Papers

Localized generative editing needs localized evaluation: full-image identity metrics are structurally confounded under hard-composited edits. We present Envisage, a FLUX.1-Fill inpainting reference pipeline for rhinoplasty goal visualization from a single frontal photograph. The pipeline combines 8 rhinoplasty clinical presets (the released framework also includes 8 blepharoplasty and 8 rhytidectomy presets), MediaPipe masks, and hard-mask compositing. The composite preserves outside-mask pixels by construction, so full-face identity scores are dominated by copied pixels rather than by the diffusion backbone. Because full-face identity metrics cannot grade localized edits, we introduce SurgicalScore, a mask-decomposed 0-1 protocol scoring edit direction, edit magnitude, masked LPIPS, realism, and outside-mask preservation; SS_raw assigns 0.919 [0.918, 0.920] to a perfect-predictor control , anchoring the ceiling. On N=211, the paired ArcFace gain (output-to-GT minus input-to-GT) is negative for all methods (Envisage -0.048 smallest, vs. ICEdit -0.139, Kontext -0.242, InstructPix2Pix -0.294; p < 1e-4), with external validation on a 457-pair ASPS/PCA corpus showing a larger negative gap. With SurgicalScore, Envisage achieves the highest score (0.599 [0.579, 0.619]) and leads on both metrics, but the all-negative ArcFace gap shows that full-face identity is poorly aligned with localized surgical accuracy under hard compositing. A 5-seed GT-oracle (an upper bound, not a deployable result) reduces the residual ArcFace gap by 73% (-0.054 to -0.015), with positive output-to-GT gain on 33.9% of cases, indicating candidate-space headroom for a learned ranker. For localized edits, progress should be measured with edit-region fidelity rather than full-face identity metrics. We release Envisage, SurgicalScore, preset definitions, and matched split manifests.

The cart-pole swing-up is a canonical benchmark for nonlinear control of underactuated systems, yet an end-to-end guarantee linking the global swing-up maneuver to the local stabilizer is seldom formalized. We present a reachability analysis of a switched energy-based/LQR controller that certifies convergence to the upright equilibrium from a compact set of initial conditions. The swing-up design exploits the phase-space geometry of the conservative pendulum: the upright equilibrium lies on the homoclinic orbit, and an energy-shaping law drives the energy error to zero, steering the pendulum onto this orbit; convergence follows from LaSalle's invariance principle. An augmented Lyapunov function additionally regulates the steady-state cart velocity to zero, and we prove almost-global convergence of the resulting closed-loop system. A local LQR with a certified ellipsoidal region of attraction stabilizes the upright equilibrium, and we verify numerically that the swing-up phase delivers the state into this region, formalizing the handoff. Numerical simulations corroborate the theoretical analysis.

Rapid and reliable post-earthquake damage assessment is critical for public safety, re-occupancy decisions, and effective emergency response. This paper presents a physics-informed, unsupervised learning framework that enables structural damage diagnosis in sparsely instrumented buildings following seismic events. The approach fuses real sensor data with physics-based simulations to create a hybrid spatiotemporal input grid, extending observability to regions without sensors. A Spatiotemporal Composite Autoencoder Network (SCAN) processes this hybrid input, learning the structure's undamaged behavior from pre-event or ambient-condition data alone. SCAN integrates convolutional layers for extracting localized spatial features and LSTM layers for modeling temporal dynamics, enabling it to recognize deviations from normal behavior caused by damage. Post-event sensor data are analyzed through the trained model, and anomalies are flagged based on elevated reconstruction and prediction errors. These error patterns are spatially mapped to localize potential damage, even in uninstrumented areas. By embedding low-fidelity physical estimates directly into the input representation, the framework enhances detection sensitivity while requiring no labeled damage examples. This hybrid AI-physics approach offers a scalable, interpretable, and data-efficient solution for real-time post-earthquake structural diagnostics, providing critical decision support for emergency managers and accelerating safe and targeted recovery efforts.

Efficient and intelligent post-earthquake structural damage assessment is critical for rapid disaster response. Although data-driven approaches have shown promise in this domain, traditional supervised learning relies on large labeled datasets that are impractical to obtain for earthquake-damaged structures. To overcome this limitation, we propose a Bayesian-optimized multisource domain adaptation framework for predicting post-earthquake structural damage on a target building without the need for any damage labels. The framework comprises three key steps. First, it extracts features from multiple source domains and the target domain and feeds them into a classifier and a domain discriminator. The classifier ensures the features remain damage-sensitive, while the discriminator promotes their invariance across domains. Second, the framework assigns a weighing factor to each source domain to balance their contributions during training. Finally, Bayesian optimization is employed to optimize these source domain weights, aiming to maximize prediction accuracy on the target domain. This framework offers a robust solution for structural damage assessment when labeled data are scarce, significantly enhancing post-earthquake damage assessment capabilities.

Neuromorphic and edge computing research has focused on reducing the inference cost of neural network controllers, yet in physical closed-loop systems the actuator can rival or exceed an efficient controller in energy. An efficient controller is therefore necessary but not sufficient, because the actuator becomes the cost worth reducing once inference no longer dominates it. Here, we introduce energy-aware learning, an approach that incorporates actuator energy directly into the reinforcement learning reward, and demonstrate it in closed-loop deep brain stimulation (DBS) for Parkinson's disease. A deep spiking Q-network, trained in a biophysical cortico-basal ganglia-thalamic circuit model, learns to suppress pathological alpha-beta oscillations by 45.2% while reducing stimulation charge by 80.0% relative to continuous DBS. Sparsity-constrained knowledge distillation compresses the policy onto the SynSense XyloAudio 3 neuromorphic processor at 0.52 mW inference power, yielding 28.1x lower energy per inference than an equivalent artificial neural network on conventional edge hardware. By co-optimizing stimulation energy and inference efficiency, the framework addresses both major power demands in implantable neuromodulation.

Ensuring the operational safety of quadrotors under partial actuator failures, lumped external disturbances, and malicious cyberattacks is a critical challenge due to the system's underactuated and highly nonlinear nature. Building on the existing result of a fault-tolerant control approach for a quadrotor experiencing a complete loss of two opposing rotors \cite{chen2024quadrotor}, this letter further addresses the additional challenge of malicious cyberattacks, which could be unknown and unbounded. While the baseline control law, rooted in proportional-derivative (PD) feedback and observer-based decoupling, effectively handles mismatched disturbances, it remains vulnerable to maliciously injected cyberattacks on the pseudo-control channels. To address this, a Resilient Control Lyapunov Function-based Quadratic Program (RCLF-QP) is developed, where a resilient compensational term with real-time online adaptation is designed in the conventional CLF to compensate for the maliciously injected unknown and unbounded attacks. Compared with the PD feedback control, the proposed QP-based constrained optimization control framework provides a systematic and extensible framework that allows new control objectives and constraints to be seamlessly integrated without altering the underlying stability guarantees. The overall proposed controller integrates a model-based extended state observer with the proposed RCLF-QP mechanism to mitigate both lumped disturbances caused by aerodynamics and strong wind, and adversarial cyberattacks injected by malicious adversaries. Simulations in a high-fidelity environment demonstrate that the proposed RCLF-QP control architecture prevents trajectory divergence and system instability in scenarios where the baseline controller fails in maintaining the stability of Quadrotors under malicious attacks.

Accurate assessment of sweetness is essential for quality control in agriculture, yet conventional methods rely on destructive sampling and are difficult to scale. This thesis presents a robotic arm-based spectral sensing system for strawberry detection, localization, approach, and non-destructive sweetness estimation. The system integrates perception, calibration, and robotic control in a closed-loop pipeline. A YOLOv11s detector is adopted for real-time strawberry detection, while RGB-ToF calibration and mask-to-depth alignment are used to obtain geometrically consistent target localization. A custom eye-in-hand hand-eye calibration workflow is developed to estimate the rigid transform between gripper_link and cam_front, enabling reliable transformation of fruit targets into the robot base frame. Based on these estimates, the robot executes a waypoint-based search and an incremental closed-loop approach strategy to position the sensor at optimal working distance for sweetness sensing. Experimental results show strong end-to-end performance (88.10% success over 42 trials), with robust detection (95.24%) and successful approach execution once a target is detected (100% conditional success). Hand-eye calibration comparisons indicate that although Andreff yields the smallest translation norm in single-run results, the Park method provides better cross-sample consistency and therefore more stable downstream robot behavior. The residual failures are concentrated in the sensing stage, especially valid-region extraction for sweetness estimation under difficult depth/reflectance conditions. Overall, this work demonstrates the feasibility of integrating RGB-ToF perception, robotic manipulation, and non-destructive sensing for practical strawberry quality assessment, and provides a scalable baseline for future integration of learning-based policies such as Vision-Language-Action models.

Positron emission tomography (PET) seeks to balance diagnostic quality with ra-diation dose. Low-count PET noise is non-Gaussian, non-stationary, and spatial-ly dependent. It scales directly with local activity and is shaped by iterative recon-struction and physical corrections. Standard denoising diffusion probabilistic models (DDPMs) ignore these PET properties. Their forward process adds iso-tropic, homoscedastic Gaussian noise to the target. Such an approach fails to cap-ture the realistic physical degradation generated by the imaging system. To ad-dress the above limitations, this study introduces a heteroscedastic residual diffu-sion model (HDDPM) for low-count brain PET recovery in which the forward corruption is itself intensity-aware. We designed a fixed, Poisson-based variance module to generate voxel-wise noise maps. These maps naturally place stronger noise perturbation on low-activity regions than high-activity ones, meanwhile the network predicts the low-to-standard-count residual under explicit dose-fraction conditioning. We evaluated our proposed model (HDDPM) alongside generative frameworks across three different scanners, using both internal and external da-tasets at various simulated dose levels (1% to 50%). HDDPM and isotropic DDPM showed comparable overall image quality, but HDDPM stood out in the lowest-dose (1%) external scans. It is highly reliable and significantly reduces measurement errors in both high- and low-activity regions, compared to the standard model. These results support that heteroscedastic noising with the pro-posed HDDPM is feasible, and it provides a physically motivated inductive bias for quantitative low-count PET recovery by reflecting the activity-dependent noise structure of PET.

Accurate and efficient parameter estimation is essential for applying electrochemical battery models in simulation, state estimation, control, and repeated model updating. However, conventional optimization methods, such as particle swarm optimization (PSO) and genetic algorithms (GA), often require many model evaluations and show considerable run-to-run variability, limiting their use in time-sensitive calibration scenarios. This study proposes a Batch-Optimized Local-to-Global Technique (BOLT) for rapid and robust parameter estimation of electrochemical battery models. BOLT combines diversified candidate initialization, batch-parallel trust-region reflective (TRF) local refinement, JIT-accelerated model evaluation, and multi-condition consistency screening within a unified calibration workflow. Comparative experiments based on a grouped single-particle model and measured data from a commercial 18650 NMC lithium-ion cell show that BOLT achieves a favorable trade-off among voltage-response accuracy, computational efficiency, and repeated-run stability. BOLT(32) achieves an average mean absolute error of \(12.4 \pm 0.1\) mV over five operating conditions, requiring only \(20636 \pm 3081\) model calls and \(8.97 \pm 1.20\) s per run. Synthetic-data validation with a known parameter vector in the grouped SPM formulation further shows that BOLT recovers the reference parameter vector under model-consistent conditions and remains robust under 1–3 mV voltage-noise perturbations, with the mean parameter absolute relative error below \(0.6\%\). These results indicate that BOLT provides a practical calibration framework for BMS parameter updating, control-oriented battery digital twins, and second-life battery screening.

Accurate correspondence matching across multiple angiographic views is the prerequisite for 3D coronary reconstruction and interventional guidance. However, the development of robust deep learning models for this task has been stifled by a fundamental data bottleneck. Obtaining ground truth for matching tasks in angiography pairs is prohibitively expensive and hard to scale. To overcome this barrier, we introduce a physically-grounded data generation framework that synthesizes high-fidelity Digital Reconstructed Radiographs (DRRs) from 3D Coronary CT Angiography (CCTA) volumes. Our framework generates dense, highly accurate 3D-to-2D projection labels by simulating realistic C-arm acquisition geometry on patient anatomy at zero human cost. Leveraging this dense supervision, we propose a Geometry-Informed Matching Module (GIMM) that integrates global feature and anatomical structure into correspondence learning. Unlike real angiography where assessment relies on subjective human annotation, our dataset provides 2D correspondence labels with paired images, allowing human-free evaluation. We comprehensively evaluate our method on the proposed CT-derived DRR dataset and demonstrate improvements over other matching baseline models.

Beamfocusing is the established near-field strategy for a large array serving a single-antenna user. We consider the single-user line-of-sight MIMO link, free of multipath, in which the user, too, carries an extended aperture, and show that the focusing prescription inverts: beyond a modest Fresnel number, focusing on the user is outperformed by far-field steering. Under fully analog, unit-modulus beamforming, we derive closed-form power gains for focusing (each aperture phase-matched to the other's center) and for steering (a planar phase ramp) in the Fresnel regime, and prove that their comparison is governed by two dimensionless quantities: the link Fresnel number, the product of the two aperture lengths normalized by wavelength and link distance, and the aperture ratio, irrespective of how many elements discretize the apertures. For equal apertures the two gains cross exactly once, at the universal value 1.947; beyond it, focusing loses ten dB per decade of Fresnel number, and the advantage celebrated in the MISO literature survives only as the receive aperture vanishes. We then derive the strategy that is order-optimal at every Fresnel number, con-focusing: both apertures aim at the common point from which they subtend equal angles. It attains the rank-one eigenbound in leading constant, needs no channel knowledge, degenerates to plain steering for equal apertures, and is acquirable within one beam-refinement round with no geometry exchange between the terminals.

Computed tomography (CT) plays a crucial role in medical diagnosis, but minimizing radiation exposure while maintaining image quality remains a critical challenge. Low-dose CT (LDCT) protocols reduce radiation risks but inevitably suffer from severe noise and artifacts that compromise diagnostic accuracy. While existing deep learning methods have achieved promising results, there remains a continuous quest for generative paradigms that intrinsically capture global-to-local structural dependencies to better preserve fine anatomical details. To this end, we propose DeVAR, a novel generative framework that applies visual autoregressive modeling (VAR) to LDCT denoising for the first time. Conditioned on global context provided by LDCT prefix tokens, DeVAR progressively generates discrete token maps of the target normal-dose CT (NDCT) via next-scale prediction. Because quantization inherently discards high-frequency information, we introduce a residual refiner to capture subtle anatomical structures beyond the capacity of a discrete codebook. Finally, empowered by a dual-representation hybrid training strategy, our hybrid NDCT decoder seamlessly integrates continuous and discrete latents to reconstruct high-fidelity, detail-preserved images. Extensive experiments on two public datasets demonstrate that DeVAR consistently achieves superior qualitative and quantitative performance compared to state-of-the-art LDCT denoising methods.

Diffusion posterior samplers for accelerated MRI can reconstruct accurately yet still disagree on the acquired k-space across samples, placing posterior variability on coefficients the scanner has already measured. We identify this measured-subspace leakage as a physical-admissibility failure. Under a hard-constraint model it violates the measurement constraint and inflates the reported uncertainty with disagreement about coefficients the scanner has already determined. To quantify this leakage, we introduce complementary measured- and unmeasured-subspace k-space dispersion metrics (MSD/USD). We then present Measured-Subspace Consistency (MSC), a training-free terminal correction that wraps any compatible image-space posterior sampler with a standard multi-coil consistency lock. The ideal lock follows classical range/null-space data consistency. Our contribution is to repurpose it as a black-box posterior audit and correction rather than a new reconstructor or learned sampler. Theoretically, we prove that the ideal transform confines pairwise sample differences to the MRI null space and bound the residual cross-subspace coupling left by practical sensitivity-weighted implementations. Across six base samplers and two MRI anatomies, including out-of-distribution transfer where a knee prior reconstructs brain, MSC substantially reduces measured-subspace dispersion for Soft samplers (a median 16.5x reduction for DPS across five brain contrasts, up to ~29x), while preserving unmeasured-subspace diversity and acting as a near-identity map for Consistent ones. Furthermore, MSC maintains or modestly improves PSNR/SSIM, with no retraining, retuning, or significant computational overhead.

Online latent state estimation constitutes a fundamental challenge within the artificial intelligence field, serving as a foundational tool for diverse applications, including sequential decision making, anomaly and change-point detection. In this paper, a novel online distributed sensing framework, where agents collaborate and exchange information to perform latent state estimation, is presented. The proposed estimator combines available partial domain knowledge with the representation capabilities of deep neural networks. In particular, the designed sensing framework incorporates prior estimates, optimized consensus weights, and Kalman-like recursive updates to perform decentralized inference, without relying on knowledge of noise statistics. Extensive experiments on linear, chaotic (Lorenz), and practical wireless tracking environments reveal that the proposed Covariance-Agnostic Neural Kalman Consensus Filter (CA-NKCF) outperforms traditional distributed Kalman and particle filters as well as purely model-free deep neural networks, exhibiting robustness even when the underlying motion and observation models are misspecified. It is also demonstrated that CA-NKCF's performance advantage remains stable across varying noise levels, random communication topologies, latent state dimensions, and observation clutter densities induced by scattering objects in wireless systems.

Fluorescence microscopy images are degraded by noise and diffraction-induced blur, which compromise structural fidelity and limit quantitative analysis. Supervised deep learning methods achieve impressive restoration performance but require large-scale paired datasets that are difficult to obtain in practice. To address this issue, we propose SDIP, a zero-shot deep image prior (DIP) framework that sequentially performs denoising and deconvolution without external training data. An aSeqDIP-based module first suppresses noise while preserving fine structures through sequential autoencoding regularization. In the deconvolution stage, a wavelet-based background correction step is incorporated before the proposed RLG-DIP module performs artifact-reduced deconvolution. RLG-DIP uses the Richardson-Lucy deconvolution result as a physically consistent guidance prior, integrating the imaging model with the implicit prior of DIP to stabilize the ill-posed deconvolution process. Experiments on the BioSR dataset across multiple cellular structures demonstrate that SDIP improves both signal-to-noise ratio and resolution, achieving superior visual quality and improved quantitative performance on most evaluated structures. The proposed framework may also provide useful insights for designing physically guided DIP methods for other inverse problems.

PAC-Bayesian bounds provide finite-sample guarantees for data-dependent randomized predictors, but applying them to learning-based control is difficult because the natural objective is a quadratic trajectory cost. Such losses are unbounded, non-Lipschitz , and lead to response-dependent Chernoff terms. We employ System Level Synthesis parameterization, which exposes the closed-loop trajectory map of a linear system directly and makes the quadratic control loss amenable to explicit certification. Moreover, we provide a set of PAC-Bayes-Chernoff certificates for posterior distributions over feasible closed-loop responses. For Gaussian disturbance trajectories with arbitrary covariance, we derive an exact one-sided Gaussian transform and a tractable quadratic upper bound expressed through closed-loop sensitivity quantities. We also derive a posterior-localized surrogate for settings where pointwise closed-loop response certificates are unavailable or have support related admissibility issues. Although PAC-Bayes certifies a non-degenerate posterior, the convex quadratic form of the SLS loss transfers the certificate to the posterior mean response. We present a deterministic mean response deployment result that is particularly suitable for control while retaining the stochastic posterior in the bound. Additionally, we provide a data-driven bound for this deployment, transitioning away from an oracle bound. Minimizing this bound naturally results in a learning algorithm for control selection from data. Numerical experiments on a double integrator show that the algorithm acts as a sensitivity-aware finite-sample regularizer, improving held-out cost and reducing closed-loop sensitivity in the low-data regime

Recently, Large Language Model (LLM)-based Text-to-Speech (TTS) models have achieved remarkable naturalness. However, the standard Supervised Fine-Tuning paradigm often converges to statistically averaged prosody, limiting emotional expressiveness. While preference-driven optimization offers a promising alternative, existing approaches suffer from two structural mismatches: information conflict, where content and emotion in a shared latent space produce conflicting gradients, leading to reward hacking and semantic degradation; and scale gap, where sparse sentence-level rewards struggle to guide dense frame-level generation. To overcome these challenges, we propose HPRO, a hierarchical progressive reward optimization framework. Within HPRO, we introduce the HD-Emo codec as a novel differentiable reward model to mitigate the information conflict. It extracts speech into distinct content and style preference tokens, structurally isolating emotional optimization from semantic content. Building upon this structured preference space, HPRO bridges the scale gap by progressively aligning frame-, word- and sentence-level objectives. Experiments demonstrate that HPRO significantly enhances emotional expressiveness, while effectively preserving linguistic intelligibility. The code and audio samples are publicly available at https://xxh333.github.io/hpro-demo/.

Temporal link prediction (TLP) is typically evaluated by predictive performance on unseen edges, but this criterion can conflate predictive accuracy with recovery of the underlying causal mechanism. In stochastic models, Fisher information governs the Cramér–Rao (CR) bound on parameter estimation error: higher Fisher information permits more accurate parameter recovery. We show that, under comonotonicity conditions between Fisher information and entropy, binary logistic models exhibit an estimation–prediction tradeoff: regimes with higher Fisher information, and hence smaller CR bounds, also have higher irreducible predictive entropy. To study this tradeoff in TLP, we introduce a probabilistic causal generator for temporal graphs with transient edges and known ground-truth causal structure, and validate the phenomenon empirically.

EEG-based emotion recognition is widely used in affective computing but suffers from poor generalization due to domain shifts caused by inter-subject variability, dataset differences, and recording conditions, especially in cross-dataset settings. Conventional unsupervised domain adaptation methods require source data, which is often unavailable due to privacy constraints. Although source-free UDA addresses this limitation, existing methods still struggle with large domain gaps, noisy pseudo-labels, and unstable adaptation. To address these challenges, we propose an enhanced source-free unsupervised domain adaptation (SF-UDA) framework for cross-dataset EEG emotion recognition. The framework introduces a non-contrastive predictive coding-based self-supervised pretraining strategy to learn robust and transferable EEG representations by modeling temporal dependencies in a reconstruction-based manner. During adaptation, we estimate target-domain structure through class-wise clustering and prediction disagreement, and optimize the model using a dual-stage strategy consisting of Multi-Loss Adaptive Regularization and Localized Consistency Learning, improving stability and neighborhood consistency under noisy pseudo-labels. We also propose a lightweight test-time training mechanism that enables selective online updates for uncertain samples using predictive reconstruction loss and entropy minimization. Experiments on DEAP, SEED, and DREAMER show consistent improvements over state-of-the-art SF-UDA methods, achieving 69.56% and 63.03% accuracy on SEED and DREAMER when trained on DEAP, and 61.38% and 68.90% when trained on SEED.

Implicit neural representations (INRs) have recently emerged as a promising approach to video compression, delivering competitive rate-distortion performance alongside rapid decoding. However, existing neural video codecs struggle to balance complexity and scalability. Lightweight models often suffer from degraded compression performance when scaled to different bitrate/quality levels, whereas high-performance models exhibit limited scalability, as their model complexity typically increases with quality. This lack of a unified architecture capable of maintaining consistent complexity across a wide range of bitrates severely limits their diverse real-world deployment. To address these challenges, we introduce NVRC++, a novel INR-based video codec that utilizes a lightweight INR with multiple high-resolution feature grids, providing high scalability at any given complexity level. This is paired with an optimization framework that enables efficient overfitting on high-resolution grids for long video sequences, thereby exploiting spatio-temporal redundancies without prohibitive computational or memory overhead. Additionally, an advanced entropy model is designed for efficiently compressing the high-dimensional grid parameters. As a result, NVRC++ provides four complexity levels (from 7kMACs/pixel to 360kMACs/pixel), each spanning wide bitrate and quality ranges while supporting real-time decoding. The experimental results show that NVRC++ offers a much faster decoding speed (up to 8.5x) compared to the SOTA INR-based video codec, NVRC, while delivering comparable performance.

Physics-informed neural networks (PINNs) have recently emerged as a promising framework for addressing the Calderón inverse problem from limited boundary data. In this work, we revisit neural Calderón inversion by introducing multiscale boundary excitations based on randomized wavelet functions and investigating the role of Fourier-feature encoding (FFE) for representing sharp conductivity variations. We propose a physics-informed reconstruction framework that represents the unknown conductivity and the associated family of electric potentials with separate neural networks conditioned on the applied boundary excitations. The governing elliptic PDE is enforced through physics-informed residuals, while finite Dirichlet-to-Neumann (DtN) data are incorporated through boundary losses. Using synthetic data from a finite-difference forward solver, we evaluate the method on conductivity fields with inclusions, sharp interfaces, smooth profiles, and heterogeneous media. Results show that the framework recovers dominant conductivity structures from finite boundary measurements with relative errors between $3\%-12\%$ approximately. We show that FFE improves the reconstruction of localized sharp features, particularly for inclusions and interfaces, but are not universally optimal, with raw-coordinate networks performing competitively for smoother fields. These results highlight coordinate representations and boundary excitation design as key factors in neural Calderón inversion.

We address the problem of robust controller synthesis for a class of linear temporal logic (LTL) specifications over families of perturbed systems using symbolic control techniques. Given a dynamical system, a specification, and a symbolic controller synthesized using the fixed-point algorithm of the specification, the objective is to find the maximal perturbation we can apply to the system while the system continues to satisfy the same specification under the same controller. We first provide general results, by demonstrating that controllers synthesized based on the symbolic model can be refined back to a perturbed version of the concrete system while preserving their correctness. Focusing on four fundamental temporal logic specifications, namely safety, reachability, persistence, and recurrence, we introduce a general measure of the maximal robustness margin. Then, for each class of specifications, we derive a customized version of the measure and establish the corresponding theoretical guarantees. Importantly, the robustness margin depends explicitly on the sequence of sets generated during the fixed-point computation, allowing for specification-dependent and less conservative bounds compared to generic abstraction-based approaches. The theoretical developments are illustrated on two examples, demonstrating the practical applicability and effectiveness of the proposed approach.

Subjective evaluation remains the most reliable way of testing speech and audio coding techniques. Crowdsourcing the listening task is a cost-efficient and fast way of conducting this evaluation, but the quality of the results tends to be inferior to that of conventional listening tests done in the controlled environment of a laboratory. In this paper, classical and neural speech codecs are evaluated to compare P.808 against P.800 DCR tests. A statistical analysis is conducted to investigate the effectiveness of selected screening methods. The analysis shows that the crowdsourced evaluation can be improved by employing postscreening methods based on anchor ordering and rating span, and continuous screening methods like traps and gold standard questions, thus giving more value to the ratings obtained for the codecs under test. Based on these outcomes, a set of suitable screenings is proposed, for cost-effective, simplified, and bias-free enhancement of listening results.

The driveability of a new heavy-truck driveline is traditionally assessed using physical prototypes. Enabling early evaluation of the driving experience in a human-in-the-loop driving simulator using a virtual prototype has the potential to significantly improve development efficiency. To enable driveability assessment using a moving-base simulator, participants must be able to perceive small differences in longitudinal acceleration. The just-noticeable difference (JND) was therefore evaluated for two variants of the classical motion-cueing algorithm (MCA) tuned specifically for tip-in/launch tests and compared to a more general variant in a driving simulator with a long linear track. Psychometric functions were fitted to responses obtained using a weighted staircase procedure and analysed using a generalized linear model. No significant differences in JND were found between the motion cueing variants. The mean JND across all participants and MCA variants was 5.4%. The mean point of subjective equality in the JND experiment was -1.9%, suggesting that participants perceived the acceleration as higher in the second stimulus of a pair. In a subjective comparison, most participants preferred the motion cueing variants that were tuned for launch manoeuvres over the general variant.

Insurance fraud remains costly and operationally difficult, particularly in call-centre workflows where many customer interactions begin at FNOL. While recent fraud detection methods mainly rely on structured data, text, or images, repeated speaker identity across calls remains underused as an investigative signal. This paper presents DG^VoiC, a voice clustering framework for customer verification and cross-profile speaker linking on anonymised real call-centre audio. The approach combines sensitive information-aligned anonymisation, speech-focused preprocessing, sliding-window speaker embedding extraction, and cosine similarity based clustering to identify repeated speakers under real telephony conditions. The method was evaluated on 121 recordings, with a curated reference subset of 56 samples in 22 human-agreed speaker clusters. used for validation. The best configuration achieved 96% AMI, 95% ARI, 98% completeness, 100% homogeneity, and 99% V-measure. These results show that speaker clustering can provide a strong additional signal for fraud investigation by helping analysts verify speaker consistency and surface repeated voices across customers.

Ensuring reliable communication for mission-critical vehicles in dynamic environments with limited infrastructure is a significant challenge due to interference and spectrum scarcity. This paper investigates a UAV-assisted vehicular communication framework that leverages multi-antenna beamforming and dynamic bandwidth allocation to provide prioritized and interference-mitigated wireless links. Vehicles are classified according to their service priority, with each class assigned a distinct frequency band to reduce interference. Within each class, optimized beamforming further minimizes transmission overlap and enhances spectral efficiency. The optimization problem is solved using an alternating optimization framework, incorporating two beamforming strategies: one based on successive convex approximation (SCA) and the other derived in closed form. Numerical results indicate that the proposed scheme outperforms baseline approaches that optimize only bandwidth allocation or beamforming in terms of overall system performance. Among the two joint optimization methods, the closed-form solution achieves higher sum rates and generally requires less transmit power, while also exhibiting lower computational complexity compared with the SCA-based approach.

Symbolic music alignment links notes in a symbolic performance to their counterparts in a score. While existing alignment encoding formats provide unique correspondences between these notes, there are various musical practices and forms such as practice repetitions in rehearsal and improvised realizations in basso continuo that require a more flexible approach to encoding their alignments. In this paper, we propose a minimal, backward-compatible extension to the Match file format to support such non-unique and semantically complex alignments. We introduce two virtual pointer notes - virtual score notes and virtual performance notes - which allow to encode multiple links between performance and score notes. In addition we expand the Match file's 'section' line to include semantically meaningful annotations of performance regions beyond score-indicated musical repetitions. We further demonstrate the utility of these extensions through two representative use-cases in piano rehearsal and basso continuo.

Neural video codecs have surpassed classical codecs in coding efficiency but remain impractical for deployment due to cross-platform incompatibility and high computational cost. Existing quantization-based solutions fail to produce deterministic results across diverse hardware platforms, leading to catastrophic decoding failures. We introduce MLVC, a hardware-robust neural video codec designed for practical cross-platform inference. The key idea is to explicitly transmit scale parameters through the hyperprior, which guarantees entropy coding consistency across devices without requiring bit-exact arithmetic. While this increases bitrate overhead, we recover most of the coding efficiency through architectural improvements (gated memory, ReGLU activation), a long-term reference recovery mechanism, and domain-specific perceptual training. On the VCD video conferencing benchmark, MLVC achieves >70% BD-rate (MOS) improvement over hardware HEVC, the strongest deployable baseline, while reaching subjective quality competitive with DCVC-RT, which cannot operate across diverse platforms. Both the encoder and decoder run at 100 FPS on average on commodity NPUs from Apple, Intel, and Qualcomm. MLVC is the first neural video codec to combine competitive compression performance, real-time speed, and cross-platform robustness across diverse consumer devices, making it suitable for widespread deployment. Code is available at https://github.com/microsoft/mlvc.

The growing penetration of inverter-based resources (IBRs) necessitates stability assessment methods that are scalable, decentralized, and model-agnostic. This paper develops a block diagonal dominance (BDD) criterion for decentralized small-signal stability of IBR-dominated power grids. The proposed approach forms the basis for an enhanced IBR connection compliance condition from a small-signal stability perspective that can be evaluated locally for IBRs to be connected to the grid. The proposed approach is shown to be much less conservative than strict diagonal dominance (SDD). Beyond mere stability, we ensure a minimum decay rate or maximum settling time for IBR-induced oscillation. Crucially, these are achieved without imposing restrictive assumptions on network or IBR models. The framework therefore, offers a practical and theoretically grounded basis for decentralized stability certificate of IBR-dominated power grids.

We propose an agentic Large Language Model (LLM) framework for active Fault-Tolerant Control (FTC) that transforms fault detection outputs into constraint-aware recovery actions grounded in plant-specific knowledge. The approach couples (i) a multi-agent workflow that decomposes operator duties into monitoring, planning, action synthesis, simulation, validation, and reprompting; (ii) a Digital Process Plant Twin (DPPT) that exposes plant data, models, and a simulation service for pre-execution testing; and (iii) a Graph Retrieval-Augmented Generation (Graph RAG) layer built on the CPSMod ontology, which organizes plant knowledge (structure, function, hybrid dynamics, control context, and fault semantics) into a graph that supports relation-aware, multi-hop retrieval for the agents. Corrective actions are generated as minimal-risk state-machine recovery paths and corresponding discrete commands or continuous setpoint adaptations, then validated deterministically against interlocks, envelopes, and dynamic feasibility before any actuation. If no acceptable plan is found within a bounded time window, control is handed to a safety fallback. The framework is evaluated in simulation on two representative benchmarks: a discrete batch Mixing Module and a Continuous Stirred-Tank Reactor (CSTR) under closed-loop PID regulation. Results with lightweight LLMs (GPT-4o-mini and GPT-4.1-mini) show that semantically grounded agents can derive valid recovery decisions within latency budgets compatible with the respective process dynamics, demonstrating a practical pathway from detection to validated corrective action across both discrete and continuous FTC tasks.