Papers

Classical guidance laws aim at minimizing the miss distance, thus implicitly determining the minimum warhead lethality radius required against nominal targets. However, nonnominal targets or scenarios might render the designed warhead insufficient, causing a significant degradation in the single-shot kill probability (SSKP). We propose a guidance methodology that shifts the interceptor's objective from minimizing the miss distance to directly maximizing the SSKP, while taking into account the warhead's probabilistic lethality model. Complying with the generalized separation theorem, the new paradigm is based on modifying deterministic differential-game-based guidance laws using Bayesian decision theory. Extensive Monte Carlo simulations demonstrate consistent SSKP improvement over the standard and recently introduced estimation-aware guidance laws, when tested against nominal and nonnominal evasively maneuvering targets.

Generative Semantic Communication (GSC) is a promising solution for image transmission over narrow-band and high-noise channels. However, existing GSC methods rely on long, indirect transport trajectories from a Gaussian to an image distribution guided by semantics, causing severe hallucination and high computational cost. To address this, we propose a general framework named Schrödinger Bridge-based GSC (SBGSC). By leveraging the Schrödinger Bridge (SB) to construct optimal transport trajectories between arbitrary distributions, SBGSC breaks Gaussian limitations and enables direct generative decoding from semantics to images. Within this framework, we design Diffusion SB-based GSC (DSBGSC). DSBGSC reconstructs the nonlinear drift term of diffusion models using Schrödinger potentials, achieving direct optimal distribution transport to reduce hallucinations and computational overhead. To further accelerate generation, we propose a self-consistency-based objective guiding the model to learn a nonlinear velocity field pointing directly toward the image, bypassing Markovian noise prediction to significantly reduce sampling steps. Simulation results demonstrate that DSBGSC outperforms state-of-the-art GSC methods, improving FID by at least 38% and SSIM by 49.3%, while accelerating inference speed by over 8 times.

The accelerating advancement of intelligent transportation systems has established consumer-oriented vehicular networks (CVNs) as a critical infrastructure for next-generation connected mobility. However, the high mobility of vehicular users (VUs) introduces significant channel state information (CSI) uncertainty, which severely undermines the performance of conventional fixed-position antenna systems. To address this, this paper explores the deployment of movable-antennas (MAs) to enhance communication robustness in CVNs under imperfect CSI conditions. We develop a joint optimization framework that dynamically coordinates the spatial positioning of MAs and transmit beamforming at the base station, with the objective of maximizing the worst-case sum rate across all VUs. The problem is formulated as a non-convex max-min optimization problem, subject to bounded CSI estimation errors, transmit power limits, and physical constraints on antenna displacement. By adopting an alternating optimization strategy, the original problem is decomposed into tractable subproblems, solved via techniques including the S-Procedure, Schur complement, and successive convex approximation. Numerical evaluations confirm that the proposed approach achieves substantial gains over existing benchmarks in terms of worst-case throughput.

Low-altitude communication networks (LACNs) serve as the critical infrastructure of the emerging low-altitude economy (LAE), supporting services such as drone delivery and infrastructure inspection. However, LACNs operate in highly dynamic three-dimensional (3D) environments characterized by high mobility and predominantly line-of-sight (LoS) propagation, creating strong coupling among key performance objectives including coverage, interference mitigation, handover management, and sensing capability. Isolated tuning of individual objectives cannot capture these cross-objective interactions, rendering conventional approaches based on experience-driven tuning and repeated field trials inefficient and costly. To address these challenges, we propose DT-MOO, a Digital Twin-based Multi-Objective Optimization framework for LACNs. By constructing a high-fidelity virtual replica that integrates realistic environmental models, electromagnetic (EM) propagation, and traffic dynamics within a unified environment, DT-MOO enables joint evaluation and systematic optimization of interdependent network parameters, scoring candidate configurations by their combined effect on multiple objectives. As the foundational validation of the framework, we report real-world experiments in a 5G-enabled LACN focusing on coverage-interference co-optimization, where DT-MOO increases the high-quality coverage rate from 14.0% to 52.9% across all evaluated altitudes compared to an operator-provisioned, experience-based baseline, while achieving a net SINR gain under stringent criteria despite local spatial trade-offs, confirming its ability to handle coupled objectives in practical LACN deployment.

We present a trajectory-based optimization framework for arrival sequencing and scheduling in the terminal maneuvering area (TMA). Unlike node-link scheduling models that reduce trajectories to time-delay variables, the proposed method computes implementable per-aircraft speed profiles and path extensions that achieve required landing separation through terminal air traffic control actions. The framework combines an analytic TMA path model, consisting of a tangent leg, a radius-to-fix turn, and a final-approach segment, with a nonlinear program (NLP) that jointly optimizes path stretch and segment speeds under a weighted objective. Three landing-order policies are examined: First-Entry-First-Serve (FEFS), First-on-Final-First-Serve (FOFFS), and FOFFS with Constrained Position Shifting (CPS) up to $k$ positions. CPS is implemented through a two-phase approach coupling mixed-integer linear programming (MILP) with NLP to select an optimized landing order before trajectory optimization. The aircraft population follows a realistic weight-class fleet mix with pair-specific wake-turbulence separation, and each scenario is perturbed by a Gaussian wind sample projected onto each segment to convert commanded airspeeds into ground speeds. An online rolling-horizon formulation commits each aircraft trajectory irrevocably upon entry, enabling real-time decision-making. Monte Carlo experiments on the simplified A80 TMA show that: (i) FOFFS consistently outperforms FEFS in delay, path stretch, and fuel burn by exploiting geometric asymmetries among arrival streams; (ii) CPS further reduces separation violations and path stretch, though with diminishing returns and rapidly increasing solver cost; (iii) fuel estimates from BADA 3 and OpenAP show consistent qualitative trends; and (iv) per-entry optimization completes in near real-time, supporting practical deployment.

Continuous-control reinforcement learning (RL) often exhibits large closed-loop variance, high-frequency control jitter, and sensitivity to disturbance injection. Existing explanations usually emphasize disturbance sources such as action noise, exploration perturbations, or policy nonsmoothness. This letter studies a complementary amplifier-side perspective: in nominally stable yet strongly non-normal closed loops, small input perturbations can undergo transient amplification and lead to disproportionately large state covariance. Motivated by this source–amplifier decomposition, we introduce an input-side variance suppression layer that operates between the learned policy and the plant input to reduce applied-input variance and step-to-step jitter. To separate mechanism from correlation, we use two control-theoretic interventions: one varies only eigenvector geometry under fixed eigenvalues and spectral radius, and the other varies only applied-input statistics under fixed strongly non-normal geometry. We then provide mechanism-consistent external validation on planar quadrotor tasks. Throughout, Koopman/ALE surrogates are used only as analysis and certification tools, not as direct performance paths. Taken together, the results support a narrower claim: in the studied settings, non-normal transient amplification is an important and under-emphasized contributor to execution-time closed-loop variance, and source-side suppression can reduce downstream covariance without changing the structural peak gain.

Reconfigurable intelligent surfaces (RISs) modify signal reflections to enhance wireless communication capabilities. Classical RIS phase optimization is highly non convex and challenging in dynamic environments due to high interference and user mobility. Here we propose a hierarchical multi-objective quantum metalearning algorithm that switches among specific quantum paths based on historical success, energy cost, and current data rate. Candidate RIS control directions are arranged as switch paths between quantum neural network layers to minimize inference, and a scoring mechanism selects the top performing paths per layer. Instead of merely storing past successful settings of the RIS and picking the closest match when a new problem is encountered, the algorithm learns how to select and recombine the best parts of different solutions to solve new scenarios. In our model, high-dimensional RIS scenario features are compressed into a quantum state using the tensor product, then superimposed during quantum path selection, significantly improving quantum computational advantage. Results demonstrate efficient performance with enhanced spectral efficiency, convergence rate, and adaptability.

We study the problem of online non-stochastic control (ONC), which is the control of a linear system under adversarial disturbances and adversarial cost functions, with the aim of minimizing the total cost incurred. A recent line of literature in ONC develops algorithms that enjoy sublinear regret with respect to a benchmark based on the set of steady-states that are attainable by a constant input. In this work, we extend this research direction by giving an algorithm that enjoys $\mathcal{O}(\sqrt{T})$ regret with respect to a richer benchmark set, namely the set of steady-states attainable under an affine controller. Since this benchmark substantially broadens the comparison class, it provides significantly stronger performance guarantees. Our proposed algorithm combines a Follow-The-Perturbed-Leader-style online non-convex optimization approach with a batching method that maintains stability despite changing policies. Although our proposed algorithm requires solving non-convex subproblems, we show that an approximate solution to this subproblem is sufficient to ensure $\mathcal{O}(\sqrt{T})$ regret. Furthermore, numerical experiments show that our algorithm enjoys lower total cost and similar computation to existing methods in certain settings.

The success of large foundation models is catalyzing a new paradigm for AI-native 6G network design: wireless foundation models for physical-layer design. However, existing models often operate on channel state information (CSI) in the spatial-temporal-frequency (STF) domain, where multipath components are superimposed and structurally entangled. This hinders the learning of a universal channel representation. Their reliance on global attention also incurs prohibitive overhead. In this paper, we propose AirFM-DDA, an Air-interface Foundation Model in the Delay-Doppler-Angle (DDA) domain. AirFM-DDA reparameterizes CSI into the DDA domain to resolve multipath components along physically meaningful axes and employs window-based attention with frame-structure-aware positional encoding. Extensive experiments demonstrate transferability across scenarios, tasks, datasets, and antenna configurations. For channel prediction and estimation, AirFM-DDA generalizes zero-shot to unseen cities, achieving average normalized mean-square error (NMSE) gains of 4.9-8.5 dB over the strongest baselines. With only 10% labeled data, it achieves average gains of 12.0 percentage points in Top-1 accuracy for beam prediction and 3.4 percentage points in F1 score for line-of-sight (LoS) identification. It further transfers across simulated datasets and adapts to measured data and different antenna arrays. Compared with global attention, window-based attention reduces training and inference costs by nearly an order of magnitude.

Data-driven MoCap-to-radar models generate plausible micro-Doppler spectrograms, but do they actually learn the underlying physics? We introduce a physics-based interpretability framework to answer this question via two proposed complementary metrics: one measures alignment between model predictions and the physics-derived Doppler frequency, while the other tests whether predictions preserve the velocity-frequency relationship under velocity intervention. Both metrics require only MoCap input and model predictions, without access to measured radar data. Experiments across several model architectures reveal that low reconstruction error does not guarantee physical consistency: some, but not all, models achieve low error yet perform poorly on the two physics-based metrics. Further analysis shows that temporal attention is critical for transformer-based models to learn the underlying physics.

In this work, we introduce a paralinguistic supervision paradigm for low-resource multilingual speech emotion recognition (LRM-SER) that leverages non-verbal vocalizations to exploit prosody-centric emotion cues. Unlike conventional SER systems that rely heavily on labeled verbal speech and suffer from poor cross-lingual transfer, our approach reformulates LRM-SER as non-verbal-to-verbal transfer, where supervision from a labeled non-verbal source domain is adapted to unlabeled verbal speech across multiple target languages. To this end, we propose NOVA ARC, a geometry-aware framework that models affective structure in the Poincaré ball, discretizes paralinguistic patterns via a hyperbolic vector-quantized prosody codebook, and captures emotion intensity through a hyperbolic emotion lens. For unsupervised adaptation, NOVA-ARC performs optimal transport based prototype alignment between source emotion prototypes and target utterances, inducing soft supervision for unlabeled speech while being stabilized through consistency regularization. Experiments show that NOVA-ARC delivers the strongest performance under both non-verbal-to-verbal adaptation and the complementary verbal-to-verbal transfer setting, consistently outperforming Euclidean counterparts and strong SSL baselines. To the best of our knowledge, this work is the first to move beyond verbal-speech-centric supervision by introducing a non-verbal-to-verbal transfer paradigm for SER.

In this study, we present Healthcare Codec-Fake Detection (HCFD), a new task for detecting codec-fakes under pathological speech conditions. We intentionally focus on codec based synthetic speech in this work, since neural codec decoding forms a core building block in modern speech generation pipelines. First, we release Healthcare CodecFake, the first pathology-aware dataset containing paired real and NAC-synthesized speech across multipl clinical conditions and codec families. Our evaluations show that SOTA codec-fake detectors trained primarily on healthy speech perform poorly on Healthcare CodecFake, highlighting the need for HCFD-specific models. Second, we demonstrate that PaSST outperforms existing speech-based models for HCFD, benefiting from its patch-based spectro-temporal representation. Finally, we propose PHOENIX-Mamba, a geometry-aware framework that models codec-fakes as multiple self-discovered modes in hyperbolic space and achieves the strongest performance on HCFD across clinical conditions and codecs. Experiments on HCFK show that PHOENIX-Mamba (PaSST) achieves the best overall performance, reaching 97.04 Acc on E-Dep, 96.73 on E-Alz, and 96.57 on E-Dys, while maintaining strong results on Chinese with 94.41 (Dep), 94.40 (Alz), and 93.20 (Dys). This geometry-aware formulation enables self-discovered clustering of heterogeneous codec-fake modes in hyperbolic space, facilitating robust discrimination under pathological speech variability. PHOENIX-Mamba achieves topmost performance on the HCFD task across clinical conditions and codecs.

This paper explores the integration of reconfigurable intelligent surfaces (RISs) into cell-free massive multiple-input-multiple-output (CF-mMIMO) networks operating in FR1 and FR3 frequency bands. We present a comprehensive framework for analyzing RIS-assisted CF-mMIMO systems under realistic propagation conditions, accounting for frequency-dependent characteristics and RIS configurations. A novel RIS-user association algorithm is proposed to optimize phase-shift settings by assigning each RIS to a single user based on line of sight (LoS) connectivity. The system model incorporates spatially correlated Ricean fading channels and employs scalable partial-minimum mean square error (P-MMSE) combining. The numerical results demonstrate that the proposed RIS-user selection strategy significantly improves the spectral efficiency compared to random or exhaustive RIS configurations, particularly when the number of RISs is moderate. We also analyze the trade-off between training overhead and performance gains, showing that excessive pilot requirements can offset benefits when RIS density or element count increases. The results highlight the potential of the FR3 bands for RIS-assisted CF-mMIMO, provided advanced channel estimation techniques are adopted to mitigate overhead. These findings emphasize the importance of intelligent RIS-user pairing and scalable estimation methods for future 6G deployments.

This paper studies scalable conjugate beamforming (CB) variants for physical-layer multicasting in cell-free massive multiple-input multiple-output (CF-mMIMO) systems. Focusing on fully distributed precoding, we analyze classical CB, normalized CB (NCB), and enhanced CB (ECB) within a subgroup-centric multicast framework. Multicast users are partitioned into subgroups based on large-scale fading similarity, which enables composite channel estimation, pilot reuse, and distributed precoding with low complexity. The performance of the different CB variants is evaluated in terms of aggregated spectral efficiency (ASE) under representative user geometries, including uniformly distributed users, spatially clustered deployments, and heterogeneous scenarios combining hotspots with more dispersed users. Monte Carlo simulations reveal a strong spatial geometry-dependent behavior: unicast transmission is preferable in uniform deployments, while subgroup-based multicasting becomes essential in clustered and heterogeneous scenarios. Among the CB-based precoders, NCB offers a robust performance-complexity trade-off across most scenarios, whereas ECB provides additional gains only when sufficient channel hardening is present. These results provide practical insights into the selection of low-complexity distributed precoders and multicast transmission modes in CF-mMIMO systems supporting broadband and multimedia services.

Future inverter-dominated power systems feature higher variability and more stressed operating conditions, which motivates the consideration of stability in operational settings. Existing approaches to stability-constrained OPF often rely on eigenvalue calculation, global model information, or dynamic evaluation inside optimization formulation, which are computationally intensive and difficult to scale. This paper proposes the first decentralized stability-constrained OPF framework for inverter-based power systems. The key novelty lies in the incorporation of a class of algebraic decentralized small-signal stability criteria that admits tractable representations in steady-state variables and is therefore suitable for optimization. The decentralized stability condition is based on local voltage differences and enables clear theoretical and practical economic interpretation of the stability contribution from each inverter. We define a Nodal Stability Shadow Price (NSSP) for each inverter, and characterize the role of these stability constraints through their associated shadow prices, enabling a nodal interpretation of their economic impacts. It is proved that under active-power-only objectives in lossless networks, binding stability constraints may occur but will admit zero shadow prices if all other operational constraints are inactive. Most importantly, we reveal the importance of considering the opportunity cost of reactive power for inverter-based resources (IBRs) that have limited capacity. When reactive power costs are considered, stability constraints can carry strictly positive shadow prices and admit meaningful economic impacts.

Financial Transmission Rights (FTRs) enable electricity market participants to hedge congestion risk in Day Ahead Market (DAM) operations, but for the market to be solvent, Independent System Operators (ISOs) must ensure that FTR payouts do not exceed the collected DAM merchandising surplus that funds them. We show that FTR underfunding (or conversely, hedging efficiency) can arise structurally from misalignment between the network models used in the FTR auction and the DAM, independent of bidding behavior. We develop a geometric framework in which both DAM merchandising surplus and the maximum supportable FTR payout are expressed as support functions of network-feasible injection polytopes. The resulting dual representation assigns nonnegative weights to transmission element-contingency constraints, enabling constraint-level attribution of model misalignment. Using this framework, we derive sharp implications for canonical FTR network modeling choices like uniform transmission element derates, and for structural sources of underfunding like unplanned DAM outages. We further show that multi-interval FTR products impose an intrinsic hedging inefficiency when DAM shadow prices vary over time, even under perfect model alignment. These results provide ISOs with rigorous tools to diagnose underfunding and quantify the efficiency cost of conservative FTR network modeling choices.

This paper develops an active sensing framework for designing the transmit and receive beamformers of a multiple-input multiple-output (MIMO) radar system with a limited number of radio-frequency (RF) chains. In the proposed framework, the transmit and receive beamformers are adaptively designed in each sensing stage based on the measurements made in the previous sensing stages, and the optimization of the beamformers are based on the minimization of the Bayesian Cramér-Rao bound (BCRB) for estimating the unknown sensing parameters. To address the exploration-exploitation tradeoff inherent in such an adaptive procedure, this paper proposes a two-phase design involving two variants of the BCRB optimization problem: an exploration-centric variant that requires multiple orthogonal beamforming directions to be probed in each sensing stage, followed by an exploitation-centric variant that optimizes the per-stage BCRB without such requirement. For both variants, the transmit and receive beamformers are optimized as two subproblems in an alternating fashion in each sensing stage. The alternating optimization is shown to converge to a stationary point of the per-stage BCRB optimization provided that each subproblem is solved to global optimality. A main contribution of this paper is to determine sufficient conditions under which global optimality can be guaranteed—despite that the two subproblems are nonconvex (due to the limited number of RF chains), by studying the structures of their optimal solutions via Lagrangian dual analysis. The conditions pertain to the multiplicity of the eigenvalues of specific direction matrices that can be analytically written in terms of the optimal dual variables. Simulation results demonstrate the advantages of the proposed design framework as compared to state-of-the-art adaptive beamforming strategies.

Modern maritime navigation and control systems rely on digital sensing, estimation, and communication pipelines that fuse GNSS, radar, inertial, and AIS data through approaches such as Kalman-filter-based estimators. While these technologies are essential for safety and efficiency, their growing interconnection also exposes vessels to faults and cyber-physical anomalies. This paper introduces a Statistical Detection Suite (SDS) to detect malicious stealthy disturbances. Specifically, the SDS operates directly on the innovations of Kalman filters, providing a lightweight yet statistically grounded layer of anomaly monitoring within maritime estimation frameworks. The SDS jointly evaluates whitened innovations through four complementary checks: (i) bias, (ii) covariance consistency via the normalized innovation squared (NIS), (iii) Gaussianity, and (iv) temporal independence via portmanteau statistics. The analysis further examines how an adversary can craft stealthy finite-impulse-response (FIR) Gaussian disturbances that can evade classical chi-square checks, formulating an optimization-based design that balances stealth and trajectory impact. An evaluation in maritime navigation scenarios illustrates how the SDS exposes colored spoofing attacks that bypass traditional methods, highlighting the role of innovation-based monitoring in strengthening maritime resilience against cyber-physical threats.

Multi-camera systems are widely employed in sports to capture the 3D motion of athletes and equipment, yet calibrating their extrinsic parameters remains costly and labor-intensive. We introduce an efficient, tool-free method for multi-camera extrinsic calibration tailored to sports involving stick-like implements (e.g., golf clubs, bats, hockey sticks). Our approach jointly exploits two complementary cues from synchronized multi-camera videos: (i) human body keypoints with unknown metric scale and (ii) a rigid stick-like implement of known length. We formulate a three-stage optimization pipeline that refines camera extrinsics, reconstructs human and stick trajectories, and resolves global scale via the stick-length constraint. Our method achieves accurate extrinsic calibration without dedicated calibration tools. To benchmark this task, we present the first dataset for multi-camera self-calibration in stick-based sports, consisting of synthetic sequences across four sports categories with 3 to 10 cameras. Comprehensive experiments demonstrate that our method delivers SOTA performance, achieving low rotation and translation errors. Our project page: https://fandulu.github.io/sport_stick_multi_cam_calib/.

Machine learning is becoming increasingly important for nonlinear system identification, including dynamical systems with spatially distributed outputs. However, classical identification and forecasting approaches become markedly less reliable in turbulent-flow regimes, where the dynamics are high-dimensional, strongly nonlinear, and highly sensitive to compounding rollout errors. Diffusion-based models have recently shown improved robustness in this setting and offer probabilistic inference capabilities, but many current implementations inherit target parameterizations from image generation, most commonly noise or velocity prediction. In this work, we revisit this design choice in the context of nonlinear spatiotemporal system identification. We consider a simple, self-contained patch-based transformer that operates directly on physical fields and use turbulent flow simulation as a representative testbed. Our results show that clean-state prediction consistently improves rollout stability and reduces long-horizon error relative to velocity- and noise-based objectives, with the advantage becoming more pronounced as the per-token dimensionality increases. These findings identify target parameterization as a key modeling choice in diffusion-based identification of nonlinear systems with spatial outputs in turbulent regimes.

Spatial correlation poses a significant challenge in massive multiple-input multiple-output (MIMO) high-altitude platform station (HAPS) systems. The inherent spatial correlation among antenna elements on the HAPS induces high correlation and interference among users' channel gains. To mitigate this issue, we propose an integrated approach that combines spatial resource allocation and user clustering. In our proposed solution, we assign the same resource blocks to users with orthogonal channel gains, while users with non-orthogonal channel gains receive different resource blocks. Additionally, we propose a sectorized antenna architecture for the massive MIMO HAPS base station, specifically designed to directly transmit three-dimensional beams to users and reduce spatial correlation among antenna elements. This work addresses the joint optimization problem of power allocation and resource allocation to maximize the overall data rate of the massive MIMO HAPS system. Simulation results revealed the role of spatial resource allocation in managing spatial correlation and interference among users.

Medical imaging AI development is fundamentally dependent on annotated datasets, yet no existing standard provides machine-enforceable validation across dataset structure, annotation provenance, quality documentation, and ML readiness within a single framework. DICOM standardizes image acquisition, storage, and communication at the individual study level. BIDS organizes neuroimaging research datasets with consistent naming conventions. Neither addresses the curation layer, viz., who annotated what, when, with what tool, and to what quality standard. This paper presents VIDS (Verified Imaging Dataset Standard), an open specification that defines folder layout, file naming, annotation provenance schemas, quality documentation, and 21 machine-enforceable validation rules across two compliance profiles. VIDS uses NIfTI as a canonical working format while preserving full DICOM metadata in sidecars for traceability, and supports export to any downstream ML framework (nnU-Net, MONAI, COCO, flat NIfTI) without loss of provenance. Twenty-two compliance dimensions are defined and four major public datasets – LIDC-IDRI, BraTS, CheXpert, and the Medical Segmentation Decathlon – are benchmarked against these dimensions. Even widely used datasets satisfy only 20–39% of these dimensions, with provenance and quality documentation as the largest systematic gaps. LIDC-Hybrid-100 is released as a 100-subject VIDS-compliant reference CT dataset with consensus segmentation masks from four radiologist annotations (mean pairwise Dice 0.7765), validating 21/21 on the Full compliance profile. VIDS is fully open source: the specification is CC BY 4.0, all tools are Apache 2.0, the reference validator is available on PyPI (pip install vids-validator), and LIDC-Hybrid-100 is published on Zenodo (https://doi.org/10.5281/zenodo.19582717).

This paper addresses the problem of adaptive reconfigurable intelligent surfaces (RIS) configuration design for user localization in rich-scattering environment (RSE), where electromagnetic waves undergo multiple interactions with dynamic scatterers and RIS elements. We propose an adaptive learning-based localization approach for a distributed RIS-assisted network in a RSE using a bidirectional long-short term memory (biLSTM) model that captures temporal correlations between observations. The proposed approach actively senses the environment using sequential pilot transmissions from the base station (BS), accounting for scattering effects, and adaptively updates the RIS configuration based on prior measurements to eventually accurately estimate and minimize the user localization error. The proposed model comprises two neural sub-networks: Scattering Estimation Network (Bi-SEN), for estimation of scattering in the environment, and Adaptive RIS-Assisted User Localization Network (Bi-ARULN), for RIS configuration and localization. Bayesian optimization is used for hyperparameter tuning of the model. The simulation results demonstrate the effectiveness of the proposed approach, achieving significantly lower localization root mean squared error (RMSE compared to random configuration, prestored codebook look-ups, and adaptive baselines in both single-input-single-output (SISO) and multiple-input-multiple output(MIMO) RIS-assisted networks in RSE. The design is generalized across configurations and scales with RIS size and network dimensions. The results highlight the strong potential of RIS deployment and of the proposed approach to enable reliable location services in RSE.

Multi-user virtual reality enables immersive interaction. However, rendering avatars for numerous participants on each headset incurs prohibitive computational overhead, limiting scalability. We introduce a framework, Privatar, to offload avatar reconstruction from headset to untrusted devices within the same local network while safeguarding attacks against adversaries capable of intercepting offloaded data. Privatar's key insight is that domain-specific knowledge of avatar reconstruction enables provably private offloading at minimal cost. (1) System level. We observe avatar reconstruction is frequency-domain decomposable via BDCT with negligible quality drop, and propose Horizontal Partitioning (HP) to keep high-energy frequency components on-device and offloads only low-energy components. HP offloads local computation while reducing information leakage to low-energy subsets only. (2) Privacy level. For individually offloaded, multi-dimensional signals without aggregation, worst-case local Differential Privacy requires prohibitive noise, ruining utility. We observe users' expression statistical distribution are slowly changing over time and trackable online, and hence propose Distribution-Aware Minimal Perturbation. DAMP minimizes noise based on each user's expression distribution to significantly reduce its effects on utility, retaining formal privacy guarantee. Combined, HP provides empirical privacy against expression identification attacks. DAMP further augments it to offer a formal guarantee against arbitrary adversaries. On a Meta Quest Pro, Privatar supports 2.37x more concurrent users at 6.5% higher reconstruction loss and 9% energy overhead, providing a better throughout-loss Pareto frontier over quantization, sparsity and local construction baselines. Privatar provides both provable privacy guarantee and stays robust against both empirical and NN-based attacks.

This paper investigates an uplink user equipment (UE) location and orientation estimation problem in an indoor rich-scattering environment (RSE) for a multiple-input-multiple-output (MIMO) narrowband reconfigurable intelligent surfaces (RIS)-assisted communication system. The localization problem in RSE is challenging as the uplink pilot signal undergoes multiple interactions with the RIS and dynamic scattering objects (SOs). This paper proposes an approach where base station (BS) adaptively senses the environment with the help of RIS. Based on this sensing, it sequentially designs RIS configuration, BS beamforming and UE beamforming vectors, using the sequence of pilot transmissions from the UE to the BS, with an objective of progressively focusing them onto the UE. Towards this end, we train a bidirectional long-short term memory (biLSTM) network based controller to capture the temporal dependencies between measurements to first adaptively sense the RSE and then design RIS, BS and UE beamforming vectors to localize the UE. We evaluate the proposed approach under various RSE conditions such as various distributed RIS installations, varying number of randomly moving SOs and sensing RIS elements. Simulation results illustrate that it effectively enables adaptive sensing to achieve low localization error with robustness in various RSEs.

Q-value iteration (Q-VI) is usually analyzed through the \(\gamma\)-contraction of the Bellman operator. This argument proves convergence to \(Q^*\), but it gives only a coarse account of when the induced greedy policy becomes optimal. We study discounted Q-VI as a switching system and focus on the practically optimal solution set (POSS), the set of \(Q\)-functions whose tie-broken greedy policies are optimal. The main result shows that Q-VI reaches the optimal action class in finite time by entering an invariant tube around \(\mathcal X_1=Q^*+\operatorname{span}(\mathbf 1)\), which is contained in the POSS. For every \(\varepsilon>0\), the distance to \(\mathcal X_1\) satisfies an exponential bound with rate \((\bar\rho+\varepsilon)^k\), where \(\bar\rho\) is the joint spectral radius of the projected switching family restricted to directions transverse to \(\mathcal X_1\). When \(\bar\rho<\gamma\), this transverse convergence is faster than the classical contraction rate. The analysis separates fast policy identification from the subsequent convergence to \(Q^*\), which may still be governed by the all-ones mode. We also give spectral and graph-theoretic conditions under which the strict inequality \(\bar\rho<\gamma\) holds or fails.

Being one of the oldest and most basic problems in image processing, image denoising has seen a resurgence spurred by rapid advances in deep learning. Yet, most modern denoising architectures make limited use of the technical knowledge acquired researching the classical denoisers that came before the mainstream use of neural networks, instead relying on depth and large parameter counts. This poses a challenge not only for understanding the properties of such networks, but also for deploying them on real devices which may present resource constraints and diverse noise profiles. Tackling both issues, we propose an architecture dedicated to RAW-to-RAW denoising that incorporates the interpretable structure of classical self-similarity-based denoisers into a fully learnable neural network. Our design centers on a novel nonlocal block that parallels the established pipeline of neighbor matching, collaborative filtering and aggregation popularized by nonlocal patch-based methods, operating on learned multiscale feature representations. This built-in nonlocality efficiently expands the receptive field, sufficing a single block per scale with a moderate number of neighbors to obtain high-quality results. Training the network on a curated dataset with clean real RAW data and modeled synthetic noise while conditioning it on a noise level map yields a sensor-agnostic denoiser that generalizes effectively to unseen devices. Both quantitative and visual results on benchmarks and in-the-wild photographs position our method as a practical and interpretable solution for real-world RAW denoising, achieving results competitive with state-of-the-art convolutional and transformer-based denoisers while using significantly fewer parameters. The code is available at https://github.com/MIA-UIB/nonlocal-matchfilter .

This paper deals with system representations in finite-sample signal subspaces and their application to data-driven fault detection. The first part addresses concepts of finite-sample image and kernel system representations and, associated with them, image and residual subspaces of finite-sample signals. On this basis, the equivalence between the fundamental lemma and finite-sample image subspace is demonstrated. While the image representation models the nominal system dynamics, the residual representation describes uncertainties in the input-output data and is essential for fault detection. This result extends the fundamental lemma and builds the basis for exploring data-driven fault detection. In the second part, a data-driven projection-based fault detection approach is developed. By means of a singular value decomposition, orthogonal projections onto the image and residual subspaces are realized in the context of a low-rank matrix approximation, leading to projection-based residual generation and evaluation. Finally, analysis of detection performance in the framework of matrix perturbation theory and comparison with existing data-driven fault detection methods are explored.

This study presents an Adaptive Transfer Learning and Thresholding-based Deep Learning Model (ATL-TDLM) for automated breathing pattern recognition using thermal imaging. Unlike conventional methods that rely on sound-based respiratory data, our approach leverages hierarchical deep feature extraction and adaptive multi-thresholding (AMT) to enhance feature segmentation. The model integrates knowledge distillation-based fine-tuning (KD-FT) to optimize learning transfer and contrastive representation learning (CRL) to improve inter-class separability between inhalation (INH) and exhalation (EXH) phases. The ATL-TDLM framework achieves an accuracy of 98.8%, significantly outperforming state-of-the-art models while ensuring computational efficiency. This approach has potential applications in respiratory disorder detection, including sleep apnea and asthma monitoring.

This paper investigates the agent design for solving the wireless resource allocation problem without sufficient channel state information (CSI), which cannot be effectively solved via conventional method. In the considered wireless agent design, we provide the general sense-repair-decide-act workflow, which can be used to intelligently solve general wireless resource allocation problem. A multi-objective optimization problem is formulated to adaptively satisfy different user requirements including both spectrum and energy efficiency. This work addresses the challenge of incomplete CSI for multiple optimization objectives. To solve this problem, we use an artificial intelligence (AI) model to predict missing channel data and construct an agent on the Coze platform, allowing the network operators to optimize multiple objectives through natural language conversations. To tackle the resource scheduling under different objectives, we develop adaptive algorithms. Simulation results validate the effectiveness of our proposed design, demonstrating that the proposed AI method reduces the root mean square error by approximately up to 67% compared to the traditional approach. Moreover, the data-driven scheduling balances system performance compared to conventional baseline approaches.