudo labels and updating anchors with an exponential moving average. This increases positive anchor coverage for small PSMA lesions and stabilizes box regression. Second, instead of a fixed confidence threshold for pseudo-label selection, we allocate size bin-wise quotas according to the estimated target domain histogram over lesion volumes. The self-training alternates between supervised learning with prior-guided pseudo labeling on PSMA and supervised learning on labeled FDG. On AutoPET 2024, adapting from 501 labeled FDG studies to 369 $^{18}$F-PSMA studies, the proposed method improves both AP and FROC over the source-only baseline and conventional self-training without label-shift mitigation, indicating that modeling target lesion prevalence and size composition is an effective path to robust cross-tracer detection.

Virtual sensing aims to infer hard-to-measure quantities from accessible measurements and is central to perception and control in physical systems. Despite rapid progress from first-principle and hybrid models to modern data-driven methods research remains siloed, leaving no established default approach that transfers across processes, modalities, and sensing configurations. We introduce MuViS, a domain-agnostic benchmarking suite for multimodal virtual sensing that consolidates diverse datasets into a unified interface for standardized preprocessing and evaluation. Using this framework, we benchmark established approaches spanning gradient-boosted decision trees and deep neural network (NN) architectures, and show that none of these provides a universal advantage, underscoring the need for generalizable virtual sensing architectures. MuViS is released as an open-source, extensible platform for reproducible comparison and future integration of new datasets and model classes.

The study explores a hybrid centralized-federated approach for Human Activity Recognition (HAR) using a Transformer-based architecture. With the increasing ubiquity of edge devices, such as smartphones and wearables, a significant amount of private data from wearable and inertial sensors is generated, facilitating discreet monitoring of human activities, including resting, sleeping, and walking. This research focuses on deploying HAR technologies using mobile sensor data and leveraging Federated Learning within the Flower framework to evaluate the training of a federated model derived from a centralized baseline. The experimental results demonstrate the effectiveness of the proposed hybrid approach in improving the accuracy and robustness of HAR models while preserving data privacy in a non-IID data scenario. The federated learning setup demonstrated comparable performance to centralized models, highlighting the potential of federated learning to strike a balance between data privacy and model performance in real-world applications.