Papers

This paper presents a robust Expectation-Maximization framework for covariance estimation in Scale-Invariant Random Vector (SIRV) models with missing data under ignorable missingness mechanisms. By adopting an inverse-gamma prior on the scale variables, the resulting observation model leads to a complex multivariate Student-t distribution and allows closed-form E-step and M-step updates. The proposed algorithm incorporates numerical robustness techniques such as computation reuse for common observation patterns, regularized matrix inversions, and explicit enforcement of Hermitian positive semidefinite structure. Experiments on synthetic data and Sentinel-1 interferograms show effective missing value reconstruction and denoising performance under both MCAR and MNAR scenarios.

Diffusion models show potential for speech enhancement but lack linguistic guidance. We condition a diffusion-based model on wav2vec 2.0 features from noisy input, injected at the U-Net bottleneck via Feature-wise Linear Modulation (FiLM). Phonetic representations from wav2vec 2.0 features of degraded speech, anchor the reverse diffusion process. While a frozen wav2vec 2.0 encoder extracts features, a learned FiLM generator produces scale and shift parameters modulating the bottleneck with minimal overhead. Motivated by the optimal Bayesian causal estimator under a linear-Gaussian state-space model, FiLM coefficients are aggregated via exponential smoothing for temporal compression. Evaluation on VoiceBank-DEMAND and LibriMix shows competitive performance against the unconditioned baseline in PESQ, STOI, SI-SDR and DNSMOS. We consistently record an improvement of 0.4 on PESQ score, suggesting self-supervised representations effectively condition diffusion-based speech enhancement.

Large data center loads introduce new resilience challenges to power systems because their disconnection and staged reconnection can induce fast voltage and frequency dynamics that are not captured by static service-status or energy-based metrics. This paper proposes a utility-side, physics-informed resilience assessment framework that evaluates these events using only grid-side dynamic models and observable post-disturbance trajectories, without requiring detailed internal data center models. An unsupervised differential algebraic equation-physics informed neural network (DAE-PINN) based on an implicit backward Euler residual is developed to jointly predict dynamic and algebraic states, enabling repeated post-disturbance trajectory evaluation while enforcing network algebraic consistency. Normalized multi-phase resilience metrics are then used to quantify disturbance, degraded-state, and restoration-period impacts and to screen data center reconnection timing and load-ramping strategies under security constraints. Case studies on a modified IEEE 33-bus feeder show that the proposed DAE-PINN accurately tracks numerical DAE solutions and substantially reduces computation time in repeated restoration screening. The proposed metrics distinguish the effects of disturbance size, data center location, and reconnection strategy, revealing the trade-off between restoration speed and transient resilience loss.

Adaptive room equalization remains challenging under time-varying acoustic conditions and complex excitation signals, such as music. In these scenarios, classical filtered-x least mean squares (Fx-LMS) methods falter due to their rigid formulation. We present a modular differentiable digital signal processing (DDSP) framework for closed-loop adaptive room equalization that recovers Fx-LMS as a special case through automatic differentiation. The framework supports interchangeable EQ structures, response estimation methods, loss functions, and optimizers. Experiments with time-varying measured room impulse responses show that frequency-domain objectives provide more stable adaptation than time-domain objectives in the considered scenarios. Relative to the non-equalized response, system distance is reduced by 70% and mel-spectral distance by 13% (worst-case scenario). We further examine how online room response estimation accuracy and frame length affect the trade-off between responsiveness and convergence stability. Overall, the framework provides a unified open-source basis for exploring synergies between classical adaptive filtering and DDSP-based optimization.

The Functional Mock-up Interface (FMI) 3.0 standard for Model Exchange is restricted to hybrid ordinary differential equations, requiring any internal algebraic equations to be solved inside the Functional Mock-up Unit (FMU) before derivatives are returned to the importer. For models originating from, e.g. Modelica, this means that non-linear algebraic equations must be solved through internal Newton iterations, which can reduce accuracy, increase computational cost, introduce hidden solver states, and cause robustness issues in downstream simulation and optimization workflows. In this article, we present a proposal for a layered standard, fmi-ls-dae, that exposes algebraic equations and their associated algebraic variables as part of a semi-explicit index-1 differential-algebraic equation. We describe the proposed extensions to the FMI XML schema and demonstrate the approach through prototype implementations: Dymola and CasADi generate FMUs that expose this semi-explicit index-1 formulation, while CasADi, FMIOPT, Simcenter Twin Activate, fmusim, and MOO (the dynamic optimization tool of OpenModelica) import them for simulation and dynamic optimization. On an industrially relevant multilink suspension corner model, the proposed DAE-FMU formulation enables the optimization routine to converge on an optimal control problem on which the equivalent ODE-FMU fails to converge. We outline ongoing work towards consistent initialization, event handling and future support for higher-index DAEs.

Classical uncertainty sets for robust optimisation impose fixed geometric shapes that cannot represent the complex dependencies present in real-world data. We propose Generative Robust Optimisation (GRO), a framework in which a deep generative model defines the uncertainty set as the image of a neural network decoder over a calibrated latent set, naturally accommodating nonlinear correlations, asymmetry, and multimodality. A five-point evaluation framework (reconstruction fidelity, distribution matching, latent regularity, robust relevance, and computational tractability) provides systematic, model-agnostic criteria for assessing any neural network-based uncertainty set. We instantiate this framework with a Wasserstein Adversarial Autoencoder employing Gaussian mixture model-guided training for latent regularity and constraint-consistency regularisation for robust relevance. Restricting the decoder to ReLU activations enables exact worst-case verification through mixed-integer programming embedding. Extensive experiments on a production planning problem across six uncertainty distributions and six generative architectures, together with a multi-period facility location study, validate the framework and demonstrate that systematic attention to all five criteria yields uncertainty sets that are simultaneously expressive, well-calibrated, and optimisation-tractable.

Simultaneous wireless information and power transfer (SWIPT) has emerged as a promising paradigm for enabling sustainable connectivity in battery-limited low-altitude wireless networks (LAWNs). This paper investigates a SWIPT-enabled LAWN system in which a multi-antenna base station (BS) simultaneously delivers control information and wireless energy to a fleet of uncrewed aircraft systems (UASs) via power splitting. In particular, the BS remotely guides the UASs to accurately track predefined reference trajectories toward their destinations while avoiding multiple mobile no-fly zones (NFZs). To guarantee collision-free path planning, we first construct smooth and safe reference trajectories using stream function theory. Then, a real-time optimization problem is formulated, which jointly takes into account the wireless control cost and energy sustainability by optimizing control inputs, transmit beamforming vectors, and the power splitting ratios. To address the resultant non-convex problem, a two-stage optimization framework is proposed. First, we develop a model predictive control (MPC)-based method to generate predictive control inputs. Subsequently, we derive a computationally efficient iterative algorithm to optimize the beamforming vectors and power splitting ratios by applying semidefinite relaxation (SDR) and successive convex approximation (SCA) techniques. We further prove that the SDR is tight for our formulation. Extensive numerical results demonstrate that our proposed design significantly outperforms benchmark schemes in terms of tracking accuracy and harvested energy, thereby validating its effectiveness for sustainable implementation in LAWN systems.

Thermal management in a battery-electric vehicle (BEV) is a coupled, vehicle-level problem: the battery pack, the passenger cabin, the heat pump, and cabin air quality compete for shared actuation and energy, yet most studies optimise a single subsystem on proprietary models, which prevents fair, reproducible comparison. We present OpenEV-ThermoSciML, an open and reproducible benchmark that couples a battery electro-thermal-aging model, a two-node cabin model, a heat-pump/HVAC model, and a CO$_2$/ventilation model under real driving cycles (EPA) and real weather (NREL TMY3, NASA POWER), scored by a multi-objective suite spanning battery health, PMV/PPD comfort, cabin air quality, and HVAC energy. The benchmark's battery thermal core is anchored and validated on real BEV battery-management-system (BMS) data; the reduced battery (two-state) and cabin (two-node) models are validated against converged higher-fidelity references and, for the cabin, independently cross-checked against EnergyPlus 25.2.0. On top of the benchmark we develop a physics-informed scientific-machine-learning (Sci-ML) surrogate – a nominal-physics prior plus a learned residual with conservation penalties – that is exact on conserved quantities and dominates black-box and Koopman surrogates out-of-distribution (overall rollout RMSE 0.014 vs 1.168 and 3.991). A shielded Sci-ML model-predictive controller (MPC) delivers statistically significant, all-positive improvements over a production-like rule-based controller across six scenarios – including a real hot-day US06 trip (energy $-15\%$, comfort RMSE $-47\%$, peak CO$_2$ $-25\%$, battery thermal-gradient $-78\%$) – and these gains transfer to an independently exported OpenModelica 8-node co-simulation plant.

A cross-fused generative site-specific beamforming (GenSSBF) framework is proposed for low-overhead beam alignment in uniform planar array (UPA) systems. A decoupled channel sensing strategy is developed, where the azimuth and elevation domains of the UPA are probed independently, and the online sweeping overhead is reduced from multiplicative to linear complexity compared to exhaustive two-dimensional codebook sweeping. However, the resulting reference signal received power (RSRP) observations only contain marginal angular power information. The explicit azimuth-elevation coupling of the UPA channel is therefore lost. Beam generation from these separate observations becomes highly ambiguous. To address this issue, a bidirectional cross-attention encoder is designed to extract and fuse the latent dependency between the azimuth and elevation sensing branches. Conditioned on the fused feature, a conditional normalizing flow generator is proposed to generate a compact set of high-fidelity beam candidates. These candidates are further verified through lightweight pilot measurements for final beam selection. A task-oriented training objective is also introduced to encourage the generated candidate set to contain at least one high-gain beam, rather than fitting the full conditional beam distribution. Simulation results based on DeepMIMO scenarios show that the proposed framework consistently outperforms deterministic beam prediction and conventional discrete Fourier transform (DFT) codebook search. Compared with the full 1024-beam two-dimensional DFT search, normalized beamforming gain improvements of 83.6%, 74.6%, and 38.1% are achieved in the I2_28, O1B_28, and Boston5G_28 scenarios, respectively, while the sweeping overhead is reduced by 93.8%.

Speech language models (SLMs) increasingly combine speech and text, often by interleaving their tokens within a single sequence. Yet how these two modalities interact in the model's latent space remains unclear. In this work, we analyze interleaved speech–text LMs from different model families and training configurations using three complementary methods. We reveal that these models pass through an implicit latent transcription phase in which the text token matching the spoken word becomes decodable in intermediate layers, despite not being trained for speech recognition. This phenomenon occurs in diverse, natural speech, and intermediate representations also encode likely text continuations. We further show that implicit transcription emerges most when combining text-LM pretraining and speech–text interleaving, and that its prevalence is positively associated with spoken factual-knowledge retrieval. Our analysis sheds light on the internal interaction between speech and text modalities in interleaved SLMs.

Modular multilevel converter (MMC) has been successfully applied in various power electronic systems owing to its high efficiency, scalability, and superior output performance. Although the half-bridge submodule (HBSM) is widely used in MMCs for its structural simplicity, it is incapable of handling direct-current (DC) short-circuit faults. The diode-clamp submodule (DCSM) addresses this limitation by providing DC fault ride-through capability. However, because its two identical capacitors are connected in series, the equivalent capacitance is halved. To overcome this drawback, an enhanced SM is proposed in this paper. For the same energy storage capacity, the proposed SM reduces the total required capacitance by 75% compared with the DCSM. In addition, the proposed SM requires one fewer diode than the DCSM, thereby lowering the overall MMC capital cost. The topology and the operating modes of the proposed SM are described in detail, and its functionality is experimentally validated. The results demonstrate that the proposed SM can suppress DC fault currents and restore normal operation without additional external protection devices.

This paper proposes a novel method to reduce the peak-to-average-power-ratio (PAPR) of affine frequency division multiplexing (AFDM) signals without side information (SI). The method is based on circularly shifting the original transmit signal in the affine domain, and selecting the shifted candidate with the lowest PAPR. Next, a maximum-likelihood-based (MLB) receiver is derived, which exploits the position of the AFDM pilot and guard band to detect the shift applied at the transmitter without SI. Simulation results show that the proposed method can achieve 2.5 to 4 dB in PAPR reduction as compared to original AFDM and existing schemes, which can be translated into significantly lower error rate, depending on the quality of power amplifiers.

Distribution networks with high penetrations of distributed energy resources (DERs) must repeatedly allocate limited network capability in two directions: under import scarcity, which flexible demand is served, and under export congestion, which generation is curtailed. Dynamic operating envelopes (DOEs) enforce hard feasibility bounds but lack intertemporal correction, while dynamic network prices (DNPs) provide an allocative signal but cannot guarantee constraint satisfaction. This paper develops a stateful cyber-physical coordination mechanism, termed an Automatic Market Maker (AMM), as an additive coordination layer for machine-to-machine DER access. The mechanism combines dual fairness states for import and export, bounded bilateral prices driven by a voltage-aware deficit signal, and feasibility-constrained matching within a two-tier MV/LV architecture. Experiments on the CSIRO MV+33LV feeder dataset compare five mechanisms and benchmark the fair-over-time DOE formulations of Moring et al. (FET, FOT, FUH). Relative to equal-allocation DOE, the AMM reduces unserved flexible demand by 76% (96.0 MWh to 23.2 MWh) with zero thermal violations and reduces export curtailment from 85.4 MWh to 64.5 MWh. Near-identical DOE and DOE-GREEDY performance confirms that heuristic choice alone does not improve repeated constrained outcomes. The AMM reaches an annual inter-feeder Jain index of 0.9998, outperforming all DOE variants from month 6 onwards. Direct benchmarking against FET/FOT/FUH shows that these mechanisms achieve higher worst-feeder equity through an explicit max-min MV objective, but operate offline over predetermined horizons and do not provide bilateral scarcity signals, real-time operation, or participant-level intertemporal correction. The two approaches address different objectives and may be combined in future work.

Purpose: To evaluate whether large language model (LLM)-assisted label cleaning can identify label-report discordance in CT-RATE, a large-scale public chest CT dataset. Materials and Methods: After report-level deduplication, 24,446 unique radiology reports were identified. Twelve reports were excluded from the primary GPT-5.4 analysis because of Microsoft Azure AI Foundry content-safety filtering, leaving 24,434 reports and 439,812 label instances across 18 abnormality categories. GPT-5.4-derived binary labels were generated from report text using structured JSON output and compared with existing CT-RATE labels. Discordant instances were adjudicated by radiologists. In addition, 100 randomly sampled reports were manually annotated to compare CT-RATE labels, individual LLM-derived labels, and multi-LLM majority-vote labels against radiologist-annotated reference labels. Results: Overall agreement between GPT-5.4-derived and CT-RATE labels was 96.4%, with Cohen's kappa of 0.884. Lymphadenopathy showed the lowest agreement and kappa. In discordance review, radiologist adjudication supported GPT-5.4-derived labels in 72 of 97 (74.2%) general discordant instances and 91 of 99 (91.9%) targeted lymphadenopathy discordant instances. Against radiologist-annotated reference labels, multi-LLM majority-vote labels achieved the highest label-macro-averaged F1 score and Cohen's kappa. Conclusion: LLM-assisted label cleaning identified clinically meaningful label-report discordance in CT-RATE and may support scalable quality improvement of public imaging datasets. The cleaned dataset will be made publicly available to support future research.

Road traffic accidents remain a critical global crisis, consistently serving as a primary driver of preventable mortality and severe injury. These incidents are frequently precipitated by human error, including overspeeding, driving under the influence of alcohol, and cognitive fatigue. To address this urgent public safety challenge, this paper presents an intelligent, Internet of Things (IoT)-based Accident Prevention and Detection System (APDS) designed to systematically mitigate driver risk and optimize post-collision emergency responses. The proposed framework features a multi-tiered architecture capable of executing continuous real-time telemetry monitoring, proactive local alarm triggering, and automated situational intervention. Furthermore, the system integrates automated emergency communication protocols that aggregate immediate spatial coordinates via GPS and dispatch targeted alerts to medical facilities in close proximity, thereby optimizing response times and reducing accident-related fatalities.

Recent generative video compression methods leverage powerful generative priors to achieve perceptually pleasing reconstructions. However, most existing approaches require additional training to adapt generative models to produce realistic reconstructions from compact representations. In this paper, we propose ZeroGVC, a zero-shot generative video compression framework that leverages pretrained autoregressive diffusion priors for low-delay video reconstruction. ZeroGVC encodes the first frame of each group of pictures (GOP) with an image codec and represents subsequent P-frames through Codebook-Guided Autoregressive Latent Compression. This design is motivated by our observation that the compression scheme of denoising diffusion codebook models is effective in few-step consistency sampling. By selecting compact combinations of reproducible codebook noise vectors, ZeroGVC steers the latent denoising trajectory toward the target P-frame while allowing the decoder to reproduce the same trajectory in only a few denoising steps. In addition, we design an optional bidirectional reference mode that mitigates error propagation by leveraging the next I-frame context without introducing any additional bitrate overhead. Extensive experiments on standard video compression benchmarks demonstrate that ZeroGVC achieves superior perceptual reconstruction quality at ultra-low bitrates without any additional training.

A-mode ultrasound (US) has emerged as a promising modality for hand and wrist motion tracking. Prior works have mainly addressed static gesture classification or regression of a few degrees of freedom (DoFs), typically relying on non-wearable systems and external computing devices, and highlight the need for strategies to ensure robustness to sensor repositioning. In this work, we propose a framework for robust whole-hand and wrist kinematic tracking via wearable A-mode US using the WULPUS platform, tackling the regression of 23 DoFs directly on the probe. First, we introduce a compact (11285 parameters) multi-output convolutional neural network combined with an incremental training strategy, which improves inter-session generalization and reduces mean absolute error by more than 17% compared to a non-incremental approach. Second, we demonstrate, for the first time, the feasibility of end-to-end hand and wrist kinematic tracking entirely on-device. We deploy the model on the WULPUS nRF52832 microcontroller, achieving 0.73 mJ per inference, 29.1 ms latency, and showing the feasibility of full operation (data acquisition, online inference, and BLE streaming of results) within 33 mW, enabling up to 36 hours of continuous use and an 88% reduction in wireless bandwidth compared to raw data transmission.

Lesion segmentation in breast ultrasound involves two related challenges. In images with lesions, speckle noise, low tissue contrast, and posterior acoustic shadowing cause boundary leakage and incomplete contour delineation. In images without lesions, those same artifacts generate false-positive activations in regions resembling solid lesion tissue. This study addresses both failure modes through a single modification to the training objective. Rather than weighting every boundary pixel equally, the proposed loss scales contour penalties by per-pixel predictive entropy and the ground-truth boundary map, concentrating gradient emphasis on lesion margin locations where the network remains uncertain. The loss was evaluated on the BUSI dataset through a controlled ablation against two baselines: a model without boundary supervision and a model with uniformly weighted boundary binary cross-entropy. Across 97 lesion-containing test images, mean Dice scores were statistically indistinguishable between the proposed method and the no-boundary baseline (0.7624 versus 0.7616, paired Wilcoxon p = 0.27), confirming that lesion segmentation quality is preserved. The primary effect appears in specificity. False-positive activations on 20 no-lesion test images fell from 14 of 20 and 19 of 20 for the two baselines to 5 of 20 with the proposed approach (McNemar p = 0.012 and 0.0005). Non-overlapping Wilson 95% confidence intervals confirm the difference is both statistically significant and practically substantial. A post-hoc spatial temperature scaling step further reduced expected calibration error from 0.0201 to 0.0095 without altering segmentation masks. Entropy-guided boundary supervision and spatial calibration thus function as complementary training-level and inference-level refinements that improve specificity and probability reliability within a U-Net framework.

Idling Vehicle Detection (IVD) seeks to determine, at the final frame of a video clip, whether any vehicle is idling, meaning the vehicle is stationary with its engine running, using synchronized video from a remote surveillance camera and multichannel audio captured by spatially distributed wireless microphones along the roadside. Prior full-image, clip-level fusion approaches tend to overfit scene background and full-frame context, produce unstable temporal decisions, and lack an explicit spatial prior to align vehicles with microphones, which makes them brittle under domain shift and data inefficient. Instead, we introduce TAVR-IVD, an audio-visual framework guided by multi-object tracking. Our method detects vehicles, links detections into tracklets, and classifies each vehicle by operating on its tracklet. This design raises the effective signal-to-noise ratio, stabilizes temporal decisions through tracklets, enforces an explicit spatial prior to align vehicles with microphones, and adapts across domains with limited calibration annotations while remaining detector agnostic and efficient. To evaluate deployment robustness, we further curate two evaluation extensions, AVIVD-LT and AVIVD-M, covering inter-day and cross-site shifts.

This paper investigates control-aware attacks against ArduPilot-based Unmanned Aerial Vehicles (UAVs), inwhich an adversary exploits the sensitivity of flight-controller dynamics to parameter changes to cause loss of control and crashes. It describes six attacks that exploit interactions among multi-layer controllers by modifying Proportional-Integral-Derivative (PID) gains, altering Extended Kalman Filter (EKF) estimation configuration, and violating failsafe assumptions, thereby forcing ArduPilot into unsafe operating conditions. We evaluate the attacks in Software-in-the-Loop (SITL) simulation and validate them on a Pixhawk 2.4.8 hardware platform. The results show that short sequences of well-formed MAVLink messages can exploit controller sensitivity to parameter values and updates frequency, affecting controller states and degrading attitude stability, angular-rate behavior, trajectory tracking, and estimator health. We demonstrate that when multiple effects are combined, the vehicle can enter an unsafe state and crashes. These findings show that security gaps in input-parameter handling, command trust, and controller-state validation can be exploited to cause loss of control and crashes in UAVs.

Sixth-generation (6G) systems target higher reliability, denser connectivity, and tighter interference control. {Within this context, rate-splitting multiple access (RSMA) is envisioned as a promising candidate to enhance interference management in future wireless networks by flexibly splitting messages into a common and a private part, while fluid antenna systems (FAS) offer the potential to improve spatial selectivity through dynamic port reconfiguration.} Combining RSMA and FAS therefore enables efficient interference control and adaptive antenna utilization in multiuser multi-input single-output (MISO) networks. However, deriving closed-form outage probability (OP) expressions and tractable user fairness optimization in this scenario remains scarce in the literature. This paper studies a multiuser MISO downlink that jointly leverages RSMA and FAS. We develop a spatial correlation model for FAS using block correlation and incorporate linear precoding with zero-forcing and maximum-ratio transmission. Within this model, we derive closed-form OP expressions using a one-factor construction and generalized Gauss-Laguerre quadrature. Building on these expressions, we formulate a fairness objective that minimizes the worst-user OP and propose a low-complexity algorithm with a linear-program feasibility check to obtain the closed-form solution per iteration. Numerical results across different port counts, channel conditions, and target rates validate the analytical analysis, show that FAS-RSMA reduces OP by up to 92% relative to the fixed-position antenna (FPA) baseline, and demonstrate that fairness-oriented design equalizes user reliability while delivering a 1 dB SNR gain for the worst user at a fixed outage level.

Ensuring safety of learning-enabled robotic manipulation across diverse embodiments and tasks still requires significant manual engineering. For collision avoidance, existing approaches typically rely on embodiment-specific learned safety approximations or simplified robot models together with stopping-based fallback controllers that compromise performance. In this paper, we propose an embodiment-agnostic safeguarding approach, X-Safe, which provides formal probabilistic guarantees for collision avoidance without model simplifications or learned safety functions. By operating in the configuration space, our safeguarding method works with different types of robot policies. X-Safe reasons locally around the current configuration to establish safe regions and uses action masking to restrict the policy's support to these safe regions. We demonstrate X-Safe on two embodiments and three policies, both in simulation and on hardware. We observe less than half of the performance degradation compared to state-of-the-art safeguarding, and empirically corroborate our formal collision avoidance guarantees.

We study trajectory optimization in mobile sensing systems that must identify which member of a finite candidate set is the true target, while maintaining reachability to all potential candidate targets, under resource constraints. Deferred-Decision Trajectory Optimization (DDTO) addresses this setting by computing trajectories that reach individual targets but remain coincident for as long as possible before separating toward different targets. We propose Active-Sensing DDTO (AS-DDTO), which extends DDTO by adding a trajectory-dependent information-acquisition term to the planning objective. The resulting planner maintains reachability to candidate targets while biasing the coincident portion of the trajectories toward regions that enable earlier target identification. The framework supports Bayesian updates and conformal candidate-set updates for distance-dependent sensing. We derive a mixed-integer conic reformulation and provide guarantees on recursive feasibility, belief concentration, and fixed-time coverage for the raw conformal candidate set. Numerical simulations show improved target identification compared with standard DDTO under distance-dependent sensing uncertainty and limited sensing budget.

Cooperative aerial transport requires controllers that respect nonlinear manifold geometry, operate without centralized coordination, and respect operational safety constraints. To address these demands, we present GPAC, a four-layer hierarchical architecture that enables $N$ quadrotors to transport a cable-suspended payload without a central coordinator or by exchanging cable states or adaptive parameters. The key insight is implicit coordination: each quadrotor independently estimates its effective load share from local cable measurements, so combined forces converge to the correct total, even without knowledge of $N$ or the payload mass; the payload position is reconstructed locally from each agent's own cable geometry, and the only inter-agent communication is a low-rate neighbor-position broadcast for collision avoidance. GPAC operates directly on the full nonlinear configuration manifold and integrates geometric position and attitude control, anti-swing regulation, an extended-state observer for wind rejection, concurrent learning-based mass estimation without persistent excitation, and a priority-ordered control barrier function (CBF)-inspired safety filter that reduces operational risk, with input-to-state safety (ISSf) margins that hold exactly under single-constraint activation. A compatibility result shows that the filter's force modifications keep the desired attitude within the almost-global stability region of the $\mathrm{SO}(3)$ attitude controller. Finally, high-fidelity simulation with flexible cables, onboard sensor fusion, and wind turbulence – with all control and estimation loops closed through the estimator – yields a mean payload-tracking RMSE of 33.8 cm (2.8% coefficient of variation over 13 seeds) at a low per-agent computational cost.

Channel State Information (CSI) feedback is vital for high spectral efficiency in wireless systems, yet high-dimensional CSI introduce significant feedback overhead. Recent deep learning (DL) approaches alleviate this issue by treating CSI as a visual image, but such "black-box" designs often lack interpretability, producing CSI that is not consistent with multipath propagation principles. To address these limitations, this paper proposes HS-PINNnet, a Hierarchical Sensing mechanism assisted Physics-Informed Neural Network for CSI Feedback. Unlike vision-inspired methods, HS-PINNnet integrates a multipath channel model into the network, reformulating high-dimensional CSI reconstruction as low-dimensional multipath parameter estimation (e.g., amplitude, angle). HS-PINNnet features a hierarchical sensing encoder to produce a compact multipath representation, and a heterogeneous decoder for parameter-specific CSI reconstruction, with dedicated branches to estimate different parameters. Moreover, a PCD module adaptively estimates the number of dominant paths in each CSI sample to enhance generalization across diverse environments. A subchannel-wise shared encoding and parallel decoding strategy is further designed to decompose high-dimensional CSI processing into low-dimensional subchannel tasks, reducing training difficulty and improving scalability of HS-PINNnet for future extremely large-scale multiple-input multiple-output (XL-MIMO) systems. Simulation results show that HS-PINNnet outperforms the state-of-the-art under different configurations, achieving a 92.8% reduction in FLOPs and exhibiting two orders of magnitude lower FPGA simulation latency.

We frame the system as diagnostic data curation for a large audio-language model: before fine-tuning, we probe Qwen3-Omni-30B-A3B-Instruct under normal, empty-audio, and shuffled-audio conditions to identify how the model's answers change when audio evidence is removed or mismatched. These model confusion patterns are used to bucket training samples into text-prior, shuffle-leak, strong audio-dependent, and hard or misleading cases. Our strongest train-only system fine-tunes only on strong-audio items, where the normal audio-question pair is correct but both counterfactual variants fail, plus a small number of empty-audio negatives and a text-only response normalizer for parse-failed generations. On the official development set, the best train-only system reaches 67.27% accuracy after response normalization, compared with 65.90% for our local Qwen3-Omni baseline. Final submissions additionally include models trained using train+development splits and a three-model ensemble.

Electroencephalogram (EEG) analysis remains the clinical gold standard for epilepsy diagnosis and seizure detection. While Deep Learning (DL) has significantly advanced automated EEG interpretation, its transition from controlled experimental settings to routine clinical deployment is severely bottlenecked by fundamental architectural flaws. Standard DL models operate as opaque black-boxes lacking clinical interpretability, demand massive amounts of balanced annotated data, and incur steep computational costs incompatible with resource-constrained wearable or implantable neuromodulation devices. This paper presents a comprehensive review of these prevailing limitations and explores Kolmogorov-Arnold Networks (KANs) as a emerging paradigm for EEG-based seizure detection. By replacing the fixed activation functions of traditional neurons with flexible, learnable functions along the network's connections, KANs bridge the critical gap between predictive accuracy and mathematical transparency. We systematically analyze how KAN architectures resolve the shortcomings of traditional DL-based models by offering exceptional parameter efficiency, inherent interpretability for physician trust, and robust performance under data scarcity. Ultimately, this review establishes KANs not merely as an incremental algorithmic update, but as a fundamental paradigm shift necessary to actualize next-generation, patient-specific, and thoroughly transparent clinical EEG monitoring systems.

We consider learning-based control in LQR setting, where the parameters associated with each mode are a priori unknown. The next mode to be activated is revealed online only at the time of switching. The objective is to determine both the switching times and the control gains for each mode such that (1) the norm of the system state remains bounded according to a prescribed criterion, and (2) the accumulated cost is minimized. To formalize the state-norm requirement, we introduce the notion of $(\alpha,\beta)$-controllability for given parameters $\alpha$ and $\beta$. We first study the problem in a known model setting and show that, under the switching mechanism described above and under the assumption that each mode is visited infinitely often, the strategy that minimizes the average expected cost consists of applying, in each mode, the feedback gain obtained from the solution of the discrete algebraic Riccati equation, while selecting dwell times that sufficiently satisfy the controllability condition. We refer to this strategy as the benchmark policy. Next, we propose an algorithm for the unknown-model setting that minimizes the regret, defined as the difference between the cumulative cost incurred by the online algorithm and that of the offline benchmark. By accurately estimating dwell-time errors, our method achieves an expected regret of $\mathcal{O}(|\mathcal{M}|^{1/4} n_s^{3/4} + n_m)$, where $n_s$ denotes the number of switches, $|\mathcal{M}|$ is the number of modes, and $n_m$ is the number of malignant switches.

This paper studies PI tuning of a first-order-plus-dead-time (FOTD) plant for the fastest strictly monotone (zero-overshoot) setpoint step response, with monotonicity imposed on the plant output only. The minimizer is shown to be neither the pole-zero cancellation design nor the multiple real dominant pole (MRDP) design. It is a non-cancellation point at which the closed loop carries a slow real mode of small residue together with a faster underdamped complex pair, with the controller zero placed near the dominant real pole. The analytical centerpiece is a single tangency identity, tan(omega tau_star + alpha) = (a - b)/omega, which states that the monotonicity boundary is the locus where the secondary complex mode just fails to drive the output slope below zero. From this identity the design reduces to nested scalar conditions, realized at three levels of fidelity: an explicit closed-form rule, an exact response-based reduction, and a simulation-free transcendental system whose only non-elementary step is a fourth-order polynomial root. Relative to the critically damped Lambert-W cancellation rule the design reduces the 2% settling time by 14 to 52 percent and lowers the load integrated absolute error by 5 to 38 percent. We report the full cost: for delay-dominated plants (T/L <= 0.55) the control stays one-pulse, so the design is itself admissible in Huba's sense and merely faster, but for larger lag ratios the control becomes two-pulse, and across the range the maximum sensitivity rises from 1.39 to between 1.44 and 1.62. The contribution is positioned not as a uniformly better tuning but as the exact characterization of a specific, well-defined operating point, with an honest multi-metric comparison against established rules.

While longitudinal brain PET imaging is the gold standard for quantifying the spatiotemporal accumulation of Beta-amyloid, its widespread clinical utility is constrained by high operational costs and cumulative radiation risks. Recent deep generative models show promise in longitudinal image synthesis; however, they often fail to capture subtle pathological progression due to identity drift and a persistent bias toward trivially replicating baseline signal intensities rather than modeling temporal transition. To this end, we propose Delta-Diffusion, a novel progression-aware framework that redefines longitudinal PET synthesis as a conditional Poisson Diffusion Bridge (PDB) process. Unlike standard diffusion models that start from Gaussian noise, our PDB formulation is mathematically anchored to the subject's baseline PET, effectively transforming the generative task into a conditional distribution transition of the amyloid trajectory. To handle heteroscedastic nature of PET imaging, we introduce a physically-grounded Poisson perturbation within a Diffusion Transformer (DiT). This architecture uses adaptive scale-shift modulation to precisely calibrate the synthesis with the elapsed clinical interval and structural MRI context. A volume-of-interest balanced objective is designed to emphasize sparse, high-risk regions of amyloid accumulation. Validated on two cohorts with 542 subjects, Delta-Diffusion demonstrates superior performance in capturing longitudinal variations in amyloid deposition compared to state-of-the-art methods, offering a robust computational framework for tracking disease progression.