Papers

Passive ultra high frequency (UHF) radio frequency identification (RFID) enables battery-free tags to communicate with a reader through backscatter. When multiple tags respond in the same time slot, their waveforms overlap at the reader, and a conventional reader that follows framed slotted ALOHA (FSA) discards the resulting collided slot. This limits the throughput of the overall protocol even though the received signal still contains recoverable information about the responding tags. To address this limitation, we propose Self-Attention Tag Recovery (SATR), a transformer-based decoding algorithm that operates directly on the baseband in-phase and quadrature (I/Q) samples received during a standard tag response. SATR uses self-attention to model the temporal structure of the modulated waveform and learns candidate tag representations. It jointly estimates the number of responding tags and, more importantly, decodes the bit sequence of each detected tag. We numerically evaluate the decoding and throughput performance of SATR over a range of collision sizes and recovery configurations, and validate it with measurements of commercial UHF-RFID tags. The results show that, with proper design and training, SATR can reliably decode collisions of up to four tags. It achieves a throughput of approximately $0.815$ tags per slot under single acknowledgment and $1.87$ tags per slot under full recovery, corresponding to $2.2$ and $5.1$ times the conventional FSA limit of $1/e \approx 0.368$ tags per slot, while approaching optimal decoding performance and outperforming existing collision recovery methods.

Active inference selects actions by minimising an expected free energy functional over predicted futures. However, adding an expectation over yet-unobserved outcomes means the free energy functional no longer has a Kullback-Leibler structure, which hinders message passing treatments of inference procedures. We propose an alternative formulation based on a Bethe free energy functional, fully supporting inference by message passing. The epistemic drive is maintained by imposing an information constraint, next to normalisation, marginalisation and form constraints, insisting that the mutual information between future observations, states and parameters given actions must be at least as large as the entropy of the goal prior. For a specific value of the corresponding Karush-Kuhn-Tucker multiplier, the stationary point of this constrained Bethe Lagrangian recovers the expected free energy solution. We show that, as the information demand is varied, the solved multiplier moves through its inactive, interior, and saturated regimes. In the inactive regime the agent's epistemic drive switches off entirely, while in the saturated regime it is maximal. We compare the performance of the constrained Bethe agent on three tasks against EFE and Q-MDP.

Hurricane debris removal is planned, contracted, and federally reimbursed on the basis of volume estimates, yet operational practice still relies on parametric forecasts with 41-90% documented over-estimation or on truck-load tallies that arrive only after hauling begins. We present DebrisHeightNet, a segmentation-conditioned monocular debris-height network that estimates spatially explicit debris volume from a single pass of post-event aerial RGB imagery, the kind of survey routinely flown within days of a hurricane landfall. We train only a lightweight 1.08 M-parameter head on top of two frozen vision foundation models. This head regresses height from a Depth Anything V2 backbone, conditioned on the debris segmentation of CLIPSeg-debris from our prior work. Because no post-hurricane debris-height ground truth exists, we synthesize the training target by confidence-weighted LiDAR-monocular fusion (CW-LMF), designed to suppress non-debris LiDAR returns. This fused target is a constructed supervision signal rather than ground truth, so we corroborate it against external references rather than claiming it as truth. A region-level power-law calibration, driven by each region's low-density debris fraction, converts model volume into an estimate of the reported hauled debris with quantified uncertainty. Across ten regions spanning five hurricanes and three states, the uncalibrated model agrees with an independent uncrewed-aerial-vehicle (UAV) survey of the training region at Spearman $\rho = 0.87$ and lands within 30% of the reported record where the Hazus and FEMA-hybrid parametric forecasts over-predict it by 2.7-4.8$\times$. Deployment requires no LiDAR, no ground access, and no second flight, so the method can produce spatially explicit volume estimates wherever single-pass post-event imagery is flown.

Circular healthcare supply chains are inherently complex, characterised by interdependencies among their actors and high uncertainty in product flows and performance. Current methods used to predict the outcomes of transitioning to circular economy (CE) are limited and mostly static. This paper demonstrates the use of simulation to assess the effect of introducing circular products and the implications across the healthcare supply chain accounting for variability. The laparoscopic scissors supply chain is chosen as a case study example. To the best of our knowledge, this is the first study that assesses the implications of introducing circular product (medical devices) designs at both the individual supply chain member and overall system level. The model can be also used to inform optimal inventory strategies for hospitals, to ensure that patient safety and hospital operations are maintained. Our findings suggest that adopting circular products can reduce the environmental impact, but to achieve significant reductions in both cost and emissions, it requires significant upfront investment. We discuss the theoretical and practical implications of our study in developing tools to support the transition to CE.

Recovering the directed acyclic graph (DAG) of a structural equation model (SEM) from observational data is a central problem in causal discovery. The iterative gradient descent and per-problem hyperparameter tuning of continuous-optimization methods are poorly suited to two practically important regimes: the sample-limited regime, where the number of samples is comparable to or smaller than the number of nodes in the DAG, and the compute-limited regime. This work proposes SURE-Ridge, a non-iterative, closed-form estimator for equal variance linear Gaussian SEM. The method performs parallel node-wise regressions with regularization parameters chosen adaptively by Stein's unbiased risk estimate (SURE), and applies an adaptive thresholding procedure to extract a DAG from the resulting soft adjacency matrix. Numerical results show that SURE-Ridge achieves the lowest structural Hamming distance in the small-sample regime and the lowest run time across all sample sizes tested, compared with NOTEARS, DAGMA, and GBNSL baselines.

Reconfigurable antennas are increasingly integrated into multi-antenna communication systems to exploit large apertures while reducing the hardware complexity, energy consumption, and implementation costs of classical massive arrays. Their reconfigurable electromagnetic (EM) properties, including dynamically varying radiation patterns and state-dependent mutual coupling, challenge the fixed-antenna and decoupled-port assumptions of conventional channel models. This motivates a physically consistent framework connecting Maxwell's equations, circuit theory, and information theory. In this tutorial, we develop a unified framework spanning three coupled dimensions: (i) reconfigurable antenna and transceiver architectures, (ii) physically consistent channel modeling, and (iii) physically consistent signal processing. We first establish a taxonomy covering tunable antennas, reconfigurable transceivers, and emerging array architectures, highlighting their reconfiguration mechanisms and hardware-performance trade-offs. We then develop modeling approaches based on Maxwell's equations, wavenumber-domain representations, multiport network theory, and computational electromagnetics, and use them to construct end-to-end channel and noise models that capture near-field propagation, mutual coupling, and circuit-level impairments. Building on these models, we examine architecture-aware channel estimation, beamforming, data detection, and channel decoding, emphasizing how physical structure reshapes algorithm design and performance-complexity trade-offs. Overall, the tutorial treats physical architecture, channel and noise models, and communication algorithms as coupled components of an end-to-end design, providing a unified foundation for physically consistent reconfigurable wireless systems.

Stochastic reach-avoid analysis of controlled dynamical systems is an important tool for safety-critical control under uncertainty, in which the reach-avoid probability is characterized by a Hamilton-Jacobi partial differential equation (PDE). However, solving this PDE using conventional numerical methods becomes computationally intractable as the system dimension increases. Physics-informed neural networks (PINNs) may converge to inaccurate local minima when trained primarily through PDE-residual minimization. Reinforcement learning (RL) offers a scalable alternative, but its learned value functions may be inaccurate or inconsistent with the governing PDE. This paper proposes a physics-informed RL (PIRL) framework that combines the complementary strengths of PINNs and RL for stochastic reach-avoid analysis. We develop a scheduled PIRL algorithm in which temporal-difference actor-critic learning first guides the critic toward a meaningful approximation of the reach-avoid value function. PDE-residual and boundary-condition losses are then introduced progressively to enforce consistency with the governing PDE and its boundary conditions. The proposed method mitigates the failure modes of conventional PINN techniques while achieving accuracy comparable to that of successfully trained PINNs. The effectiveness of the proposed framework is demonstrated through two case studies.

A very low complexity feature extractor called next iRDT is proposed and evaluated for the problem of keyword spotting (KWS). Unlike any other types of feature extractors including the widely used MFCC, or adaptive, CNN-based ones, our algorithm is multiplier-free and it employs only simple, energy-efficient arithmetic operators. Since keyword-spotting of speech commands (KWS) is a typical application for TinyML platforms requiring low complexity for the signal classification chain, we consider it as a case study to evaluate complexity and functional performance. If properly tuned, iRDT demonstrates similar accuracy to solutions based on MFCC or CNN-based extractors using baseline classifiers on Google's KWS 12-classes dataset. With a different classifier the system achieved 94.7% validation accuracy. Processing times on CPU for the proposed feature extractor, are at least one order of magnitude smaller than for the MFCC. The proposed algorithm has a very low hardware footprint, making it ideal for ultra-low power edge devices. Code and demo are available [18].

Modern multimodal foundation models (MFMs) have made rapid progress on tasks requiring integrated perception across speech, vision, and language, including emotion recognition. However, it remains unclear whether they recognize speech and facial emotion through shared affective functional units or modality-specific pathways. We explore emotion-sensitive neurons (ESNs), sparse decoder neurons selectively associated with emotion categories, in three MFMs: Gemma-4-12B-it, MiniCPM-o-4.5, and Qwen2.5-Omni-7B. Using speech emotion recognition and facial expression recognition as complementary probes, we identify acoustic and visual ESNs. Visual ESNs are causally meaningful: deactivating them selectively impairs recognition of the associated facial emotion, whereas steering their activations selectively enhances recognition of that emotion relative to other emotion categories. Acoustic and visual ESNs further show emotion-matched overlap and similar layer-wise distributions, indicating partial structural alignment between affective representations across speech and faces. Finally, cross-modal interventions reveal bidirectional causal transfer: ESNs identified from one modality produce emotion-specific effects when applied to the other. Our findings provide one of the first cross-modality activation-level analyses of affective functional units in MFMs, suggesting that speech and facial emotion recognition partially converge onto sparse decoder-level components that can be localized and manipulated without training.

Linear temporal logic is a modal extension of propositional logic that allows one to state how a system should behave over time. Its canonical domain is the booleans, but discretely-valued judgements are of little use in steering softly-valued systems (neural policies, adaptive controllers, sequence models, etc). In such cases, the goal formula's (dis)satisfaction becomes a training signal, and differentiability becomes a prime concern. Candidate differentiable semantics abound, but navigating them is tricky. Implementations, where available, are shallow embeddings, demanding an upfront commitment to a single semantic algebra and its (usually implicit) conduct. The paper casts the reader as a functional programmer asked to come to terms with this predicament, and refusing. Out of that refusal comes an evaluation engine that is algebra-generic and amenable to differentiation, together with an executable specification of the algebras it can accept. Various algebras are implemented and audited for their behavior, both forward and backward. Each algebra turns out to be a choice of which direction to disappoint, and how. Everything described (and more) is part of the PyTorch library telos, to be found at https://github.com/konstantinosKokos/telos.

Hardware optimization is critical in the design of efficient wireless communication systems. Wireless communication hardware often consumes a significant fraction of the total system's power budget, with much of this power used in circuits that reduce various types of noise, particularly in the analog front end. The Single-Chip Micro Mote, or SCuM, uses a crystal-free radio architecture and makes design trade-offs that favor power consumption over noise performance while maintaining standards compatibility with popular Internet-of-Things (IoT) protocols such as IEEE 802.15.4 and Bluetooth Low Energy. In the continued development of SCuM, we recognize that the digital baseband hardware developed can be more closely optimized with the architecture of the chip. In this paper, we present an extensible Python-based RTL generator that is closely linked to simulation and testing environments. This approach provides flexibility for use on different hardware platforms, such as tape-outs and FPGA implementations, and has promise in AI-assisted design workflows.

The global transition toward green hydrogen is driving the deployment of gigawatt-scale electrolysis centers, introducing a novel, converter-dominated load class to the bulk power system. Unlike conventional industrial loads, these facilities utilize extensive power electronics interfaces with fast dynamics comparable to Inverter-Based Resources (IBRs). This paper presents a comprehensive grid impact assessment of large-scale hydrogen hubs, focusing on harmonic injection, voltage stability in low Short Circuit Ratio (SCR) environments, and frequency response capabilities. Adopting a "full-spectrum" open-source modeling approach, the study utilizes PandaPower for large-scale steady-state contingency assessment; ANDES for electromechanical dynamic simulations to evaluate Fast Frequency Response (FFR); and ParaEMT for high-fidelity electromagnetic transient analysis of harmonic distortion and Low Voltage Ride-Through (LVRT). A critical finding of this study is that standard load models, including the generic PERC1 (data center) model, are insufficient for hydrogen hubs. The paper recommends specific structural modifications to the PERC1 model - specifically regarding process safety latches and restart voltage thresholds - to accurately capture the risk of prolonged plant tripping. Based on these findings, the paper proposes a set of standardized connection requirements to ensure secure integration.

Blueberry ripeness is judged by berry colour, cluster composition, and the distribution of maturity stages within a plant, however, public green house image resources with dense ripeness-stage masks remain limited. We present AerialYield-B2D, where B2D denotes BlueBerry Dataset, acurated real-image resource containing 514 RGB images and 30,195 annotated blueberry instances across five ripeness stages: green immature, pale pink, pink-turns-purple, fully ripe and over-ripe. The release provides class-specific binary masks, overall berry masks, semantic label maps, image-level count tables, SHA-256 hashes, source metadata, recommended train/validation/test splits and technical validations. AerialYield is the broader project name; this release does not provide harvest weight, fruit mass or per-area yield measurements, and the count labels should therefore be interpreted as image-level berry counts rather than yield estimates. The images include 424 smartphone greenhouse images, 67 video-derived frames, and 23 DJI Fly video-frame samples, providing a reproducible dataset for ripeness segmentation, berry counting, and class-imbalance analysis in controlled-environment blueberry production.

Magnetic resonance imaging (MRI) is central to prostate cancer assessment, yet its acquisition is costly and time-consuming, making it a major bottleneck in the patient care pathway. Personalized MRI synthesis is a promising direction because it may allow the generation of clinically realistic images tailored to individual patients while reducing the dependence on scanner-acquired imaging. In this work, we in-vestigate whether patient-personalized MRI synthesis produces images that experts perceive as clinically plausible and how such personaliza-tion influences expert interpretation. We synthesize patient-personalized prostate MRI using a 3D diffusion model conditioned on routine pre-imaging clinical variables age and PSA. Using a large publicly available dataset, we first verify that synthesized images are technically compara-ble to scanner-acquired MRIs using standard image similarity metrics. We then conduct a pilot, single-blinded, randomized expert study on a subset of cases sampled to reflect variation in patient clinical pro-files. A radiologist compared personalized and unconditioned synthetic MRIs using paired assessments and Likert-scale ratings of clinical plausibility, confidence, and risk of misleading interpretation, supplemented by qualitative feedback. Although personalized synthetic MRIs appeared technically plausible, expert interpretation highlighted variability in con-fidence and potential risks of misleading cues when routine clinical data was embedded into image generation. These findings suggest that while personalized synthesis may be technically feasible, careful assessment is needed to understand how generated cues influence clinician interpreta-tion before such systems can be safely integrated into clinical workflows.

Identifying individuals at risk of Alzheimer's disease (AD), particularly in the preclinical and early stages, remains challenging. Although deep learning approaches based on structural MRI show promise as a non-invasive biomarker, existing multimodal models require task-specific training and depend on biomarkers that are not routinely available in clinical practice. Here, we propose a zero-shot multimodal framework based on SigLIP that combines structural MRI embeddings with text embeddings of routinely collected clinical variables for early AD risk stratification in individuals at preclinical or mild cognitive impairment (MCI) stages. We evaluated the approach in 416 individuals from the ADNI cohort (age: 72.73 +- 6.7). SigLIP was used without fine-tuning to extract MRI and clinical text embeddings, which were combined into multimodal representations for individual-level AD risk prediction within 4 years. We further compared the model performance in a single-visit and two-visit settings to assess the value of longitudinal information and framework scalability. In the single-visit setting, combining MRI embeddings with MMSE, age, and sex achieved an AUC of 0.91 +- 0.02, outperforming both a CSF Ae̱ta}42-based model (AUC 0.73 +- 0.08) and an MMSE-based model (AUC 0.85 +- 0.22). In the two-visit setting, performance was maintained or improved, supporting the scalability of the approach to longitudinal data. These findings suggest that zero-shot multimodal fusion of structural MRI and routinely collected clinical variables may provide a practical and scalable strategy for early AD risk stratification without task-specific retraining.

This paper investigates a movable-element simultaneously transmitting and reflecting reconfigurable intelligent surface (ME-STARS) assisted rate-splitting multiple access (RSMA) system under imperfect channel state information (CSI). Unlike conventional STARS with fixed element positions, the elements of ME-STARS can be repositioned within a predefined region, providing additional spatial degrees of freedom for improving the cascaded transmitter–STARS–user channels. To exploit this flexibility while accounting for CSI uncertainty, we formulate a robust sum-rate maximization problem that jointly optimizes the transmit beamforming, common-rate allocation, reflection and transmission coefficients, and ME-STARS element positions, subject to transmit-power, user-rate, minimum inter-element spacing, and movement-region constraints. The resulting problem is highly non-convex due to the strong coupling among the design variables and the position-dependent channels. To address this challenge, an iterative optimization framework is developed in which the transmit beamforming, STARS coefficients, and element positions are successively optimized through tractable convex reformulations. In particular, the element positions are updated sequentially using a majorization–minimization (MM) framework, where quadratic surrogate functions are constructed from the first- and second-order derivatives of the position-dependent channels while preserving the minimum inter-element spacing constraint. Simulation results demonstrate that the proposed ME-STARS design consistently outperforms the considered benchmark schemes. Moreover, the performance gains remain significant under increasing CSI uncertainty, highlighting the effectiveness of element repositioning for robust RSMA transmission.

Time-critical interactive systems increasingly require ultra-low-latency device identification for multiple users, yet prevailing approaches such as passwords, QR codes, and RFID/NFC are constrained by human input, frame-based sensing, or near-contact range. This paper presents ECO-ID, an event-camera-based optical system for multi-user, ultra-low-latency identification over visible light communication (VLC). Leveraging microsecond-resolution, asynchronous observations of brightness transitions, ECO-ID employs a spatiotemporal coding design: disjoint LED subsets provide spatial separation among users, while user-specific timing delays encode identities without inter-user synchronization. The optical channel and event-driven sensing reduce full-scene capture relative to frame cameras and limit the RF attack surface, while enabling rapid token verification with freshness and replay protection. We implement a prototype and demonstrate that ECO-ID can practically achieve approximately 99.8% localization and 98.7% identification with 0.64 ms mean latency, while theoretically supporting identification at the scale of tens of concurrent users. Overall, ECO-ID provides a fast, privacy-conscious, and security-aware alternative for scalable multi-user identification in time-critical interactive environments.

Interconnected systems have been widely studied, with a focus on interconnected systems whose subsystems are solely input-to-state stable (ISS) or passive systems. The focus of this work is on the interconnected systems that appear in adaptive gradient methods. In adaptive gradient methods, one subsystem seeks to move parameters of a cost function towards a minimizer while the other subsystem works to estimate some derivative information about the cost function to help determine the direction of the parameter update. This work studies instances of such interconnected systems and gives various Lyapunov function constructions for them using different techniques. Global asymptotic stability (GAS) of adaptive gradient optimizers follows from small-gain theory; the Lyapunov functions constructed here certify it explicitly.

Lung nodule segmentation in computed tomography is essential for extracting clinically relevant information for lung cancer assessment and treatment planning. Foundation models have shown notable segmentation capabilities, but state-of-the-art approaches often depend on input prompts, such as points or boxes, making their performance sensitive to prompt quality and placement. Understanding the limitations and constraints of prompt-based foundation models is therefore essential for designing reliable medical image segmentation solutions. In this work, we investigate how prompt quality affects foundation models performance for lung nodule segmentation. We further propose a synthetic prompt-generation model to test if the dependence on manually provided prompts can be mitigated by generating synthetic prompts that can also improve segmentation performance. Perturbation experiments show that bounding box prompts generally outperform point prompts, while latest specialized medical imaging models achieve better performance than general purpose ones. The proposed approach obtains a Dice coefficient of 0.85, suggesting that synthetic prompt generation as a promising strategy for lung nodule segmentation with foundation models.

We investigate power-efficient multiuser integrated sensing and communication (ISAC) assisted by an element-grouping extremely large-scale intelligent reflecting surface (EG-XL-IRS). The grouping pattern is designed using slowly varying statistical channel state information (S-CSI), so that both IRS-related channel acquisition and online passive beamforming operate in the group domain rather than the element domain. We reveal a fundamental gain-rank tradeoff induced by element grouping: phase-consistent grouping can coherently enhance selected deterministic propagation components, while excessive concentration on a common deterministic mode can reduce the effective spatial rank of the multiuser channel and, for extended targets, the diversity of desired-scatterer responses. Motivated by this observation, we develop a task-adaptive rank-aware grouping strategy that balances weak-user enhancement and target-scatterer illumination while preserving task-relevant spatial dimensions. For each candidate grouping pattern, the transmit covariances and group-wise reflection phases are jointly optimized under communication and sensing quality-of-service constraints, followed by physical phase recovery and feasibility verification. Numerical results show that the proposed design substantially reduces the required transmit power compared with representative grouping benchmarks under the same grouping dimension and online optimization budget.

Design structure matrices (DSMs) are used to comprehensively represent complex systems. They visualize and describe the dependencies between various variables, processes, states, and events. As such they are used in several system engineering approaches, such as requirement and interface management, fault detection, and supervisory control. Currently, a DSM is typically built from knowledge of experts. This may lead to an incomplete or imbalanced DSMs. For instance, elements and links might be missing or superfluous. In this article, we propose a novel method to acquire the DSM using state-of-the-art network identification methods. This demonstrates a proof-of-principle of identifying DSMs from data as an additional tool to the standard heuristic approach. In the future, we plan to embed DSMs in system design and supervisory controllers. We apply this technique to identify the DSM of a fusion reactor modelled by a five-chamber plasma model describing the transport in a tokamak.

Individually measuring head-related transfer functions (HRTFs) at scale remains a central challenge for personalised spatial audio, motivating growing interest in synthetic HRTFs. We evaluated the numerical, computational, and behavioural validity of synthetic HRTFs, generated through the boundary element method simulation using Mesh2HRTF, against measured and KEMAR HRTFs using the Extended SONICOM dataset. Across 200 subjects, synthetic HRTFs deviated less from measured than KEMAR in interaural time and level differences, but residual errors, together with elevated spectral distortion, concentrated at low, rear elevations. This is consistent with the omission of torso geometry from the synthesis pipeline. Two computational models revealed a corresponding pattern of predicted localisation errors, with synthetic HRTFs positioned between measured and KEMAR. In a virtual reality localisation task (N = 20), synthetic HRTFs matched measured on every polar metric, while KEMAR was significantly worse. However, behavioural error clustered around the front-back midline regardless of condition, not at the low elevations implicated numerically or by the models. A separate spatial release from masking task (N = 18) showed no effect of HRTF type. Together, these results indicate that high-resolution synthetic HRTFs preserve behavioural localisation performance, despite discrepancies between the numerical/model-predicted bias and the spatial pattern of behavioural error.

This paper presents a real-time orthogonal frequency-division multiplexing (OFDM) radar embedded in the OpenAirInterface (OAI) 5G base-station process. The radar removes communication symbols by regularized element-wise division and performs range-Doppler processing and ordered-statistic constant-false-alarm-rate detection online without modifying the 5G waveform. The implemented system provides 2.57 m nominal range resolution and 0.28 m/s velocity resolution. Hardware measurements identify and mitigate several implementation-specific limitations, most notably a deterministic carrier-dependent transmit-receive phase rotation on a Universal Software Radio Peripheral (USRP) X300. Selecting a tuning-grid-aligned carrier improves mean-removal clutter suppression from -16.4 dB to 38.0 dB and reduces coherent-integration loss from 19.81 dB to 0.27 dB. The measured processing gain closely agrees with its predicted value. Instrumented worker timing confirms real-time operation, with a conservative 58.4 percent utilization bound and no dropped soundings. A custom E2 service model, E2SM-RADAR, exports detections and a compact slow-time product to a near-real-time RAN Intelligent Controller. Live end-to-end operation demonstrates reliable delivery and supports controller-side tracking, micro-Doppler analysis, and classification. With a commercial user equipment connected on the same carrier, measurements show no measurable difference in downlink throughput estimate with sensing enabled, while the radar sensing bandwidth follows the scheduler allocation. Video demonstration can be found in youtube: youtu.be/GaYAcTJ8RKI

An accurate estimation of the state of health (SOH) underpins safe and optimized use of the battery system. Although compelling, data-driven SOH estimation models typically require large amounts of high-quality labeled cycling data, while in practice such labels are often sparse in both quantity and coverage. Therefore, in this work, we propose a degradation-aligned self-supervised learning (SSL) framework based on a convolutional neural network-gated recurrent unit (CNN-GRU) model, which learns aging-consistent representations from unlabeled data through a cycle-order ranking objective as the pretext task for pretraining, thereby enabling robust SOH estimation after fine-tuning on sparsely labeled data. Test results showcase that the proposed ranking-based SSL approach proves to endow the pretrained model with degradation awareness from unlabeled data, and after fine-tuning the model can carry out accurate, robust SOH estimation, even when only an extremely limited amount of 1% of unevenly distributed labeled training data is available, where the MAE of 1.718% and RMSE of 2.329% can be achieved on the test cell. In addition, in-depth analyses are presented regarding the influences of label distribution and cross-cell robustness. We believe this work could shed new light on label-efficient SOH estimation of lithium-ion batteries, addressing a practical need in battery management.

Oscillation is a critical issue that power systems have long faced. Especially over the past two decades, with the large-scale inte-gration of renewable energy into the grid, oscillation problems have posed a serious threat to the secure operation of power systems. However, the current literature has not fully explained the oscillation mechanism of renewable energy integrated power systems (REIPSs). In this paper, the underlying mechanism of stimulated oscillations is explored, with novel analytical methods proposed. Firstly, it is explained from both mathematical formu-las and physical interpretations that for an oscillation mode characterized by a pair of complex conjugate poles, the oscilla-tion risk under disturbance depends on the relative positional relationship between the corresponding poles and all other poles and zeros on the complex plane, rather than their standalone locations, i.e., the stability perceived by classical theory. Then the underlying mechanism of high amplitude oscillations induced by closely-located poles under even slight disturbance is clarified. On this basis, a theoretical framework for stimulated oscilla-tions applicable to REIPSs, covering its definition, mechanism, and methods, is proposed. Finally, this paper discusses the rela-tionship between the stimulated oscillation theory proposed herein and the classical stability-based theory, revealing that the research findings surpass rather than negate the classical theo-ries.

Objective assessment of learning remains a fundamental challenge in education. Electroencephalography (EEG) provides a direct, non-invasive window into the neural correlates of knowledge acquisition, including cognitive familiarity. This study benchmarks fifteen machine learning (ML) and deep learning (DL) models for EEG-based familiarity prediction across two cognitive domains: faces (factual knowledge) and mathematical equations (conceptual knowledge). Using continuous EEG data from 23 participants, we extract spectral features (Power Spectral Density) across six frequency bands. We show that while standard stratified cross-validation yields artificially high classification performance (up to 0.9853 F1-score using CNN) due to temporal leakage across neighboring epochs, a rigorous trial-independent validation (Group K-Fold) drops the peak performance to 0.6038 F1-score (using CNN), which is still statistically significant above the 25% chance level. This highlights the critical necessity of trial-independent evaluation to avoid overestimating model generalizability. Furthermore, feature importance and SHAP analysis reveal that temporal and frontal Gamma and Beta oscillations are the most critical biomarkers for familiarity. This work establishes a realistic benchmark for EEG-based cognitive monitoring in educational technologies.

Large Audio Language Models (LALMs) have made rapid progress on standardized benchmarks, yet their deployment in practical media workflows, curation, archival indexing, and content distribution remains largely unrealized. We identify automated audio chapterization, the task of segmenting continuous audio streams into thematically coherent chapters, as a demanding and commercially consequential setting that exposes this gap. Chapterization is challenging because boundaries are defined less by objective acoustic events than by subjective editorial judgment, requiring models to reason sequentially over long acoustic contexts and approximate creator-authored boundary decisions. We present AudioChaps, a post-training framework for aligning end-to-end LALMs for this task via Group Relative Policy Optimization (GRPO) guided by Chain-of-Thought (CoT) reasoning. To support training and evaluation, we curate three datasets: AudioChaps-Alignment, derived from creator-annotated chapter boundaries on YouTube; AudioChaps-CoT, which provides structured supervision for well-formatted, high-quality, and evidence-grounded boundary reasoning; and AudioChaps-Eval, a held-out benchmark for audio chapterization. Applying GRPO directly without a Supervised Fine-Tuning (SFT) cold start, AudioChaps-R1-Zero already improves average F1 by 33 points over the state-of-the-art LALM Audio-Flamingo-3-Think. The AudioChaps framework produces our final aligned LALM, AudioChaps-R1, which improves average F1 by 49 points. These results demonstrate that GRPO-trained LALMs can reliably transform unstructured auditory streams into navigable, structured media. Our code, models, and dataset resources will be released upon acceptance at https://github.com/ta012/AudioChaps.

Near-field antenna measurements underpin the characterization of electrically large apertures, yet the fidelity of the Near-Field to Far-Field (NF-FF) transformation depends on the reconstruction algorithm's assumptions and robustness to real-world imperfections, including those from drone-based scanning platforms. Classical FFT-based modal expansion is efficient on uniformly sampled canonical grids but fails when phase-coherent acquisition cannot be maintained. We address this via a phaseless NF-FF algorithm reconstructing the far field from amplitude-only data through iterative phase retrieval. When sampling becomes sparse or irregular, even amplitude-based methods break down, motivating the $\textbf{Adaptive Sparse Inverse Radiation Estimator (ASPIRE)}$, a full-complex inverse source framework that solves a Method-of-Moments problem over RWG basis functions via cascaded rSVD and regularized shrinkage. Mutual coupling between basis functions is explicitly resolved, improving reconstruction fidelity beyond coupling-agnostic inverse-source formulations. The solver is accelerated via a Multilevel Fast Multipole Method engine with Numba just-in-time compilation, achieving a $1.2\times$ reduction in matrix-vector product time and up to $15\times$ lower memory usage relative to dense evaluation at N=100K. Across frequency bands and positioning/truncation error scenarios, the pipeline sustains algorithmic stability and achieves sub-degree beamwidth reconstruction error. These results establish an error-aware framework for algorithm selection across fixed and drone-based near-field measurement platforms.

This paper outlines a sonification design to support fault detection in the transmission of I2S transport signals. I2S is a protocol for communicating real-time digital audio between integrated circuits that, while in wide and general use, does not include built-in error detection. Moreover, given the nature of the protocol transmission faults affecting timing, framing and alignment can be difficult to identify using conventional visual methods. The proposed design addresses this with an approach informed by Audification, wherein oversampling controls temporal rescaling to render protocol structure (SCK and WS) and payload data (SD) across separate stereo channels. A preliminary computational feasibility study was carried out to measure feature-space separability of I2S faults in the generated auditory representations as opposed to listener performance. It evaluates the design across several payload types and error conditions including jitter, bit-slip, and word-length errors. Class separability was assessed through clustering analyses of extracted features. The evaluation results show that while oversampling produces systematic changes in feature values, it does not meaningfully improve separability between error classes. However, a modest but consistent improvement in separability is observed as a function of the joint representation of structural and payload information across channels. The findings suggest that feature-space separability in sonified communication protocol data may be dependent on the integration of complementary information streams, rather than on signal scaling alone.

Modern power systems face growing operational complexity driven by the integration of renewable energy sources, decentralization, and the need for real-time decision-making across a wide range of timescales. Addressing these challenges traditionally relies on model-based methods that, while accurate, can be too slow for operational demands. Machine learning (ML) has therefore emerged as a faster, data-driven alternative. As grid topology plays a central role in power system operation, graph machine learning (GML) methods offer a natural framework for incorporating topological dependencies as an inductive bias. We survey nearly 800 papers at the intersection of GML and power systems, covering forecasting, state estimation, optimization, control, fault diagnosis, and cybersecurity. Power systems constitute an unusually rich benchmark setting for GML, as they combine hard physical constraints, multi-scale dynamics, safety-critical requirements, and scarce labeled data within a single, well-defined domain. Conversely, power systems can benefit from utilizing GML to complement classical solvers, as GML provide scalable, topology-aware approximations with promising generalization and computational efficiency. We identify open challenges, including limited real-world deployment and the need for interpretable models in safety-critical settings. Despite the rapidly growing number of publications, standardized benchmarks and open datasets remain scarce, leaving many results difficult to reproduce and undermining the long-term scientific credibility of the field. We further derive a structured requirements catalog for ML-ready power grid benchmarks, intended to guide future dataset development and improve reproducibility across studies. We call on the community to prioritize dedicated benchmark studies and the release of open datasets and models.