Papers

This paper presents a comprehensive review of dielectric barrier discharge (DBD) power supply topologies, aiming to bridge the gap between DBD applications and power electronics design. Two key aspects are examined: the dependence of the DBD electrical model on reactor geometry, and application-driven requirements for injected waveform characteristics, including shapes, voltage amplitude, frequency, and modulation techniques. On this basis, the paper systematically reviews two major categories of power supplies: sinusoidal types comprising transformerless and transformer-based resonant inverters, and pulsed power supplies (PPSs). The review summarizes performance trade-offs, highlights untested topologies and emerging applications, and offers guidance for advancing high-performance DBD power supply design for next-generation systems.

Out-of-distribution (OOD) detection is essential for the safe deployment of neural networks, as it enables the identification of samples outside the training domain. We present FOODER, a real-time, privacy-preserving radar-based framework that integrates OOD-based facial authentication with facial expression recognition. FOODER operates using low-cost frequency-modulated continuous-wave (FMCW) radar and exploits both range-Doppler and micro range-Doppler representations. The authentication module employs a multi-encoder multi-decoder architecture with Body Part (BP) and Intermediate Linear Encoder-Decoder (ILED) components to classify a single enrolled individual as in-distribution while detecting all other faces as OOD. Upon successful authentication, an expression recognition module is activated. Concatenated radar representations are processed by a ResNet block to distinguish between dynamic and static facial expressions. Based on this categorization, two specialized MobileViT networks are used to classify dynamic expressions (smile, shock) and static expressions (neutral, anger). This hierarchical design enables robust facial authentication and fine-grained expression recognition while preserving user privacy by relying exclusively on radar data. Experiments conducted on a dataset collected with a 60 GHz short-range FMCW radar demonstrate that FOODER achieves an AUROC of 94.13% and an FPR95 of 18.12% for authentication, along with an average expression recognition accuracy of 94.70%. FOODER outperforms state-of-the-art OOD detection methods and several transformer-based architectures while operating efficiently in real time.

In this paper, the feasibility of recently developed higher order delayed sliding mode controllers is addressed. With this aim the robustness against the measurement noise and mismatched perturbations for the systems governed by such controllers is established using ISS implicit Lyapunov-Razumikhin function approach. To illustrate proposed results, a simulation example validating the efficiency of the method is provided.

Imitation learning (IL) enables autonomous behavior by learning from expert demonstrations. While more sample-efficient than comparative alternatives like reinforcement learning, IL is sensitive to compounding errors induced by distribution shifts. There are two significant sources of distribution shifts when using IL-based feedback laws on systems: distribution shifts caused by policy error and distribution shifts due to exogenous disturbances and endogenous model errors due to lack of learning. Our previously developed approaches, Taylor Series Imitation Learning (TaSIL) and $\mathcal{L}_1$ -Distributionally Robust Adaptive Control (\ellonedrac), address the challenge of distribution shifts in complementary ways. While TaSIL offers robustness against policy error-induced distribution shifts, \ellonedrac offers robustness against distribution shifts due to aleatoric and epistemic uncertainties. To enable certifiable IL for learned and/or uncertain dynamical systems, we formulate Distributionally Robust Imitation Policy (DRIP) architecture, a Layered Control Architecture (LCA) that integrates TaSIL and~\ellonedrac. By judiciously designing individual layer-centric input and output requirements, we show how we can guarantee certificates for the entire control pipeline. Our solution paves the path for designing fully certifiable autonomy pipelines, by integrating learning-based components, such as perception, with certifiable model-based decision-making through the proposed LCA approach.

This paper analyzes the performance of spectro-temporal unitary transforms for coherent optical modulation. Unlike conventional IQ modulation, such transforms are based on a cascade of phase modulators and dispersive elements, so are theoretically lossless and not limited by the bandwidth of the constituent modulators. We analyse the performance limits and design trade-offs of this scheme: estimating how the number of stages, amount of dispersion, modulator bandwidth, symbol block length and electrical signal power impacts the achievable signal-to-distortion ratio (SDR). Importantly, we show that high (>30 dB) SDRs suitable for modern >200 GBd class coherent optical communications are achievable with a low (<6) number of stages and reasonable parameters for driver power, modulator bandwidth and on-chip dispersion. Finally we address the SDR penalties associated with potential phase, amplitude, or dispersion errors, and limited DAC resolution.

Recent advancements in robotic rehabilitation therapy have provided modular exercise systems for post-stroke muscle recovery with basic control schemes. But these systems struggle to adapt to patients' complex and ever-changing behaviour, and to operate within mobile settings, such as heat and power. To aid this, we present NeuRehab: an end-to-end framework consisting of a training and inference pipeline with AI-based automation, co-designed with neuromorphic computing-based control systems that balance action performance, power consumption, and observed latency. The framework consists of 2 partitions. One is designated for the rehabilitation device based on ultra-low power spiking networks deployed on dedicated neuromorphic hardware. The other resides on stationary hardware that can accommodate computationally intensive hardware for fine-tuning on a per-patient basis. By maintaining a communication channel between both the modules and splitting the algorithm components, the power and latency requirements of the movable system have been optimised, while retaining the learning performance advantages of compute- and power-hungry hardware on the stationary machine. As part of the framework, we propose (a) the split machine learning processes for efficiency in architectural utilisation, and (b) task-specific temporal optimisations to lower edge-inference control latency. This paper evaluates the proposed methods on a reference stepper motor-based shoulder exercise. Overall, these methods offer comparable performance uplifts over the State-of-the-art for neuromorphic deployment, while achieving over 60% savings in both power and latency during inference compared to standard implementations.

This paper proposes a novel preamble design and detection method for multiuser asynchronous massive MIMO LoRa networks. Unlike existing works, which only consider the preamble detection for a single end device (ED), we propose to simultaneously detect the preambles of multiple EDs that asynchronously transmit their uplink (UL) packets to a multiple-antenna gateway (GW). First, we show that the preamble detection in multiuser LoRa networks with the conventional single-chirp preamble suffers from the so-called preamble resemblance effect. This means that the preamble of any single ED can resemble the preambles of all EDs in the network and make it impossible to determine to which ED a preamble belongs. To address this problem, a novel double-chirp preamble design and a preamble assignment method are proposed, which can mitigate the preamble resemblance effect by making the preamble of each ED unique and recognizable. Next, a maximum-likelihood (ML) based detection scheme for the proposed double-chirp preamble is derived. Finally, since the proposed algorithm requires the calculation of the discrete Fourier transform (DFT) every sampling period, we propose a low-complexity technique to calculate the DFT recursively to reduce the complexity of our proposed design. Simulation shows that the proposed preamble design and detection require just about 2 dB more power to increase the number of EDs from one to 15 in the Rayleigh fading channel while achieving the same preamble detection error performance.

Ultra-reliable low latency communication (URLLC) is a key part of 5G wireless systems. Achieving low latency necessitates codes with short blocklengths for which polar codes with successive cancellation list (SCL) decoding typically outperform message-passing (MP)-based decoding of low-density parity-check (LDPC) codes. However, SCL decoders are known to exhibit high latency and poor area efficiency. In this paper, we propose a new short-blocklength multi-rate binary LDPC code that outperforms the 5G-LDPC code for the same blocklength and is suitable for URLLC applications using fully parallel MP. To demonstrate our code's efficacy, we present a 0.44mm$^2$ GlobalFoundries 22FDX LDPC decoder ASIC which supports three rates and achieves the lowest-in-class decoding latency of 14ns while reaching an information throughput of 9Gb/s at 62pJ/b energy efficiency for a rate-1/2 code with 128-bit blocklength.

Accurate estimation of subsurface material properties, such as soil moisture, is critical for wildfire risk assessment and precision agriculture. Ground-penetrating radar (GPR) is a non-destructive geophysical technique widely used to characterize subsurface conditions. Data-driven parameter estimation methods typically require large amounts of labeled training data, which is expensive to obtain from real-world GPR scans under diverse subsurface conditions. A physics-based GPR model using the finite-difference time-domain (FDTD) method can be employed to generate large synthetic datasets through simulations across varying material parameters, which are then utilized to train data-driven models. A key limitation, however, is that simulated data (source domain) and real-world data (target domain) often follow different distributions, which can cause data-driven models trained on simulations to underperform in real-world scenarios. To address this challenge, this study proposes a novel physics-guided hierarchical domain adaptation framework with deep adversarial learning for robust subsurface material property estimation from GPR signals. The proposed framework is systematically evaluated through the laboratory tests for single- and two-layer materials, as well as the field tests for single- and two-layer materials, and is benchmarked against state-of-the-art methods, including the one-dimensional convolutional neural network (1D CNN) and domain adversarial neural network (DANN). The results demonstrate that the proposed framework achieves higher correlation coefficients R and lower Bias between the predicted and measured parameter values, along with smaller standard deviations in the estimations, thereby validating their effectiveness in bridging the domain gap between simulated and real-world radar signals and enabling efficient subsurface material property retrieval.

Local Electricity Communities (communautés électriques locales, CEL) will become operational in Switzerland in 2026, allowing prosumers, consumers, and storage operators within the same municipality and distribution system operator (DSO) area to exchange electricity over the public grid with reduced distribution tariffs. This report examines a rural Swiss case study to explore the techno-economic implications of CELs for both participants and the local DSO. The findings indicate that CELs can enhance the local use of renewable generation, particularly photovoltaics, and offer modest financial gains, with outcomes strongly shaped by community size, composition, and tariff design. Larger and more heterogeneous communities achieve better internal matching of supply and demand, though the overall incentive remains limited because the tariff reduction applies only to distribution charges. The study further shows that internal energy exchange is maximized when local PV generation covers roughly 1-2 times the community load. For DSOs, CELs reduce grid imports (27-46%), resulting in a substantial reduction in distribution tariff revenues (17-36%), necessitating regulatory adaptation. While centralized batteries provide economic value to members, their technical impact on the grid remains modest due to their small, economically optimized capacity. Larger centralized storage is shown to reduce transformer peak power, but risks increasing line loading, suggesting a need for careful sizing and placement.

Large-scale supervised pretraining is rapidly reshaping 3D medical image segmentation. However, existing efforts focus primarily on increasing dataset size and overlook the question of whether the backbone network is an effective representation learner at scale. In this work, we address this gap by revisiting ConvNeXt-based architectures for volumetric segmentation and introducing MedNeXt-v2, a compound-scaled 3D ConvNeXt that leverages improved micro-architecture and data scaling to deliver state-of-the-art performance. First, we show that routinely used backbones in large-scale pretraining pipelines are often suboptimal. Subsequently, we use comprehensive backbone benchmarking prior to scaling and demonstrate that stronger from scratch performance reliably predicts stronger downstream performance after pretraining. Guided by these findings, we incorporate a 3D Global Response Normalization module and use depth, width, and context scaling to improve our architecture for effective representation learning. We pretrain MedNeXt-v2 on 18k CT volumes and demonstrate state-of-the-art performance when fine-tuning across six challenging CT and MR benchmarks (144 structures), showing consistent gains over seven publicly released pretrained models. Beyond improvements, our benchmarking of these models also reveals that stronger backbones yield better results on similar data, representation scaling disproportionately benefits pathological segmentation, and that modality-specific pretraining offers negligible benefit once full finetuning is applied. In conclusion, our results establish MedNeXt-v2 as a strong backbone for large-scale supervised representation learning in 3D Medical Image Segmentation. Our code and pretrained models are made available with the official nnUNet repository at: https://www.github.com/MIC-DKFZ/nnUNet

Aim: This study investigates treatment response prediction to neoadjuvant chemotherapy (NACT) in breast cancer patients, using longitudinal contrast-enhanced magnetic resonance images (CE-MRI) and clinical data. The goal is to develop machine learning (ML) models to predict pathologic complete response (PCR binary classification) and 5-year relapse-free survival status (RFS binary classification). Method: The proposed framework includes tumour segmentation, image registration, feature extraction, and predictive modelling. Using the image registration method, MRI image features can be extracted and compared from the original tumour site at different time points, therefore monitoring the intratumor changes during NACT process. Four feature extractors, including one radiomics and three deep learning-based (MedicalNet, Segformer3D, SAM-Med3D) were implemented and compared. In combination with three feature selection methods and four ML models, predictive models are built and compared. Results: The proposed image registration-based feature extraction consistently improves the predictive models. In the PCR and RFS classification tasks logistic regression model trained on radiomic features performed the best with an AUC of 0.88 and classification accuracy of 0.85 for PCR classification, and AUC of 0.78 and classification accuracy of 0.72 for RFS classification. Conclusions: It is evidenced that the image registration method has significantly improved performance in longitudinal feature learning in predicting PCR and RFS. The radiomics feature extractor is more effective than the pre-trained deep learning feature extractors, with higher performance and better interpretability.

Microelectromechanical systems (MEMS) speakers are compact, scalable alternatives to traditional voice coil speakers, promising improved sound quality through precise semiconductor manufacturing. This review provides an overview of the research landscape, covering baseband-displacement, ultrasound-based, and thermoacoustic sound generation concepts, classifying MEMS speakers by their actuation principles as electrodynamic, piezoelectric, or electrostatic devices. A comparative analysis of performance indicators from 1990 to 2026 highlights the dominance of piezoelectric MEMS with baseband displacement, focusing on miniaturization and efficiency. The review outlines upcoming research challenges and identifies potential candidates for achieving full-spectrum audio performance. A focus on innovative approaches could lead to widespread adoption of MEMS-only speakers.

Research results on Affine Frequency Division Multiplexing (AFDM) reveal that it experiences the same Peak-to-Average Power Ratio (PAPR) problem as conventional Orthogonal Frequency-Division Multiplexing (OFDM). On the other side, some references and also our studies demonstrate that AFDM involves an unneeded matrix, which is based on a parameter typically represented by $c_2$, for signalling. Hence, in this paper, an augmented AFDM scheme, referred to as A$^2$FDM, is proposed to mitigate the PAPR problem of AFDM, which is achieved by replacing the $c_2$ matrix in AFDM by a new unitary matrix that performs both sub-block-based Discrete Fourier Transform (DFT) and symbol mapping. Two symbol mapping schemes, namely interleaved mapping and localized mapping, are proposed for implementing A$^2$FDM, yielding the Interleaved A$^2$FDM and Localized A$^2$FDM. The input-output relationships of these schemes are derived and the complexity and the effects of system parameters on the performance of A$^2$FDM along with AFDM systems are analyzed. Furthermore, simulation results are provided to demonstrate and compare comprehensively the performance of the considered schemes in conjunction with different system settings and various operational conditions. Our studies and results demonstrate that, while A$^2$FDM is capable of circumventing the PAPR problem faced by AFDM, it is capable of attaining the achievable diversity gain, when AFDM is operated in its undesirable conditions resulting in the loss of the diversity gain available.

In this paper, we investigate an integrated sensing-and-communication (ISAC) network enabled by an unmanned aerial vehicle (UAV). The UAV is supposed to fly along a periodical circular trajectory at a fixed height for ISAC service supply from the sky. We consider on-demand sensing services, where on-demand detection and on-demand localization requests may be activated at any time toward any position within the targeted serving region. While guaranteeing satisfactory accuracy for both on-demand sensing tasks, we aim at maximizing the minimum achievable throughput among all communication users, via joint optimizing the UAV trajectory and communication user scheduling. To address the complicated problem with infinite sensing constraints, we characterize the on-demand detection constraint as a restricted deployment area for UAV and the on-demand localization constraint as Cramer-Rao Bound (CRB) constraints over finite reference target points, based on which the original problem is simplified to more tractable one. Afterwards, particularly aiming to ensure no violations of CRB constraints, we propose a convex approximation for the reformulated problem, where tight approximation is guaranteed at given local solution. The construction strategy for convex problem approximation allows an efficient iterative algorithm with verified convergence to a superior suboptimal solution. At last, with simulations, we verified the applicability of our developed optimization scheme in strictly fulfilling the on-demand sensing constraints and the effectiveness of our proposed solution for simultaneously enhancing the communication throughput in UAV-enabled ISAC.

This work introduces SkinGenBench, a systematic biomedical imaging benchmark that investigates how preprocessing complexity interacts with generative model choice for synthetic dermoscopic image augmentation and downstream melanoma diagnosis. Using a curated dataset of $14,116$ dermoscopic images from HAM10000 and MILK10K across five lesion classes, we evaluate the two representative generative paradigms: StyleGAN2-ADA and Denoising Diffusion Probabilistic Models (DDPMs) under basic geometric augmentation and advanced artifact removal pipelines. Synthetic melanoma images are assessed using established perceptual and distributional metrics (FID, KID, IS), feature space analysis, and their impact on diagnostic performance across five downstream classifiers. Experimental results demonstrate that generative architecture choice has a stronger influence on both image fidelity and diagnostic utility than preprocessing complexity. StyleGAN2-ADA consistently produced synthetic images more closely aligned with real data distributions, achieving the lowest FID ($\approx 65.5$) and KID ($\approx 0.05$), while diffusion models generated higher variance samples at the cost of reduced perceptual fidelity and class anchoring. Advanced artifact removal yielded only marginal improvements in generative metrics and provided limited downstream diagnostic gains, suggesting possible suppression of clinically relevant texture cues. In contrast, synthetic data augmentation substantially improved melanoma detection with $8$-$15$% absolute gains in melanoma F1-score, and ViT-B/16 achieving F1 $\approx 0.88$ and ROC-AUC $\approx 0.98$, representing an improvement of approximately $14\%$ over non-augmented baselines. Our code can be found at https://github.com/adarsh-crafts/SkinGenBench

Speech enhancement methods are commonly believed to improve the performance of automatic speech recognition (ASR) in noisy environments. However, the effectiveness of these techniques cannot be taken for granted in the case of modern large-scale ASR models trained on diverse, noisy data. We present a systematic evaluation of MetricGAN-plus-voicebank denoising on four state-of-the-art ASR systems: OpenAI Whisper, NVIDIA Parakeet, Google Gemini Flash 2.0, Parrotlet-a using 500 medical speech recordings under nine noise conditions. ASR performance is measured using semantic WER (semWER), a normalized word error rate (WER) metric accounting for domain-specific normalizations. Our results reveal a counterintuitive finding: speech enhancement preprocessing degrades ASR performance across all noise conditions and models. Original noisy audio achieves lower semWER than enhanced audio in all 40 tested configurations (4 models x 10 conditions), with degradations ranging from 1.1% to 46.6% absolute semWER increase. These findings suggest that modern ASR models possess sufficient internal noise robustness and that traditional speech enhancement may remove acoustic features critical for ASR. For practitioners deploying medical scribe systems in noisy clinical environments, our results indicate that preprocessing audio with noise reduction techniques might not just be computationally wasteful but also be potentially harmful to the transcription accuracy.

This work presents a 28nm 13.93mm2 CNN-Transformer accelerator for semantic segmentation, achieving 3.86-to-10.91x energy reduction over previous designs. It features a hybrid attention unit, layer-fusion scheduler, and cascaded feature-map pruner, with peak energy efficiency of 52.90TOPS/W (INT8).

This study investigates the stabilization of interlayer temperature in the laser powder bed fusion process through a novel switched layer-to-layer closed-loop feedback controller. The controller architecture aims to measure the interlayer temperature by a laterally positioned thermal camera and maintain a preset reference temperature by switching between the heating mode through dynamic laser power adjustment and the cooling mode by assigning interlayer dwell time to allow cooling between layers. The switching controller employs a feedback optimization control algorithm for the heating mode to adjust the laser power, and a triggering algorithm that increases the interlayer dwell time until the interlayer temperature reaches the reference value. Additionally, the study compares the performance of the proposed controller in both supported and unsupported overhanging parts to evaluate the effect of support structures on the controller performance as well as the thermal behavior of overhanging parts. Results demonstrate the controller's effectiveness in stabilizing interlayer temperature across varying cross-sectional areas while remaining within the material's stable processing zone. In the heating mode, the controller efficiently stabilizes temperature, even in geometries with significant cross-section variation. The study also identifies trade-offs among process efficiency, energy consumption, and build time. Supported parts exhibit reduced overheating but consume more energy and material, while unsupported parts stabilize interlayer temperature faster but with longer build times due to increased dwell time assignments. The research highlights notable improvements in interlayer temperature control for geometries prone to excessive thermal stresses. Moreover, the introduction of interlayer dwell time offers a practical solution to maintaining thermal stability in complex geometries.

Optimal transmission switching (OTS) improves optimal power flow (OPF) by selectively opening transmission lines, but its mixed-integer formulation increases computational complexity, especially on large grids. To address this, we propose a dispatch-aware deep neural network (DA-DNN) that accelerates DC-OTS without relying on pre-solved labels, eliminating costly OTS label generation that becomes impractical at scale. DA-DNN predicts line states and passes them through an embedded differentiable DC-OPF layer, using the resulting generation cost as the loss function so that physical network constraints are enforced throughout training and inference. To stabilize training, we adopt a customized weight and bias initialization that keeps the embedded DC-OPF feasible from the first epoch. To improve inference robustness, we incorporate a binary regularization term that reduces ambiguity in the relaxed line-status outputs prior to thresholding. Once trained, DA-DNN produces a feasible topology and dispatch pair with highly predictable computation time comparable to a single DC-OPF solve, while conventional MIP solvers can become intractable. Moreover, the embedded OPF layer enables DA-DNN to generalize to untrained system configurations, such as changes in line flow limits, and to support post-contingency corrective operation. As a result, the proposed method captures the economic advantages of OTS while maintaining scalability and generalization ability.

Medical image classification is a critical task in healthcare, enabling accurate and timely diagnosis. However, deploying deep learning models on resource-constrained edge devices presents significant challenges due to computational and memory limitations. This research investigates a resource-efficient approach to medical image classification by employing model quantization techniques. Quantization reduces the precision of model parameters and activations, significantly lowering computational overhead and memory requirements without sacrificing classification accuracy. The study focuses on the optimization of quantization-aware training (QAT) and post-training quantization (PTQ) methods tailored for edge devices, analyzing their impact on model performance across medical imaging datasets. Experimental results demonstrate that quantized models achieve substantial reductions in model size and inference latency, enabling real-time processing on edge hardware while maintaining clinically acceptable diagnostic accuracy. This work provides a practical pathway for deploying AI-driven medical diagnostics in remote and resource-limited settings, enhancing the accessibility and scalability of healthcare technologies.

Reconstructing high-quality images from substantially undersampled k-space data for accelerated MRI presents a challenging ill-posed inverse problem. While supervised deep learning has revolutionized this field, it relies heavily on large datasets of fully sampled ground-truth images, which are often impractical or impossible to acquire in clinical settings due to long scan times. Despite advances in self-supervised/unsupervised MRI reconstruction, their performance remains inadequate at high acceleration rates. To bridge this gap, we introduce UPMRI, an unsupervised reconstruction framework based on Projected Conditional Flow Matching (PCFM) and its unsupervised transformation. Unlike standard generative models, PCFM learns the prior distribution of fully sampled parallel MRI data by utilizing only undersampled k-space measurements. To reconstruct the image, we establish a novel theoretical link between the marginal vector field in the measurement space, governed by the continuity equation, and the optimal solution to the PCFM objective. This connection results in a cyclic dual-space sampling algorithm for high-quality reconstruction. Extensive evaluations on the fastMRI brain and CMRxRecon cardiac datasets demonstrate that UPMRI significantly outperforms state-of-the-art self-supervised and unsupervised baselines. Notably, it also achieves reconstruction fidelity comparable to or better than leading supervised methods at high acceleration factors, while requiring no fully sampled training data.

Voice-based human-machine interaction has become a primary means of accessing intelligent systems, yet individuals with dysarthria are systematically excluded by persistent gaps in recognition accuracy. Although automatic speech recognition (ASR) achieves word error rates (WER) below 5% on typical speech, performance degrades sharply for dysarthric speakers, while the zero-shot behaviour of multimodal large language models (MLLMs) on such speech remains unclear. We evaluate eight commercial speech-to-text services on the TORGO dysarthric speech corpus: four conventional ASR systems (AssemblyAI, Whisper large-v3, Deepgram Nova-3, Nova-3 Medical) and four MLLM-based systems (GPT-4o, GPT-4o Mini, Gemini 2.5 Pro, Gemini 2.5 Flash), using lexical accuracy, semantic preservation, and cost-latency measures. Recognition degraded consistently with severity. Mild dysarthria reached low single-digit WER, around 1-2% for the leading systems, whereas severe dysarthria exceeded 51% WER for every system, with no MLLM advantage over conventional ASR under default settings. A four-condition prompt ablation showed architecture-specific effects: for the OpenAI models, verbatim-transcription prompts reduced severe-tier WER mainly by suppressing non-target-language drift, lowering GPT-4o from 60.1% to 52.9% and GPT-4o Mini from 66.0% to about 55%; Gemini models showed no consistent benefit and sometimes degraded. Semantic metrics correlated strongly with WER and were largely redundant in aggregate, but identified cases where communicative intent was partly preserved despite poor lexical accuracy. These severity-stratified, per-speaker baselines provide a reusable reference for evidence-based technology selection in assistive voice interfaces.

We present an algorithm based on the alternating direction method of multipliers (ADMM) for solving nonlinear matrix decompositions (NMD). Given an input matrix $X \in \mathbb{R}^{m \times n}$ and a factorization rank $r \ll \min(m, n)$, NMD seeks matrices $W \in \mathbb{R}^{m \times r}$ and $H \in \mathbb{R}^{r \times n}$ such that $X \approx f(WH)$, where $f$ is an element-wise nonlinear function. We evaluate our method on several representative nonlinear models: the rectified linear unit activation $f(x) = \max(0, x)$, suitable for nonnegative sparse data approximation, the component-wise square $f(x) = x^2$, applicable to probabilistic circuit representation, and the MinMax transform $f(x) = \min(b, \max(a, x))$, relevant for recommender systems. The proposed framework flexibly supports diverse loss functions, including least squares, $\ell_1$ norm, and the Kullback-Leibler divergence, and can be readily extended to other nonlinearities and metrics. We illustrate the applicability, efficiency, and adaptability of the approach on real-world datasets, highlighting its potential for a broad range of applications.

Fetal brain magnetic resonance imaging (MRI) is crucial for assessing neurodevelopment in utero. However, fetal MRI analysis remains technically challenging due to fetal motion, low signal-to-noise ratio, and the need for complex multi-step processing pipelines. These pipelines typically include motion correction, super-resolution reconstruction, tissue segmentation, and cortical surface extraction. While specialized tools exist for each individual processing step, integrating them into a robust, reproducible, and user-friendly end-to-end workflow remains difficult. This fragmentation limits reproducibility across studies and hinders the adoption of advanced fetal neuroimaging methods in both research and clinical contexts. Fetpype addresses this gap by providing a standardized, modular, and reproducible framework for fetal brain MRI preprocessing and analysis, enabling researchers to process raw T2-weighted acquisitions through to derived volumetric and surface-based outputs within a unified workflow. Fetpype is publicly available on GitHub at https://github.com/fetpype/fetpype.

Optimal AP clustering and power allocation are critical in user-centric cell-free massive MIMO systems. Existing deep learning models lack flexibility to handle dynamic network configurations. Furthermore, many approaches overlook pilot contamination and suffer from high computational complexity. In this paper, we propose a lightweight transformer model that overcomes these limitations by jointly predicting AP clusters and powers solely from spatial coordinates of user devices and AP. Our model is architecture-agnostic to users load, handles both clustering and power allocation without channel estimation overhead, and eliminates pilot contamination by assigning users to AP within a pilot reuse constraint. We also incorporate a customized linear attention mechanism to capture user-AP interactions efficiently and enable linear scalability with respect to the number of users. Numerical results confirm the model's effectiveness in maximizing the minimum spectral efficiency and providing near-optimal performance while ensuring adaptability and scalability in dynamic scenarios.

As wireless communication systems continue to grow rapidly, high-performance antennas become increasingly crucial for expanding coverage, improving capacity, and enhancing transmission quality. In light of this, research has focused considerable attention on liquid antennas due to their unique characteristics, which include small size, flexibility, reconfigurability and transparency. Recently, graphene liquid has been explored for numerous applications due to its low cost, high conductivity, flexibility, and ease of processing. Specifically for antenna applications, graphene liquid performs better than conventional liquid metal. This paper presents a graphene-liquid antenna with beam reconfiguration ability for sub-6 GHz communication system. The graphene-liquid movement within the microfluidic channel is taken into consideration by the reconfiguration mechanism. The antenna achieves beam reconfiguration in 360{\deg} directions with 6 dBi of gain at 5.5 GHz, featuring a wideband impedance bandwidth of 24%. The antenna main beam is specifically reconfigured into six directions (0{\deg}, 45{\deg}, 135{\deg}, 180{\deg}, 225{\deg} and 315{\deg}) at 5.5 GHz. Additionally, in all six reconfigurable scenarios at 5.5 GHz, the antenna provides a stable reflection coefficient. Therefore, for the next generation of wireless communication systems, this novel design of graphene-liquid-based reconfigurable sub-6 GHz antennas holds promise.

Deep learning models for Electrocardiogram (ECG) analysis have achieved expert-level performance but remain vulnerable to adversarial attacks. However, applying Universal Adversarial Perturbations (UAP) to ECG signals presents a unique challenge: standard imperceptible noise constraints (e.g., 10 uV) fail to generate effective universal attacks due to the high inter-subject variability of cardiac waveforms. Furthermore, traditional "invisible" attacks are easily dismissed by clinicians as technical artifacts, failing to compromise the human-in-the-loop diagnostic pipeline. In this study, we propose SCAR (Semantic Cardiac Adversarial Representation), a novel UAP framework tailored to bypass the clinical "Human Firewall." Unlike traditional approaches, SCAR integrates spatiotemporal smoothing (W=25, approx. 50ms), spectral consistency (<15 Hz), and anatomical amplitude constraints (<0.2 mV) directly into the gradient optimization manifold. Results: We benchmarked SCAR against a rigorous baseline (Standard Universal DeepFool with post-hoc physiological filtering). While the baseline suffers a performance collapse (~16% success rate on transfer tasks), SCAR maintains robust transferability (58.09% on ResNet) and achieves 82.46% success on the source model. Crucially, clinical analysis reveals an emergent targeted behavior: SCAR specifically converges to forging Myocardial Infarction features (90.2% misdiagnosis) by mathematically reconstructing pathological ST-segment elevations. Finally, we demonstrate that SCAR serves a dual purpose: it not only functions as a robust data augmentation strategy for Hybrid Adversarial Training, offering optimal clinical defense, but also provides effective educational samples for training clinicians to recognize low-cost, AI-targeted semantic forgeries.

Rydberg atom-based RF sensors offer distinct advantages over conventional dipole antennas for electric field detection. This paper presents a system model and performance analysis of a Rydberg atom-based quantum radar, which employs optical readout via lasers and photon detectors instead of circuit-based receivers. We derive the signal-to-noise ratio (SNR), compare it with classical radar, and estimate Doppler frequency using an invariant function-based method. Simulations show that the quantum radar achieves higher SNR and lower RMSE in velocity estimation than conventional radar.

In multi-attribute decision-making problems where the attribute values are interval grey numbers, a simplified form based on kernels and the degree of greyness is presented. Combining fuzzy graph theory with the kernel and the degree of greyness of interval grey numbers, grey graphs and their corresponding operation rules are presented. This paper presents a new multi-attribute decision-making method based on grey graph theory. We analyzed and evaluated the alternative schemes using grey graph. Lastly, a numerical example was conducted in order to demonstrate the effectiveness and feasibility of the proposed method.