This paper presents a novel motion feature guided diffusion model for unpaired video-to-video translation (MFD-V2V), designed to synthesize dynamic, high-contrast cine cardiac magnetic resonance (CMR) from lower-contrast, artifact-prone displacement encoding with stimulated echoes (DENSE) CMR sequences. To achieve this, we first introduce a Latent Temporal Multi-Attention (LTMA) registration network that effectively learns more accurate and consistent cardiac motions from cine CMR image videos. A multi-level motion feature guided diffusion model, equipped with a specialized Spatio-Temporal Motion Encoder (STME) to extract fine-grained motion conditioning, is then developed to improve synthesis quality and fidelity. We evaluate our method, MFD-V2V, on a comprehensive cardiac dataset, demonstrating superior performance over the state-of-the-art in both quantitative metrics and qualitative assessments. Furthermore, we show the benefits of our synthesized cine CMRs improving downstream clinical and analytical tasks, underscoring the broader impact of our approach. Our code is publicly available at this https URL.

Accurate and scalable surrogate models for AC power flow are essential for real-time grid monitoring, contingency analysis, and decision support in increasingly dynamic and inverter-dominated power systems. However, most existing surrogates fall short of practical deployment due to their limited capacity to capture long-range nonlinear dependencies in meshed transmission networks and their weak enforcement of physical laws. These models often require extensive hyperparameter tuning, exhibit poor generalization under topology changes or large load swings, and typically do not quantify uncertainty or scale well beyond a few hundred buses. To address these challenges, this paper proposes a \textit{gated graph neural network (GGNN)} surrogate for AC power-flow estimation under topological uncertainty. The model is trained across multiple IEEE benchmark networks of varying size and complexity, each incorporating randomized line contingencies and up to 40\% load variation. To improve robustness and generalization, we explore both conventional supervised learning and physics-informed self-supervised training strategies. Comparative evaluations show that the proposed GGNN consistently outperforms prior GNN-based surrogates, achieving predictions closely aligned with Newton--Raphson solutions. By embedding operational constraints directly into the architecture and loss function, the model ensures physical consistency and delivers a lightweight, accurate, and scalable tool for real-time grid operations.

In this paper, we present a robust and adaptive model predictive control (MPC) framework for uncertain nonlinear systems affected by bounded disturbances and unmodeled nonlinearities. We use Gaussian Processes (GPs) to learn the uncertain dynamics based on noisy measurements, including those collected during system operation. As a key contribution, we derive robust predictions for GP models using contraction metrics, which are incorporated in the MPC formulation. The proposed design guarantees recursive feasibility, robust constraint satisfaction and convergence to a reference state, with high probability. We provide a numerical example of a planar quadrotor subject to difficult-to-model ground effects, which highlights significant improvements achieved through the proposed robust prediction method and through online learning.

Synthesizing second-language (L2) speech is potentially highly valued for L2 language learning experience and feedback. However, due to the lack of L2 speech synthesis datasets, it is difficult to synthesize L2 speech for low-resourced languages. In this paper, we provide a practical solution for editing native speech to approximate L2 speech and present PPG2Speech, a diffusion-based multispeaker Phonetic-Posteriorgrams-to-Speech model that is capable of editing a single phoneme without text alignment. We use Matcha-TTS's flow-matching decoder as the backbone, transforming Phonetic Posteriorgrams (PPGs) to mel-spectrograms conditioned on external speaker embeddings and pitch. PPG2Speech strengthens the Matcha-TTS's flow-matching decoder with Classifier-free Guidance (CFG) and Sway Sampling. We also propose a new task-specific objective evaluation metric, the Phonetic Aligned Consistency (PAC), between the edited PPGs and the PPGs extracted from the synthetic speech for editing effects. We validate the effectiveness of our method on Finnish, a low-resourced, nearly phonetic language, using approximately 60 hours of data. We conduct objective and subjective evaluations of our approach to compare its naturalness, speaker similarity, and editing effectiveness with TTS-based editing. Our source code is published at this https URL.

This paper presents a novel partitioning method designed to minimize control performance degradation resulting from partitioning a system for distributed control while maintaining the computational benefits of these methods. A game-theoretic performance metric, the modified Price of Anarchy, is introduced and is used in a generalizable partitioning metric to quantify optimality losses in a distributed controller. By finding the partition that minimizes the partitioning metric, the best-performing distributed control design is chosen. The presented partitioning metric is control-design agnostic, making it broadly applicable to many control design problems. In this paper, the developed metric is used to minimize the performance losses in the distributed control of a demand-flexible District Heating Network. The final distributed controller is provably feasible and stable. In simulation, this novel partitioning performed similarly to the centralized controller, increasing overall heat losses by only 1.9%, as compared to a similarly-sized baseline partition, which resulted in a 22% increase in losses.

We propose a novel iterative phase estimation framework, termed multi-source Griffin-Lim algorithm (MSGLA), for speech enhancement (SE) under additive noise conditions. The core idea is to leverage the ad-hoc consistency constraint of complex-valued short-time Fourier transform (STFT) spectrograms to address the sign ambiguity challenge commonly encountered in geometry-based phase estimation. Furthermore, we introduce a variant of the geometric constraint framework based on the law of sines and cosines, formulating a new phase reconstruction algorithm using noise phase estimates. We first validate the proposed technique through a series of oracle experiments, demonstrating its effectiveness under ideal conditions. We then evaluate its performance on the VB-DMD and WSJ0-CHiME3 data sets, and show that the proposed MSGLA variants match well or slightly outperform existing algorithms, including direct phase estimation and DNN-based sign prediction, especially in terms of background noise suppression.

This work addresses the general problem of control synthesis for continuous-space, discrete-time stochastic systems with probabilistic guarantees via finite abstractions. While established methods exist, they often trade off accuracy for tractability. We propose a unified abstraction framework that improves both the tightness of probabilistic guarantees and computational efficiency. First, we introduce multi-interval MDPs (MI-MDPs), a generalization of interval-valued MDPs (IMDPs), which allows multiple, possibly overlapping clusters of successor states. This results in tighter abstractions but with increased computational complexity. To mitigate this, we further propose a generalized form of MDPs with set-valued transition probabilities (SMDPs), which model transitions as a fixed probability to a state cluster, followed by a non-deterministic choice within the cluster, as a sound abstraction. We show that control synthesis for MI-MDPs reduces to robust dynamic programming via linear optimization, while SMDPs admit even more efficient synthesis algorithms that avoid linear programming altogether. Theoretically, we prove that, given the partitioning of the state and disturbance spaces, both MI-MDPs and SMDPs yield tighter probabilistic guarantees than IMDPs, and that SMDPs are tighter than MI-MDPs. Extensive experiments across several benchmarks validate our theoretical results and demonstrate that SMDPs achieve favorable trade-offs among tightness, memory usage, and computation time.

Reconfigurable antennas, including reconfigurable intelligent surface (RIS), movable antenna (MA), fluid antenna (FA), and other advanced antenna techniques, have been studied extensively in the context of reshaping wireless propagation environments for 6G and beyond wireless communications. Nevertheless, how to reconfigure/optimize the real-time controllable coefficients to achieve a favorable end-to-end wireless channel remains a substantial challenge, as it usually requires accurate modeling of the complex interaction between the reconfigurable devices and the electromagnetic waves, as well as knowledge of implicit channel propagation parameters. In this paper, we introduce a derivative-free optimization (a.k.a., zeroth-order (ZO) optimization) technique to directly optimize reconfigurable coefficients to shape the wireless end-to-end channel, without the need of channel modeling and estimation of the implicit environmental propagation parameters. We present the fundamental principles of ZO optimization and discuss its potential advantages in wireless channel reconfiguration. Two case studies for RIS and movable antenna-enabled single-input single-output (SISO) systems are provided to show the superiority of ZO-based methods as compared to state-of-the-art techniques. Finally, we outline promising future research directions and offer concluding insights on derivative-free optimization for reconfigurable antenna technologies.

The task of separating a superposition of signals into its individual components is a common challenge encountered in various signal processing applications, especially in domains such as audio and radar signals. A previous paper by Chui and Mhaskar proposes a method called Signal Separation Operator (SSO) to find the instantaneous frequencies and amplitudes of such superpositions where both of these change continuously and slowly over time. In this paper, we amplify and modify this method in order to separate chirp signals in the presence of crossovers, a very low SNR, and discontinuities. We give a theoretical analysis of the behavior of SSO in the presence of noise to examine the relationship between the minimal separation, minimal amplitude, SNR, and sampling frequency. Our method is illustrated with a few examples, and numerical results are reported on a simulated dataset comprising 7 simulated signals.

Myocardial infarction (MI) is a leading cause of death worldwide. Late gadolinium enhancement (LGE) and T2-weighted cardiac magnetic resonance (CMR) imaging can respectively identify scarring and edema areas, both of which are essential for MI risk stratification and prognosis assessment. Although combining complementary information from multi-sequence CMR is useful, acquiring these sequences can be time-consuming and prohibitive, e.g., due to the administration of contrast agents. Cine CMR is a rapid and contrast-free imaging technique that can visualize both motion and structural abnormalities of the myocardium induced by acute MI. Therefore, we present a new end-to-end deep neural network, referred to as CineMyoPS, to segment myocardial pathologies, \ie scars and edema, solely from cine CMR images. Specifically, CineMyoPS extracts both motion and anatomy features associated with MI. Given the interdependence between these features, we design a consistency loss (resembling the co-training strategy) to facilitate their joint learning. Furthermore, we propose a time-series aggregation strategy to integrate MI-related features across the cardiac cycle, thereby enhancing segmentation accuracy for myocardial pathologies. Experimental results on a multi-center dataset demonstrate that CineMyoPS achieves promising performance in myocardial pathology segmentation, motion estimation, and anatomy segmentation.

The emerging Web3 has great potential to provide worldwide decentralized services powered by global-range data-driven networks in the future. To ensure the security of Web3 services among diverse user entities, a decentralized identity (DID) system is essential. Especially, a user's access request to Web3 services can be treated as a DID transaction within the blockchain, executed through a consensus mechanism. However, a critical implementation issue arises in the current Web3, i.e., how to deploy network nodes to serve users on a global scale. To address this issue, emerging Low Earth Orbit (LEO) satellite communication systems, such as Starlink, offer a promising solution. With their global coverage and high reliability, these communication satellites can complement terrestrial networks as Web3 deployment infrastructures. In this case, this paper develops three hybrid satellite-ground modes to deploy the blockchain-enabled DID system for Web3 users. Three modes integrate ground nodes and satellites to provide flexible and continuous DID services for worldwide users. Meanwhile, to evaluate the effectiveness of the present hybrid deployment modes, we analyze the complete DID consensus performance of blockchain on three hybrid satellite-ground modes. Moreover, we conduct numerical and simulation experiments to verify the effectiveness of three hybrid satellite-ground modes. The impacts of various system parameters are thoroughly analyzed, providing valuable insights for implementing the worldwide Web3 DID system in real-world network environments.

Grid-connected inverter control is challenging to implement due to the difficulty of obtaining and maintaining an accurate grid model. Direct Data-Driven Predictive Control provides a model-free alternative to traditional model-based control methods. This paper describes how the recently-proposed Transient Predictive Control (TPC) can be used for real-world, plug-and-play inverter control. The following hypotheses were tested: 1) The TPC algorithm can be run online using standard hardware, and 2) TPC, which is derived using Linear Time-Invariant assumptions, is effective for grid-connected inverter control, which is a nonlinear and time-varying system. Experiments conducted on a two-converter benchtop setup and at the CoSES Laboratory on a 25 kVA converter connected to the Munich grid support these hypotheses.

This study examines an integrated sensing and communication (ISAC) transceiver featuring a simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) and a receiver equipped with a passive electronically scanned array (PESA) and a single digital channel. By utilizing a periodic pulsed signal emitted by a feeder, we introduce at the STAR-RIS a space modulation to illuminate two angular directions observed by the radar receiver, one in each half-space, and a time modulation to distinguish the corresponding echoes from prospective moving targets and embed communication messages. The proposed time modulation employs orthogonal binary codebooks with different trade-offs in transmission and error rates, while having minimal impact on the radar performance, evaluated by probability of detection and root mean square error in the radial velocity estimation.

Pinching-antenna systems (PASS) have been recently proposed to improve the performance of wireless networks by reconfiguring both the large-scale and small-scale channel conditions. However, existing studies ignore the physical constraints of antenna placement and assume fixed antenna radiation power. To fill this research gap, this paper investigates the design of PASS taking into account the motion power consumption of pinching-antennas (PAs) and the impact of adjustable antenna radiation power. To that end, we minimize the average power consumption for a given quality-of-service (QoS) requirement, by jointly optimizing the antenna positions, antenna radiation power ratios, and transmit beamforming. To the best of the authors' knowledge, this is the first work to consider radiation power optimization in PASS, which provides an additional degree of freedom (DoF) for system design. The cases with both continuous and discrete antenna placement are considered, where the main challenge lies in the fact that the antenna positions affect both the magnitude and phase of the channel coefficients of PASS, making system optimization very challenging. To tackle the resulting unique obstacles, an alternating direction method of multipliers (ADMM)-based framework is proposed to solve the problem for continuous antenna movement, while its discrete counterpart is formulated as a mixed integer nonlinear programming (MINLP) problem and solved by the block coordinate descent (BCD) method. Simulation results validate the performance enhancement achieved by incorporating PA movement power assumption and adjustable radiation power into PASS design, while also demonstrating the efficiency of the proposed optimization framework. The benefits of PASS over conventional multiple-input multiple-output (MIMO) systems in mitigating the large-scale path loss and inter-user interference is also revealed.

The optimal management of a building's microclimate to satisfy the occupants' needs and objectives in terms of comfort, energy efficiency, and costs is particularly challenging. This complexity arises from the non-linear, time-dependent interactions among all the variables of the control problem and the changing internal and external constraints. Focusing on the accurate modeling of the indoor temperature, we propose a data-driven approach to address this challenge. We account for thermal inertia, non-linear effects, small perturbations of the indoor climate dynamics caused by ventilation and weather variations, as well as for the stochastic nature of the control system due to the observed noise in the input signal. Since the prohibitive cost of quality data acquisition and processing limits the implementation of data-driven approaches for real-life problems, we applied a method that merges several Bayesian machine learning and deep learning architectures that are suitable for predicting complex system dynamics, while relaxing the dataset quality requirements. Our framework includes a built-in deep Kalman filter, which makes it deployable even with low-accuracy temperature sensors. It achieves state-of-the-art performance, best performing with a 150-minute prediction horizon with an RMSE of 0.2455, an MAE of 0.162, and an $R^2$ of 0.926. The model's performance remains consistent even when exposed to highly noisy data. Finally, we show how our approach can be extended to other applications including demand response event duration prediction and equipment failure detection.

This study considers a base station equipped with sensing and communication capabilities, which serves a ground user and scans a portion of the sky via a passive reconfigurable intelligent surface. To achieve more favorable system tradeoffs, we utilize a multi-frame radar detector, comprising a detector, a plot-extractor, and a track-before-detect processor. The main idea proposed here is that user spectral efficiency can be enhanced by increasing the number of scans jointly processed by the multi-frame radar detector while maintaining the same sensing performance. A numerical analysis is conducted to verify the effectiveness of the proposed solution and to evaluate the achievable system tradeoffs.

Dynamic positron emission tomography (PET) and kinetic modeling are pivotal in advancing tracer development research in small animal studies. Accurate kinetic modeling requires precise input function estimation, traditionally achieved via arterial blood sampling. However, arterial cannulation in small animals like mice, involves intricate, time-consuming, and terminal procedures, precluding longitudinal studies. This work proposes a non-invasive, fully convolutional deep learning-based approach (FC-DLIF) to predict input functions directly from PET imaging, potentially eliminating the need for blood sampling in dynamic small-animal PET. The proposed FC-DLIF model includes a spatial feature extractor acting on the volumetric time frames of the PET sequence, extracting spatial features. These are subsequently further processed in a temporal feature extractor that predicts the arterial input function. The proposed approach is trained and evaluated using images and arterial blood curves from [$^{18}$F]FDG data using cross validation. Further, the model applicability is evaluated on imaging data and arterial blood curves collected using two additional radiotracers ([$^{18}$F]FDOPA, and [$^{68}$Ga]PSMA). The model was further evaluated on data truncated and shifted in time, to simulate shorter, and shifted, PET scans. The proposed FC-DLIF model reliably predicts the arterial input function with respect to mean squared error and correlation. Furthermore, the FC-DLIF model is able to predict the arterial input function even from truncated and shifted samples. The model fails to predict the AIF from samples collected using different radiotracers, as these are not represented in the training data. Our deep learning-based input function offers a non-invasive and reliable alternative to arterial blood sampling, proving robust and flexible to temporal shifts and different scan durations.

Low-altitude wireless networks (LAWNs) have been envisioned as flexible and transformative platforms for enabling delay-sensitive control applications in Internet of Things (IoT) systems. In this work, we investigate the real-time wireless control over a LAWN system, where an aerial drone is employed to serve multiple mobile automated guided vehicles (AGVs) via finite blocklength (FBL) transmission. Toward this end, we adopt the model predictive control (MPC) to ensure accurate trajectory tracking, while we analyze the communication reliability using the outage probability. Subsequently, we formulate an optimization problem to jointly determine control policy, transmit power allocation, and drone trajectory by accounting for the maximum travel distance and control input constraints. To address the resultant non-convex optimization problem, we first derive the closed-form expression of the outage probability under FBL transmission. Based on this, we reformulate the original problem as a quadratic programming (QP) problem, followed by developing an alternating optimization (AO) framework. Specifically, we employ the projected gradient descent (PGD) method and the successive convex approximation (SCA) technique to achieve computationally efficient sub-optimal solutions. Furthermore, we thoroughly analyze the convergence and computational complexity of the proposed algorithm. Extensive simulations and AirSim-based experiments are conducted to validate the superiority of our proposed approach compared to the baseline schemes in terms of control performance.

This study investigates a communication-centric integrated sensing and communication (ISAC) system that utilizes orthogonal time frequency space (OTFS) modulated signals emitted by low Earth orbit (LEO) satellites to estimate the parameters of space targets experiencing range migration, henceforth referred to as high-speed targets. Leveraging the specific signal processing performed by OTFS transceivers, we derive a novel input-output model for the echo generated by a high-speed target in scenarios where ideal and rectangular shaping filters are employed. Our findings reveal that the target response exhibits a sparse structure in the delay-Doppler domain, dependent solely upon the initial range and range-rate; notably, range migration causes a spread in the target response, marking a significant departure from previous studies. Utilizing this signal structure, we propose an approximate implementation of the maximum likelihood estimator for the target's initial range, range-rate, and amplitude. The estimation process involves obtaining coarse information on the target response using a block orthogonal matching pursuit algorithm, followed by a refinement step using a bank of matched filters focused on a smaller range and range-rate region. Finally, numerical examples are provided to evaluate the estimation performance.

Echocardiography (echo) plays an indispensable role in the clinical practice of heart diseases. However, ultrasound imaging typically provides only two-dimensional (2D) cross-sectional images from a few specific views, making it challenging to interpret and inaccurate for estimation of clinical parameters like the volume of left ventricle (LV). 3D ultrasound imaging provides an alternative for 3D quantification, but is still limited by the low spatial and temporal resolution and the highly demanding manual delineation. To address these challenges, we propose an innovative framework for reconstructing personalized 3D heart anatomy from 2D echo slices that are frequently used in clinical practice. Specifically, a novel 3D reconstruction pipeline is designed, which alternatively optimizes between the 3D pose estimation of these 2D slices and the 3D integration of these slices using an implicit neural network, progressively transforming a prior 3D heart shape into a personalized 3D heart model. We validate the method with two datasets. When six planes are used, the reconstructed 3D heart can lead to a significant improvement for LV volume estimation over the bi-plane method (error in percent: 1.98\% VS. 20.24\%). In addition, the whole reconstruction framework makes even an important breakthrough that can estimate RV volume from 2D echo slices (with an error of 5.75\% ). This study provides a new way for personalized 3D structure and function analysis from cardiac ultrasound and is of great potential in clinical practice.

Owing to the use of attention mechanism to leverage the dependency across tokens, Transformers are efficient for natural language processing. By harnessing permutation properties broadly exist in resource allocation policies, each mapping measurable environmental parameters (e.g., channel matrix) to optimized variables (e.g., precoding matrix), graph neural networks (GNNs) are promising for learning these policies efficiently in terms of scalability and generalizability. To reap the benefits of both architectures, there is a recent trend of incorporating attention mechanism with GNNs for learning wireless policies. Nevertheless, is the attention mechanism really needed for resource allocation? In this paper, we strive to answer this question by analyzing the structures of functions defined on sets and numerical algorithms, given that the permutation properties of wireless policies are induced by the involved sets (say user set). In particular, we prove that the permutation equivariant functions on a single set can be recursively expressed by two types of functions: one involves attention, and the other does not. We proceed to re-express the numerical algorithms for optimizing several representative resource allocation problems in recursive forms. We find that when interference (say multi-user or inter-data stream interference) is not reflected in the measurable parameters of a policy, attention needs to be used to model the interference. With the insight, we establish a framework of designing GNNs by aligning with the structures. By taking reconfigurable intelligent surface-aided hybrid precoding as an example, the learning efficiency of the proposed GNN is validated via simulations.

This paper quantifies and addresses the bias of subspace-based Data-Driven Predictive Control (DDPC) for linear, time-invariant (LTI) systems. The primary focus is the bias that arises when the training data is gathered with a feedback controller in closed-loop with the system. First, the closed-loop bias of Subspace Predictive Control is quantified using the training data innovations. Next, the bias of direct, subspace-based DDPC methods DeePC and $\gamma$-DDPC is shown to consist of two parts--the Subspace Bias, which arises from closed-loop data, and an Optimism Bias, which arises from DeePC/$\gamma$-DDPC's "optimistic" adjustment of the output trajectory. We show that, unlike subspace-based DDPC methods, Transient Predictive Control does not suffer from Subspace Bias or Optimism Bias. Double integrator experiments demonstrate that Subspace and Optimism Bias are responsible for poor reference tracking by the subspace-based DDPC methods.

We present an open-source system designed for multilingual translation and speech regeneration, addressing challenges in communication and accessibility across diverse linguistic contexts. The system integrates Whisper for speech recognition with Voice Activity Detection (VAD) to identify speaking intervals, followed by a pipeline of Large Language Models (LLMs). For multilingual applications, the first LLM segments speech into coherent, complete sentences, which a second LLM then translates. For speech regeneration, the system uses a text-to-speech (TTS) module with voice cloning capabilities to replicate the original speaker's voice, maintaining naturalness and speaker identity. The system's open-source components can operate locally or via APIs, offering cost-effective deployment across various use cases. These include real-time multilingual translation in Zoom sessions, speech regeneration for public broadcasts, and Bluetooth-enabled multilingual playback through personal devices. By preserving the speaker's voice, the system ensures a seamless and immersive experience, whether translating or regenerating speech. This open-source project is shared with the community to foster innovation and accessibility. We provide a detailed system performance analysis, including latency and word accuracy, demonstrating its potential to enable inclusive, adaptable communication solutions in real-world multilingual scenarios.

The Feedback-based Algorithm for Quantum Optimization (FALQON) is a Lyapunov inspired quantum algorithm proposed to tackle combinatorial optimization problems. In this paper, we examine the robustness of FALQON against coherent control errors, a class of multiplicative errors that affect the control input. We show that the algorithm is asymptotically robust with respect to systematic errors, and we derive robustness bounds for independent errors. Finally, we propose a robust version of FALQON which minimizes a regularized Lyapunov function. Our theoretical results are supported through simulations.

Achieving rapid and time-deterministic stabilization for complex systems characterized by strong nonlinearities and parametric uncertainties presents a significant challenge. Traditional model-based control relies on precise system models, whereas purely data-driven methods often lack formal stability guarantees, limiting their applicability in safety-critical systems. This paper proposes a novel control framework that synergistically integrates data-driven modeling with model-based control. The framework first employs the Extended Dynamic Mode Decomposition with Control (EDMDc) to identify a high-dimensional Koopman linear model and quantify its bounded uncertainty from data. Subsequently, a novel Prescribed-Time Adaptive Backstepping (PTAB) controller is synthesized based on this data-driven model. The design leverages the structural advantages of Koopman linearization to systematically handle model errors and circumvent the "explosion of complexity" issue inherent in traditional backstepping. The proposed controller is validated through simulations on the classic Van der Pol oscillator. The results demonstrate that the controller can precisely stabilize the system states to a small neighborhood of the origin within a user-prescribed time, regardless of the initial conditions, while ensuring the boundedness of all closed-loop signals. This research successfully combines the flexibility of data-driven approaches with the rigor of Lyapunov-based analysis. It provides a high-performance control strategy with quantifiable performance and pre-assignable settling time for nonlinear systems, showcasing its great potential for controlling complex dynamics.

Tables of optimal transition coefficients and peak sidelobe level (PSL, in dB) associated with frequency sampling filter (FSF) design were published by Rabiner et al. (Jun 1970), and reproduced, for example, in the book Digital Signal Processing by Proakis and Manolakis (4/e, 2007). A set of values are also given in Appendix H of Understanding Digital Signal Processing} by Lyons (3/e, 2011), but there are significant differences between these two sets. For example, for $N=16$ and BW$=4$, two different transition coefficient values have been reported, viz., $0.38916626$ (Rabiner, et al.) and $0.34918551$ (Lyons). Neither is optimal, for we find the optimum value to be $0.40474097$. The published values of the corresponding PSLs were also found to be incorrect. In this paper we give the optimal values of the transition coefficients and PSL values as estimated by our program for the lowpass and bandpass filters listed in Rabiner et al. and Lyons.

Current deep neural network (DNN) based speech separation faces a fundamental challenge -- while the models need to be trained on short segments due to computational constraints, real-world applications typically require processing significantly longer recordings with multiple utterances per speaker than seen during training. In this paper, we investigate how existing approaches perform in this challenging scenario and propose a frequency-temporal recurrent neural network (FTRNN) that effectively bridges this gap. Our FTRNN employs a full-band module to model frequency dependencies within each time frame and a sub-band module that models temporal patterns in each frequency band. Despite being trained on short fixed-length segments of 10 s, our model demonstrates robust separation when processing signals significantly longer than training segments (21-121 s) and preserves speaker association across utterance gaps exceeding those seen during training. Unlike the conventional segment-separation-stitch paradigm, our lightweight approach (0.9 M parameters) performs inference on long audio without segmentation, eliminating segment boundary distortions while simplifying deployment. Experimental results demonstrate the generalization ability of FTRNN for multi-utterance speech separation and speaker association.

Switching communication topologies can cause instability in vehicle platoons, as vehicle information may be lost during the dynamic switching process. This highlights the need to design a controller capable of maintaining the stability of vehicle platoons under dynamically changing topologies. However, capturing the dynamic characteristics of switching topologies and obtaining complete vehicle information for controller design while ensuring stability remains a significant challenge. In this study, we propose an observer-based distributed model predictive control (DMPC) method for vehicle platoons under directed Markovian switching topologies. Considering the stochastic nature of the switching topologies, we model the directed switching communication topologies using a continuous-time Markov chain. To obtain the leader vehicle's information for controller design, we develop a fully distributed adaptive observer that can quickly adapt to the randomly switching topologies, ensuring that the observed information is not affected by the dynamic topology switches. Additionally, a sufficient condition is derived to guarantee the mean-square stability of the observer. Furthermore, we construct the DMPC terminal update law based on the observer and formulate a string stability constraint based on the observed information. Numerical simulations demonstrate that our method can reduce tracking errors while ensuring string stability.

In this paper, a novel pinching-antenna assisted index modulation (PA-IM) scheme is proposed for improving the spectral efficiency without increasing the hardware complexity, where the information bits are conveyed not only by the conventional M-ary quadrature amplitude modulation (QAM) symbols but also by the indices of pinching antenna (PA) position patterns. To realize the full potential of this scheme, this paper focuses on the comprehensive transceiver design, addressing key challenges in signal detection at the receiver and performance optimization at thetransmitter. First, a comprehensive channel model is formulated for this architecture, which sophisticatedly integrates the deterministic in-waveguide propagation effects with the stochastic nature of wireless channels, including both largescale path loss and small-scale fading. Next, to overcome the prohibitive complexity of optimal maximum likelihood (ML) detection, a low-complexity box-optimized sphere decoding (BOSD) algorithm is designed, which adaptively prunes the search space whilst preserving optimal ML performance. Furthermore, an analytical upper bound on the bit error rate (BER) is derived and validated by the simulations. Moreover, a new transmit precoding method is designed using manifold optimization, which minimizes the BER by jointly optimizing the complex-valued precoding coefficients across the waveguides for the sake of maximizing the minimum Euclidean distance of all received signal points. Finally, the simulation results demonstrate that the proposed PA-IM scheme attains a significant performance gain over its conventional counterparts and that the overall BER of the pinching-antenna system is substantially improved by the proposed precoding design.

Colorectal polyp segmentation is critical for early detection of colorectal cancer, yet weak and low contrast boundaries significantly limit automated accuracy. Existing deep models either blur fine edge details or rely on handcrafted filters that perform poorly under variable imaging conditions. We propose MEGANet-W, a Wavelet Driven Edge Guided Attention Network that injects directional, parameter free Haar wavelet edge maps into each decoder stage to recalibrate semantic features. Our two main contributions are: (1) a two-level Haar wavelet head for multi orientation edge extraction; and (2) Wavelet Edge Guided Attention (WEGA) modules that fuse wavelet cues with reverse and input branches. On five public polyp datasets, MEGANetW consistently outperforms existing methods, improving mIoU by up to 2.3% and mDice by 1.2%, while introducing no additional learnable parameters.

The Princess Marijke lock complex is a large lock and water-protection installation in the Netherlands between the river Rhine and the Amsterdam-Rijnkanaal -- a large waterway connecting the Rhine to the port of Amsterdam. The lock complex consists of two independent locks and a moveable flood-protection barrier. Ensuring safe control of the lock complex is of utmost importance to guarantee both flood-protection and reliable ship operations. This paper gives a precise, formal description of the software control of the lock complex in less than 400 lines of mCRL2 code. This description can act as a blueprint on how the software of this lock complex needs to be constructed. Moreover, using model checking, 53 software requirements are shown to be valid, ensuring that the formal description of the behaviour is correct with regard to these properties and is unlikely to contain mistakes and oversights.

Auditory scene analysis (ASA) aims to retrieve information from the acoustic environment, by carrying out three main tasks: sound source location, separation, and classification. These tasks are traditionally executed with a linear data flow, where the sound sources are first located; then, using their location, each source is separated into its own audio stream; from each of which, information is extracted that is relevant to the application scenario (audio event detection, speaker identification, emotion classification, etc.). However, running these tasks linearly increases the overall response time, while making the last tasks (separation and classification) highly sensitive to errors of the first task (location). A considerable amount of effort and computational complexity has been employed in the state-of-the-art to develop techniques that are the least error-prone possible. However, doing so gives rise to an ASA system that is non-viable in many applications that require a small computational footprint and a low response time, such as bioacoustics, hearing-aid design, search and rescue, human-robot interaction, etc. To this effect, in this work, a multi-agent approach is proposed to carry out ASA where the tasks are run in parallel, with feedback loops between them to compensate for local errors, such as: using the quality of the separation output to correct the location error; and using the classification result to reduce the localization's sensitivity towards interferences. The result is a multi-agent auditory scene analysis (MASA) system that is robust against local errors, without a considerable increase in complexity, and with a low response time. The complete proposed MASA system is provided as a framework that uses open-source tools for sound acquisition and reproduction (JACK) and inter-agent communication (ROS2), allowing users to add their own agents.

We introduce DeSTA2.5-Audio, a general-purpose Large Audio Language Model (LALM) designed for robust auditory perception and instruction-following, without requiring task-specific audio instruction-tuning. Recent LALMs typically augment Large Language Models (LLMs) with auditory capabilities by training on large-scale, manually curated or LLM-synthesized audio-instruction datasets. However, these approaches have often suffered from the catastrophic forgetting of the LLM's original language abilities. To address this, we revisit the data construction pipeline and propose DeSTA, a self-generated cross-modal alignment strategy in which the backbone LLM generates its own training targets. This approach preserves the LLM's native language proficiency while establishing effective audio-text alignment, thereby enabling zero-shot generalization without task-specific tuning. Using DeSTA, we construct DeSTA-AQA5M, a large-scale, task-agnostic dataset containing 5 million training samples derived from 7,000 hours of audio spanning 50 diverse datasets, including speech, environmental sounds, and music. DeSTA2.5-Audio achieves state-of-the-art or competitive performance across a wide range of audio-language benchmarks, including Dynamic-SUPERB, MMAU, SAKURA, Speech-IFEval, and VoiceBench. Comprehensive comparative studies demonstrate that our self-generated strategy outperforms widely adopted data construction and training strategies in both auditory perception and instruction-following capabilities. Our findings underscore the importance of carefully designed data construction in LALM development and offer practical insights for building robust, general-purpose LALMs.

Recent works on deep non-linear spatially selective filters demonstrate exceptional enhancement performance with computationally lightweight architectures for stationary speakers of known directions. However, to maintain this performance in dynamic scenarios, resource-intensive data-driven tracking algorithms become necessary to provide precise spatial guidance conditioned on the initial direction of a target speaker. As this additional computational overhead hinders application in resource-constrained scenarios such as real-time speech enhancement, we present a novel strategy utilizing a low-complexity tracking algorithm in the form of a particle filter instead. Assuming a causal, sequential processing style, we introduce temporal feedback to leverage the enhanced speech signal of the spatially selective filter to compensate for the limited modeling capabilities of the particle filter. Evaluation on a synthetic dataset illustrates how the autoregressive interplay between both algorithms drastically improves tracking accuracy and leads to strong enhancement performance. A listening test with real-world recordings complements these findings by indicating a clear trend towards our proposed self-steering pipeline as preferred choice over comparable methods.

Hybrid beamforming has been widely employed in mmWave communications such as vehicular-to-everything (V2X) scenarios, as a compromise between hardware complexity and spectral efficiency. However, the inherent coupling between analog and digital precoders in hybrid array architecture significantly limits the computational and spectral efficiency of existing algorithms. To address this issue, we propose an alternating residual error elimination (AREE) algorithm, which decomposes the hybrid beamforming problem into two low-dimensional subproblems, each exhibiting a favorable matrix structure that enables effective decoupling of analog and digital precoders from the matrix product formulation. These subproblems iteratively eliminate each other's residual errors, driving the original problem toward the optimal hybrid beamforming performance. The proposed initialization ensures rapid convergence, while a low-complexity geometric channel SVD algorithm is developed by transforming the high-dimensional sparse channel into a low-dimensional equivalent, thereby simplifying the derivation of subproblems. Simulation results demonstrate that the AREE algorithm effectively decouples analog and digital precoders with low complexity, achieves fast convergence, and offers higher spectral efficiency than existing beamforming methods.

Personalized head-related transfer functions (HRTFs) are essential for ensuring a realistic auditory experience over headphones, because they take into account individual anatomical differences that affect listening. Most machine learning approaches to HRTF personalization rely on a learned low-dimensional latent space to generate or select custom HRTFs for a listener. However, these latent representations are typically learned in a manner that optimizes for spectral reconstruction but not for perceptual compatibility, meaning they may not necessarily align with perceptual distance. In this work, we first study whether traditionally learned HRTF representations are well correlated with perceptual relations using auditory-based objective perceptual metrics; we then propose a method for explicitly embedding HRTFs into a perception-informed latent space, leveraging a metric-based loss function and supervision via Metric Multidimensional Scaling (MMDS). Finally, we demonstrate the applicability of these learned representations to the task of HRTF personalization. We suggest that our method has the potential to render personalized spatial audio, leading to an improved listening experience.

In this paper, the precoding design is investigated for maximizing the throughput of millimeter wave (mmWave) multiple-input multiple-output (MIMO) systems with obstructed direct communication paths. In particular, a reconfigurable intelligent surface (RIS) is employed to enhance MIMO transmissions, considering mmWave characteristics related to line-of-sight (LoS) and multipath effects. The traditional exhaustive search (ES) for optimal codewords in the continuous phase shift is computationally intensive and time-consuming. To reduce computational complexity, permuted discrete Fourier transform (DFT) vectors are used for finding codebook design, incorporating amplitude responses for practical or ideal RIS systems. However, even if the discrete phase shift is adopted in the ES, it results in significant computation and is time-consuming. Instead, the trained deep neural network (DNN) is developed to facilitate faster codeword selection. Simulation results show that the DNN maintains sub-optimal spectral efficiency even as the distance between the end-user and the RIS has variations in the testing phase. These results highlight the potential of DNN in advancing RIS-aided systems.

This work investigates the benefits of implementing a systematic approach to social isolation policies during epidemics. We develop a mixed integer data-driven model predictive control (MPC) scheme based on an SIHRD model which is identified from available data. The case of the spread of the SARS-CoV-2 virus (also known as COVID-19) in Mauritius is used as a reference point with data obtained during the period December 2021 to May 2022. The isolation scheme is designed with the control decision variable taking a finite set of values corresponding to the desired level of isolation. The control input is further restricted to shifting between levels only after a minimum amount of time. The simulation results validate our design, showing that the need for hospitalisation remains within the capacity of the health centres, with the number of deaths considerably reduced by raising the level of isolation for short periods of time with negligible social and economic impact. We also show that the introduction of additional isolation levels results in a smoother containment approach with a considerably reduced hospitalisation burden.

The Internet of Vehicles (IoV) transforms the transportation ecosystem promising pervasive connectivity and data-driven approaches. Deep learning and generative Artificial Intelligence (AI) have the potential to significantly enhance the operation of applications within IoV by facilitating efficient decision-making and predictive capabilities, including intelligent navigation, vehicle safety monitoring, accident prevention, and intelligent traffic management. Nevertheless, efficiently transmitting and processing the massive volumes of data generated by the IoV in real-time remains a significant challenge, particularly in dynamic and unpredictable wireless channel conditions. To address these challenges, this paper proposes a semantic communication framework based on channel perception to improve the accuracy and efficiency of data transmission. The semantic communication model extracts and compresses the information to be transmitted. In addition, the wireless channel is estimated by using a generative diffusion model, which is employed to predict the dynamic channel states, thereby improving the quality of IoV service. In dynamic scenarios, however, the channel estimation performance may be degraded when substantially new scenarios take place, which will adversely affect user experience. To mitigate this limitation, we employ a large model to fine-tune the channel generation model to enhance its adaptability for varying scenarios. The performance and reliability of the proposed framework are evaluated on the two public datasets.

The adaptive Iterative Soft-Thresholding Algorithm (ISTA) has been a popular algorithm for finding a desirable solution to the LASSO problem without explicitly tuning the regularization parameter $\lambda$. Despite that the adaptive ISTA is a successful practical algorithm, few theoretical results exist. In this paper, we present the theoretical analysis on the adaptive ISTA with the thresholding strategy of estimating noise level by median absolute deviation. We show properties of the fixed points of the algorithm, including scale equivariance, non-uniqueness, and local stability, prove the local linear convergence guarantee, and show its global convergence behavior.

We introduce PANAMA, an active learning framework for the training of end-to-end parametric guitar amp models using a WaveNet-like architecture. With \model, one can create a virtual amp by recording samples that are determined by an active learning strategy to use a minimum amount of datapoints (i.e., amp knob settings). We show that gradient-based optimization algorithms can be used to determine the optimal datapoints to sample, and that the approach helps under a constrained number of samples.

This article investigates the robustness of gradient descent algorithms under perturbations. The concept of small-disturbance input-to-state stability (ISS) for discrete-time nonlinear dynamical systems is introduced, along with its Lyapunov characterization. The conventional linear Polyak-Lojasiewicz (PL) condition is then extended to a nonlinear version, and it is shown that the gradient descent algorithm is small-disturbance ISS provided the objective function satisfies the generalized nonlinear PL condition. This small-disturbance ISS property guarantees that the gradient descent algorithm converges to a small neighborhood of the optimum under sufficiently small perturbations. As a direct application of the developed framework, we demonstrate that the LQR cost satisfies the generalized nonlinear PL condition, thereby establishing that the policy gradient algorithm for LQR is small-disturbance ISS. Additionally, other popular policy gradient algorithms, including natural policy gradient and Gauss-Newton method, are also proven to be small-disturbance ISS.

The goal of this paper is to estimate the directions of arrival (DoAs) for hybrid analog/digital (HAD) receivers when the number of snapshots is too small for statistical averaging to be reliable. This goal is achieved in fully-digital receivers by employing the matrix pencil method (MPM). Unfortunately, the MPM cannot be directly applied in HAD receivers because of the entanglement induced by the underlying analog combiners on the output signals. Furthermore, these analog combiners project the received signal onto a low-dimensional space, jeopardizing the reception of signals arriving from particular DoA ranges. To circumvent these difficulties, we propose two approaches to enable the MPM to extract the DoAs in HAD receivers. The two approaches avoid severe attenuation induced by low-dimensional projection by cycling over an exhaustive set of analog combiners, collectively spanning the entire space. The first approach can be applied to both fully-connected (FC) and partially-connected (PC) HADs and relies on the availability of periodic, potentially unknown, signals to disentangle the output of the HAD receiver. The second approach applies to PC-HADs only, and eliminates contingency on periodic signals by exploiting the underlying block diagonal structure. The superiority of the proposed approaches is demonstrated via numerical simulations and comparisons with the Cramér-Rao lower bound.

Physics-consistent theoretical studies on RIS-parametrized wireless channels use models from multiport-network theory (MNT) to capture mutual-coupling (MC) effects. However, in practice, RIS design and radio environment are partially or completely unknown. We fill a research gap on how to estimate the MNT model parameters in such experimentally relevant scenarios. Our technique efficiently combines closed-form and gradient-descent steps, and it can be applied to multi-bit-programmable RIS elements. We discuss inevitable (but operationally irrelevant) parameter ambiguities. We experimentally validate our technique in an unknown rich-scattering environment parametrized by eight 8-bit-programmable RIS elements of unknown design. We experimentally evaluate the performance of RIS configurations optimized with the estimated MNT model and an MC-unaware cascaded model. While the models differ in accuracy by up to 17 dB, the end-to-end performance differences are small.

Modeling voice identity is challenging due to its multifaceted nature. In generative speech systems, identity is often assessed using automatic speaker verification (ASV) embeddings, designed for discrimination rather than characterizing identity. This paper investigates which aspects of a voice are captured in such representations. We find that widely used ASV embeddings focus mainly on static features like timbre and pitch range, while neglecting dynamic elements such as rhythm. We also identify confounding factors that compromise speaker similarity measurements and suggest mitigation strategies. To address these gaps, we propose U3D, a metric that evaluates speakers' dynamic rhythm patterns. This work contributes to the ongoing challenge of assessing speaker identity consistency in the context of ever-better voice cloning systems. We publicly release our code.

Utilizing hyperspectral remote sensing technology enables the extraction of fine-grained land cover classes. Typically, satellite or airborne images used for training and testing are acquired from different regions or times, where the same class has significant spectral shifts in different scenes. In this paper, we propose a Bi-directional Domain Adaptation (BiDA) framework for cross-domain hyperspectral image (HSI) classification, which focuses on extracting both domain-invariant features and domain-specific information in the independent adaptive space, thereby enhancing the adaptability and separability to the target scene. In the proposed BiDA, a triple-branch transformer architecture (the source branch, target branch, and coupled branch) with semantic tokenizer is designed as the backbone. Specifically, the source branch and target branch independently learn the adaptive space of source and target domains, a Coupled Multi-head Cross-attention (CMCA) mechanism is developed in coupled branch for feature interaction and inter-domain correlation mining. Furthermore, a bi-directional distillation loss is designed to guide adaptive space learning using inter-domain correlation. Finally, we propose an Adaptive Reinforcement Strategy (ARS) to encourage the model to focus on specific generalized feature extraction within both source and target scenes in noise condition. Experimental results on cross-temporal/scene airborne and satellite datasets demonstrate that the proposed BiDA performs significantly better than some state-of-the-art domain adaptation approaches. In the cross-temporal tree species classification task, the proposed BiDA is more than 3\%$\sim$5\% higher than the most advanced method. The codes will be available from the website: this https URL.

General-purpose audio representations have proven effective across diverse music information retrieval applications, yet their utility in intelligent music production remains limited by insufficient understanding of audio effects (Fx). Although previous approaches have emphasized audio effects analysis at the mixture level, this focus falls short for tasks demanding instrument-wise audio effects understanding, such as automatic mixing. In this work, we present Fx-Encoder++, a novel model designed to extract instrument-wise audio effects representations from music mixtures. Our approach leverages a contrastive learning framework and introduces an "extractor" mechanism that, when provided with instrument queries (audio or text), transforms mixture-level audio effects embeddings into instrument-wise audio effects embeddings. We evaluated our model across retrieval and audio effects parameter matching tasks, testing its performance across a diverse range of instruments. The results demonstrate that Fx-Encoder++ outperforms previous approaches at mixture level and show a novel ability to extract effects representation instrument-wise, addressing a critical capability gap in intelligent music production systems.

Simulators are useful tools for testing automated driving controllers. Vehicle-in-the-loop (ViL) tests and digital twins (DTs) are widely used simulation technologies to facilitate the smooth deployment of controllers to physical vehicles. However, conventional ViL tests rely on full-size vehicles, requiring large space and high expenses. Also, physical-model-based DT suffers from the reality gap caused by modeling imprecision. This paper develops a comprehensive and practical simulator for testing automated driving controllers enhanced by scaled physical cars and AI-powered DT models. The scaled cars allow for saving space and expenses of simulation tests. The AI-powered DT models ensure superior simulation fidelity. Moreover, the simulator integrates well with off-the-shelf software and control algorithms, making it easy to extend. We use a filtered control benchmark with formal safety guarantees to showcase the capability of the simulator in validating automated driving controllers. Experimental studies are performed to showcase the efficacy of the simulator, implying its great potential in validating control solutions for autonomous vehicles and intelligent traffic.

JoyTTS is an end-to-end spoken chatbot that combines large language models (LLM) with text-to-speech (TTS) technology, featuring voice cloning capabilities. This project is built upon the open-source MiniCPM-o and CosyVoice2 models and trained on 2000 hours of conversational data. We have also provided the complete training code to facilitate further development and optimization by the community. On the testing machine seed-tts-zh, it achieves a SS (speaker similarity) score of 0.73 and a WER (Word Error Rate) of 5.09. The code and models, along with training and inference scripts, are available at this https URL.

We explore unsupervised speech enhancement using diffusion models as expressive generative priors for clean speech. Existing approaches guide the reverse diffusion process using noisy speech through an approximate, noise-perturbed likelihood score, combined with the unconditional score via a trade-off hyperparameter. In this work, we propose two alternative algorithms that directly model the conditional reverse transition distribution of diffusion states. The first method integrates the diffusion prior with the observation model in a principled way, removing the need for hyperparameter tuning. The second defines a diffusion process over the noisy speech itself, yielding a fully tractable and exact likelihood score. Experiments on the WSJ0-QUT and VoiceBank-DEMAND datasets demonstrate improved enhancement metrics and greater robustness to domain shifts compared to both supervised and unsupervised baselines.

Most existing automatic speech recognition (ASR) research evaluate models using in-domain datasets. However, they seldom evaluate how they generalize across diverse speech contexts. This study addresses this gap by benchmarking seven Akan ASR models built on transformer architectures, such as Whisper and Wav2Vec2, using four Akan speech corpora to determine their performance. These datasets encompass various domains, including culturally relevant image descriptions, informal conversations, biblical scripture readings, and spontaneous financial dialogues. A comparison of the word error rate and character error rate highlighted domain dependency, with models performing optimally only within their training domains while showing marked accuracy degradation in mismatched scenarios. This study also identified distinct error behaviors between the Whisper and Wav2Vec2 architectures. Whereas fine-tuned Whisper Akan models led to more fluent but potentially misleading transcription errors, Wav2Vec2 produced more obvious yet less interpretable outputs when encountering unfamiliar inputs. This trade-off between readability and transparency in ASR errors should be considered when selecting architectures for low-resource language (LRL) applications. These findings highlight the need for targeted domain adaptation techniques, adaptive routing strategies, and multilingual training frameworks for Akan and other LRLs.

Green ammonia is emerging as a strategic intermediary within green energy supply chains, serving effectively as both an industrial commodity and hydrogen carrier. This study provides a techno-economic analysis of green ammonia supply chains, comparing cost-effective pathways from global production to European consumers, and evaluates ammonia alongside alternative hydrogen carriers. Gaseous hydrogen consistently remains the most economical import option for Europe, though ammonia holds a narrowing cost advantage over liquid hydrogen (from 16 % in 2030 to 10 % by 2040). Competitive ammonia suppliers, notably Morocco, the United States, and the United Arab Emirates, benefit from low renewable energy costs, with significant price reductions expected by 2040, driven by falling costs for electricity, electrolysers, and conversion technologies. Optimal transport modes vary by consumer demand and distance: trucks are ideal for low demands at all distances, rail for medium ranges, and pipelines for high-demand scenarios. By 2040, ammonia will primarily serve direct-use applications, as hydrogen consumers increasingly shift to direct hydrogen supplies. Policymakers should prioritize pipeline infrastructure for hydrogen distribution, cautiously invest in ammonia's short- to medium-term infrastructure advantages, and limit long-term reliance on ammonia as a hydrogen carrier to mitigate stranded asset risks.

Structural crack detection is a critical task for public safety as it helps in preventing potential structural failures that could endanger lives. Manual detection by inexperienced personnel can be slow, inconsistent, and prone to human error, which may compromise the reliability of assessments. The current study addresses these challenges by introducing a novel deep-learning architecture designed to enhance the accuracy and efficiency of structural crack detection. In this research, various configurations of residual U-Net models were utilized. These models, due to their robustness in capturing fine details, were further integrated into an ensemble with a meta-model comprising convolutional blocks. This unique combination aimed to boost prediction efficiency beyond what individual models could achieve. The ensemble's performance was evaluated against well-established architectures such as SegNet and the traditional U-Net. Results demonstrated that the residual U-Net models outperformed their predecessors, particularly with low-resolution imagery, and the ensemble model exceeded the performance of individual models, proving it as the most effective. The assessment was based on the Intersection over Union (IoU) metric and DICE coefficient. The ensemble model achieved the highest scores, signifying superior accuracy. This advancement suggests way for more reliable automated systems in structural defects monitoring tasks.

In autonomous driving, recent research has increasingly focused on collaborative perception based on deep learning to overcome the limitations of individual perception systems. Although these methods achieve high accuracy, they rely on high communication bandwidth and require unrestricted access to each agent's object detection model architecture and parameters. These constraints pose challenges real-world autonomous driving scenarios, where communication limitations and the need to safeguard proprietary models hinder practical implementation. To address this issue, we introduce a novel late collaborative framework for 3D multi-source and multi-object fusion, which operates solely on shared 3D bounding box attributes-category, size, position, and orientation-without necessitating direct access to detection models. Our framework establishes a new state-of-the-art in late fusion, achieving up to five times lower position error compared to existing methods. Additionally, it reduces scale error by a factor of 7.5 and orientation error by half, all while maintaining perfect 100% precision and recall when fusing detections from heterogeneous perception systems. These results highlight the effectiveness of our approach in addressing real-world collaborative perception challenges, setting a new benchmark for efficient and scalable multi-agent fusion.

Test-Time Adaptation (TTA) has emerged as a promising solution for adapting a source model to unseen medical sites using unlabeled test data, due to the high cost of data annotation. Existing TTA methods consider scenarios where data from one or multiple domains arrives in complete domain units. However, in clinical practice, data usually arrives in domain fragments of arbitrary lengths and in random arrival orders, due to resource constraints and patient variability. This paper investigates a practical Free-Form Test-Time Adaptation (F$^{2}$TTA) task, where a source model is adapted to such free-form domain fragments, with shifts occurring between fragments unpredictably. In this setting, these shifts could distort the adaptation process. To address this problem, we propose a novel Image-level Disentangled Prompt Tuning (I-DiPT) framework. I-DiPT employs an image-invariant prompt to explore domain-invariant representations for mitigating the unpredictable shifts, and an image-specific prompt to adapt the source model to each test image from the incoming fragments. The prompts may suffer from insufficient knowledge representation since only one image is available for training. To overcome this limitation, we first introduce Uncertainty-oriented Masking (UoM), which encourages the prompts to extract sufficient information from the incoming image via masked consistency learning driven by the uncertainty of the source model representations. Then, we further propose a Parallel Graph Distillation (PGD) method that reuses knowledge from historical image-specific and image-invariant prompts through parallel graph networks. Experiments on breast cancer and glaucoma classification demonstrate the superiority of our method over existing TTA approaches in F$^{2}$TTA. Code is available at this https URL.

Shared control combines human intention with autonomous decision-making, from low-level safety overrides to high-level task guidance, enabling systems that adapt to users while ensuring safety and performance. This enhances task effectiveness and user experience across domains such as assistive robotics, teleoperation, and autonomous driving. However, existing shared control methods, based on e.g. Model Predictive Control, Control Barrier Functions, or learning-based control, struggle with feasibility, scalability, or safety guarantees, particularly since the user input is unpredictable. To address these challenges, we propose an assistive controller framework based on Constrained Optimal Control Problem that incorporates an offline-computed Control Invariant Set, enabling online computation of control actions that ensure feasibility, strict constraint satisfaction, and minimal override of user intent. Moreover, the framework can accommodate structured class of non-convex constraints, which are common in real-world scenarios. We validate the approach through a large-scale user study with 66 participants--one of the most extensive in shared control research--using a computer game environment to assess task load, trust, and perceived control, in addition to performance. The results show consistent improvements across all these aspects without compromising safety and user intent.

Current methods for restoring underexposed images typically rely on supervised learning with paired underexposed and well-illuminated images. However, collecting such datasets is often impractical in real-world scenarios. Moreover, these methods can lead to over-enhancement, distorting well-illuminated regions. To address these issues, we propose IGDNet, a Zero-Shot enhancement method that operates solely on a single test image, without requiring guiding priors or training data. IGDNet exhibits strong generalization ability and effectively suppresses noise while restoring illumination. The framework comprises a decomposition module and a denoising module. The former separates the image into illumination and reflection components via a dense connection network, while the latter enhances non-uniformly illuminated regions using an illumination-guided pixel adaptive correction method. A noise pair is generated through downsampling and refined iteratively to produce the final result. Extensive experiments on four public datasets demonstrate that IGDNet significantly improves visual quality under complex lighting conditions. Quantitative results on metrics like PSNR (20.41dB) and SSIM (0.860dB) show that it outperforms 14 state-of-the-art unsupervised methods. The code will be released soon.

Purpose: The primary aim of this study is to enhance fault diagnosis in induction machines by leveraging the Padé Approximant Neuron (PAON) model. While accelerometers and microphones are standard in motor condition monitoring, deep learning models with nonlinear neuron architectures offer promising improvements in diagnostic performance. This research addresses the question: Can Padé Approximant Neural Networks (PadéNets) outperform conventional Convolutional Neural Networks (CNNs) and Self-Organized Operational Neural Networks (Self-ONNs) in diagnosing electrical and mechanical faults using vibration and acoustic data? Methods: We evaluate and compare the diagnostic capabilities of three deep learning architectures: one-dimensional CNNs, Self-ONNs, and PadéNets. These models are tested on the University of Ottawa's publicly available constant-speed induction motor datasets, which include both vibration and acoustic sensor data. The PadéNet model is designed to introduce enhanced nonlinearity and is compatible with unbounded activation functions such as Leaky ReLU. Results and Conclusion: PadéNets consistently outperformed the baseline models, achieving diagnostic accuracies of 99.96%, 98.26%, 97.61%, and 98.33% for accelerometers 1, 2, 3, and the acoustic sensor, respectively. The enhanced nonlinearity of PadéNets, together with their compatibility with unbounded activation functions, significantly improves fault diagnosis performance in induction motor condition monitoring.

The rapid advancement of speech generation models has heightened privacy and security concerns related to voice cloning (VC). Recent studies have investigated disrupting unauthorized voice cloning by introducing adversarial perturbations. However, determined attackers can mitigate these protective perturbations and successfully execute VC. In this study, we conduct the first systematic evaluation of these protective perturbations against VC under realistic threat models that include perturbation purification. Our findings reveal that while existing purification methods can neutralize a considerable portion of the protective perturbations, they still lead to distortions in the feature space of VC models, which degrades the performance of VC. From this perspective, we propose a novel two-stage purification method: (1) Purify the perturbed speech; (2) Refine it using phoneme guidance to align it with the clean speech distribution. Experimental results demonstrate that our method outperforms state-of-the-art purification methods in disrupting VC defenses. Our study reveals the limitations of adversarial perturbation-based VC defenses and underscores the urgent need for more robust solutions to mitigate the security and privacy risks posed by VC. The code and audio samples are available at this https URL.

This paper presents a complete signal-processing chain for multistatic integrated sensing and communications (ISAC) using 5G Positioning Reference Signal (PRS). We consider a distributed architecture in which one gNB transmits a periodic OFDM-PRS waveform while multiple spatially separated receivers exploit the same signal for target detection, parameter estimation and tracking. A coherent cross-ambiguity function (CAF) is evaluated to form a range-Doppler map from which the bistatic delay and radial velocity are extracted for every target. For a single target, the resulting bistatic delays are fused through nonlinear least-squares trilateration, yielding a geometric position estimate, and a regularized linear inversion of the radial-speed equations yields a two-dimensional velocity vector, where speed and heading are obtained. The approach is applied to 2D and 3D settings, extended to account for time synchronization bias, and generalized to multiple targets by resolving target association. The sequence of position-velocity estimates is then fed to standard and extended Kalman filters to obtain smoothed tracks. Our results show high-fidelity moving-target detection, positioning, and tracking using 5G PRS signals for multistatic ISAC.

This paper proposes a novel non-anticipatory long-short-term coordinated dispatch framework for isolated microgrid with hybrid short-long-duration energy storages (LDES). We introduce a convex hull approximation model for nonconvex LDES electrochemical dynamics, facilitating computational tractability and accuracy. To address temporal coupling in SoC dynamics and long-term contracts, we generate hindsight-optimal state-of-charge (SoC) trajectories of LDES and netloads for offline training. In the online stage, we employ kernel regression to dynamically update the SoC reference and propose an adaptive online convex optimization (OCO) algorithm with SoC reference tracking and expert tracking to mitigate myopia and enable adaptive step-size optimization. We rigorously prove that both long-term and short-term policies achieve sublinear regret bounds over time, which improves with more regression scenarios, stronger tracking penalties, and finer convex approximations. Simulation results show that the proposed method outperforms state-of-the-art methods, reducing costs by 73.4%, eliminating load loss via reference tracking, and achieving an additional 2.4% cost saving via the OCO algorithm. These benefits scale up with longer LDES durations, and the method demonstrates resilience to poor forecasts and unexpected system faults.

We formulate a generalization of the Restricted Isometry Property (RIP) referred to as the Restricted Quasiconvexity Isometry Property (RQIP) for alpha stable random projections with $0<\alpha<1$. A lower bound on the number of rows for RQIP to hold for random matrices whose entries are drawn from a symmetric $\alpha$-stable ($S\alpha S$) distribution is derived. The proof leverages two key components: a concentration inequality for empirical fractional moments of $S\alpha S$ variables and a covering number bound for sparse $\ell_\alpha$ balls. The resulting sample complexity reflects the polynomial tail behavior of the concentration and reinforces an observation made in the literature that the RIP framework may have to be replaced with other sparse recovery formulations in practice, such as those based on the null space property.

In recent advancements in audio self-supervised representation learning, the standard Transformer architecture has emerged as the predominant approach, yet its attention mechanism often allocates a portion of attention weights to irrelevant information, potentially impairing the model's discriminative ability. To address this, we introduce a differential attention mechanism, which effectively mitigates ineffective attention allocation through the integration of dual-softmax operations and appropriately tuned differential coefficients. Experimental results demonstrate that our ASDA model achieves state-of-the-art (SOTA) performance across multiple benchmarks, including audio classification (49.0% mAP on AS-2M, 41.5% mAP on AS20K), keyword spotting (98.3% accuracy on SPC-2), and environmental sound classification (96.1% accuracy on ESC-50). These results highlight ASDA's effectiveness in audio tasks, paving the way for broader applications.

Deep Learning (DL) has revolutionized medical imaging, yet its adoption is constrained by data scarcity and privacy regulations, limiting access to diverse datasets. Federated Learning (FL) enables decentralized training but suffers from high communication costs and is often restricted to a single downstream task, reducing flexibility. We propose a data-sharing method via Differentially Private (DP) generative models. By adopting foundation models, we extract compact, informative embeddings, reducing redundancy and lowering computational overhead. Clients collaboratively train a Differentially Private Conditional Variational Autoencoder (DP-CVAE) to model a global, privacy-aware data distribution, supporting diverse downstream tasks. Our approach, validated across multiple feature extractors, enhances privacy, scalability, and efficiency, outperforming traditional FL classifiers while ensuring differential privacy. Additionally, DP-CVAE produces higher-fidelity embeddings than DP-CGAN while requiring $5{\times}$ fewer parameters.

Future wireless networks are expected to incorporate diverse services that often lack general mathematical models. To address such black-box network management tasks, the large language model (LLM) optimizer framework, which leverages pretrained LLMs as optimization agents, has recently been promoted as a promising solution. This framework utilizes natural language prompts describing the given optimization problems along with past solutions generated by LLMs themselves. As a result, LLMs can obtain efficient solutions autonomously without knowing the mathematical models of the objective functions. Although the viability of the LLM optimizer (LLMO) framework has been studied in various black-box scenarios, it has so far been limited to numerical simulations. For the first time, this paper establishes a theoretical foundation for the LLMO framework. With careful investigations of LLM inference steps, we can interpret the LLMO procedure as a finite-state Markov chain, and prove the convergence of the framework. Our results are extended to a more advanced multiple LLM architecture, where the impact of multiple LLMs is rigorously verified in terms of the convergence rate. Comprehensive numerical simulations validate our theoretical results and provide a deeper understanding of the underlying mechanisms of the LLMO framework.

This article focuses on integrating path-planning and control with specializing on the unique needs of robotic unicycles. A unicycle design is presented which is capable of accelerating/breaking and carrying out a variety of maneuvers. The proposed path-planning method segments the path into straight and curved path sections dedicated for accelerating/breaking and turning maneuvers, respectively. The curvature profiles of the curved sections are optimized while considering the control performance and the slipping limits of the wheel. The performance of the proposed integrated approach is demonstrated via numerical simulations.

Integrated sensing and communication (ISAC) is a key feature of future cellular systems, enabling applications such as intruder detection, monitoring, and tracking using the same infrastructure. However, its potential for structural health monitoring (SHM), which requires the detection of slow and subtle structural changes, remains largely unexplored due to challenges such as multipath interference and the need for ultra-high sensing precision. This study introduces a novel theoretical framework for SHM via ISAC by leveraging reconfigurable intelligent surfaces (RIS) as reference points in collaboration with base stations and users. By dynamically adjusting RIS phases to generate distinct radio signals that suppress background multipath interference, measurement accuracy at these reference points is enhanced. We theoretically analyze RIS-aided collaborative sensing in three-dimensional cellular networks using Fisher information theory, demonstrating how increasing observation time, incorporating additional receivers (even with self-positioning errors), optimizing RIS phases, and refining collaborative node selection can reduce the position error bound to meet SHM's stringent accuracy requirements. Furthermore, we develop a Bayesian inference model to identify structural states and validate damage detection probabilities. Both theoretical and numerical analyses confirm ISAC's capability for millimeter-level deformation detection, highlighting its potential for high-precision SHM applications.

Efficient chip-scale interconnects are essential for modern microelectronic-photonic systems, supporting high bandwidth and low-latency processing. Traditional wired links face high resistivity and latency, while millimeter-wave wireless solutions suffer from bandwidth congestion and interference. Terahertz (THz) plasmonic communication, based on surface plasmon polaritons (SPPs), offers high data rates and broad bandwidth, and is compatible with nanophotonic platforms. This work introduces a Binary Field-Driven Meta-Routing Method supported by a semi-analytical framework that models the tunable interaction between THz plasmonic phenomena and graphene's electromagnetic properties. By modulating graphene's impedance, the method enables dynamic coupling and routing of localized surface plasmon resonances (LSPRs) across a meta-network, facilitating real-time beam steering in chip-scale systems. Combining analytical conductivity models, coupled-mode theory, and algorithmic control, the approach enables predictive configuration of LSPR-based steering in reconfigurable graphene metasurfaces. Four meta-pixel antenna configurations Y-MetaRouter, MetaSwitcher, Penta-MetaEmitter, and CP-MetaCore are designed to support unidirectional radiation, bi-directional steering, frequency-driven transitions, and circular polarization, respectively. Chemical potential modulation creates reconfigurable LSPR pathways and virtual SPP channels. A Coupled-Mode Theory for Field-Driven LSPR Meta-Networks is proposed to model current distributions and predict far-field characteristics. Results show strong agreement between theory and full-wave simulations. A point-to-point meta-wireless link is analyzed, demonstrating scalability for low-latency, high-performance THz communication in WiNoC and chiplet applications. System-level metrics confirm feasibility for space-constrained, high-speed interconnects.

Recent advancements in the reasoning capabilities of large language models (LLMs) show that employing group relative policy optimization (GRPO) algorithm for reinforcement learning (RL) training allows the models to use more thinking/reasoning tokens for generating better responses. However, LLMs can generate only a finite amount of tokens while maintaining attention to the previously generated tokens. This limit, also known as the context size of an LLM, is a bottleneck in LLM reasoning with arbitrarily large number of tokens. To think beyond the limit of context size, an LLM must employ a modular thinking strategy to reason over multiple rounds. In this work, we propose $\textbf{MOTIF: Modular Thinking via Reinforcement Finetuning}$ -- an RL training method for generating thinking tokens in multiple rounds, effectively allowing the model to think with additional context size. We trained the open-source model Qwen2.5-3B-Instruct on GSM8K dataset via parameter efficient fine-tuning and tested its accuracy on MATH500 and AIME2024 benchmarks. Our experiments show 3.8\% and 3.3\% improvements over vanilla GRPO based training in the respective benchmarks. Furthermore, this improvement was achieved with only 15\% of samples, thus demonstrating sample efficiency of MOTIF. Our code and models are available at this https URL and this https URL, respectively.

Controlling inter-area oscillation (IAO) across wide areas is crucial for the stability of modern power systems. Recent advances in deep learning, combined with the extensive deployment of phasor measurement units (PMUs) and generator sensors, have catalyzed the development of data-driven IAO damping controllers. In this paper, a novel IAO damping control framework is presented by modeling the control problem as a Markov Decision Process (MDP) and solving it through deep reinforcement learning (DRL). The DRL-based controller is trained in the state space with continuous eigenvalue configurations. To optimize control performance and cost-efficiency, only a subset of generators, identified by global participation factors, are selected for control. In addition, a switching control strategy (SCS) is introduced that effectively integrates the DRL-based controller with power system stabilizers (PSSs) to enhance overall performance. The simulation results on the IEEE 39-bus New England power system show that the proposed method outperforms two benchmark methods regarding the transient response. The DRL-based controller trained on the linear state-space environment can be directly tested in the nonlinear differential-algebraic environment. The robustness of the proposed method against communication delays has been thoroughly investigated.

The investigation reported in this document focuses on identifying systems with symmetries using equivariant autoregressive reservoir computers. General results in structured matrix approximation theory are presented, exploring a two-fold approach. Firstly, a comprehensive examination of generic symmetry-preserving nonlinear time delay embedding is conducted. This involves analyzing time series data sampled from an equivariant system under study. Secondly, sparse least-squares methods are applied to discern approximate representations of the output coupling matrices. These matrices play a critical role in determining the nonlinear autoregressive representation of an equivariant system. The structural characteristics of these matrices are dictated by the set of symmetries inherent in the system. The document outlines prototypical algorithms derived from the described techniques, offering insight into their practical applications. Emphasis is placed on the significant improvement on structured identification precision when compared to classical reservoir computing methods for the simulation of equivariant dynamical systems.

One of the primary catalysts fueling advances in artificial intelligence (AI) and machine learning (ML) is the availability of massive, curated datasets. A commonly used technique to curate such massive datasets is crowdsourcing, where data are dispatched to multiple annotators. The annotator-produced labels are then fused to serve downstream learning and inference tasks. This annotation process often creates noisy labels due to various reasons, such as the limited expertise, or unreliability of annotators, among others. Therefore, a core objective in crowdsourcing is to develop methods that effectively mitigate the negative impact of such label noise on learning tasks. This feature article introduces advances in learning from noisy crowdsourced labels. The focus is on key crowdsourcing models and their methodological treatments, from classical statistical models to recent deep learning-based approaches, emphasizing analytical insights and algorithmic developments. In particular, this article reviews the connections between signal processing (SP) theory and methods, such as identifiability of tensor and nonnegative matrix factorization, and novel, principled solutions of longstanding challenges in crowdsourcing -- showing how SP perspectives drive the advancements of this field. Furthermore, this article touches upon emerging topics that are critical for developing cutting-edge AI/ML systems, such as crowdsourcing in reinforcement learning with human feedback (RLHF) and direct preference optimization (DPO) that are key techniques for fine-tuning large language models (LLMs).

Incurable diseases continue to pose major challenges to global healthcare systems, with their prevalence shaped by lifestyle, economic, social, and genetic factors. Among these, kidney disease remains a critical global health issue, requiring ongoing research to improve early diagnosis and treatment. In recent years, deep learning (DL) has shown promise in medical imaging and diagnostics, driving significant progress in automatic kidney cancer (KC) detection. However, the success of DL models depends heavily on the availability of high-quality, domain-specific datasets, which are often limited and expensive to acquire. Moreover, DL models demand substantial computational power and storage, restricting their real-world clinical use. To overcome these barriers, transfer learning (TL) has emerged as an effective approach, enabling the reuse of pre-trained models from related domains to enhance KC diagnosis. This paper presents a comprehensive survey of DL-based TL frameworks for KC detection, systematically reviewing key methodologies, their advantages, and limitations, and analyzing their practical performance. It further discusses challenges in applying TL to medical imaging and highlights emerging trends that could influence future research. This review demonstrates the transformative role of TL in precision medicine, particularly oncology, by improving diagnostic accuracy, lowering computational demands, and supporting the integration of AI-powered tools in healthcare. The insights provided offer valuable guidance for researchers and practitioners, paving the way for future advances in KC diagnostics and personalized treatment strategies.

Deep Learning (DL) in neuroimaging has become increasingly relevant for detecting neurological conditions and neurodegenerative disorders. One of the most predominant biomarkers in neuroimaging is represented by brain age, which has been shown to be a good indicator for different conditions, such as Alzheimer's Disease. Using brain age for weakly supervised pre-training of DL models in transfer learning settings has also recently shown promising results, especially when dealing with data scarcity of different conditions. On the other hand, anatomical information of brain MRIs (e.g. cortical thickness) can provide important information for learning good representations that can be transferred to many downstream tasks. In this work, we propose AnatCL, an anatomical foundation model for brain MRIs that i.) leverages anatomical information in a weakly contrastive learning approach, and ii.) achieves state-of-the-art performances across many different downstream tasks. To validate our approach we consider 12 different downstream tasks for the diagnosis of different conditions such as Alzheimer's Disease, autism spectrum disorder, and schizophrenia. Furthermore, we also target the prediction of 10 different clinical assessment scores using structural MRI data. Our findings show that incorporating anatomical information during pre-training leads to more robust and generalizable representations. Pre-trained models can be found at: this https URL.

Recently, medical image synthesis gains more and more popularity, along with the rapid development of generative models. Medical image synthesis aims to generate an unacquired image modality, often from other observed data modalities. Synthesized images can be used for clinical diagnostic assistance, data augmentation for model training and validation or image quality improving. In the meanwhile, the flow-based models are among the successful generative models for the ability of generating realistic and high-quality synthetic images. However, most flow-based models require to calculate flow ordinary different equation (ODE) evolution steps in synthesis process, for which the performances are significantly limited by heavy computation time due to a large number of time iterations. In this paper, we propose a novel flow-based model, namely bi-directional Discrete Process Matching (Bi-DPM) to accomplish the bi-modality image synthesis tasks. Different to other flow matching based models, we propose to utilize both forward and backward ODE flows and enhance the consistency on the intermediate images over a few discrete time steps, resulting in a synthesis process maintaining high-quality generations for both modalities under the guidance of paired data. Our experiments on three datasets of MRI T1/T2 and CT/MRI demonstrate that Bi-DPM outperforms other state-of-the-art flow-based methods for bi-modality image synthesis, delivering higher image quality with accurate anatomical regions.

Vibration-based damage detection relies on identifying modal parameters, such as natural frequencies and mode shapes, to be used as structural integrity indicators. However, traditional comparisons of these parameters are often ambiguous in complex systems, complicating damage detection and assessment. The modified total modal assurance criterion (MTMAC), a metric well-known in the field of finite element model updating, is extended to address this challenge and is proposed as a metric for damage identification and severity assessment. To support the requirement for precise and robust modal identification of Structural Health Monitoring (SHM), the improved Loewner Framework (iLF), known for its reliability and computational performance, is pioneeringly employed within SHM. Since the MTMAC is proposed solely as a damage identification and severity assessment metric, the coordinate modal assurance criterion (COMAC), also a well-established tool, but for damage localisation using mode shapes, is used for completeness. The iLF SHM capabilities are validated through comparisons with traditional methods, including least-squares complex exponential (LSCE) and stochastic subspace identification with canonical variate analysis (SSI-CVA) on a numerical case study of a cantilever beam. Furthermore, the MTMAC is validated against the traditional vibration-based approach, which involves directly comparing natural frequencies and mode shapes. Finally, an experimental dataset from a BAE Systems Hawk T1A trainer jet ground vibration tests is used to demonstrate the iLF and MTMAC capabilities on real-life, real-size SHM problem, showing their effectiveness in detecting and assessing damage.

Recent developments in Self-Supervised Learning (SSL) have demonstrated significant potential for Speaker Verification (SV), but closing the performance gap with supervised systems remains an ongoing challenge. SSL frameworks rely on anchor-positive pairs, constructed from segments of the same audio utterance. Hence, positives have channel characteristics similar to those of their corresponding anchors, even with extensive data-augmentation. Therefore, this positive sampling strategy is a fundamental limitation as it encodes too much information regarding the recording source in the learned representations. This article introduces Self-Supervised Positive Sampling (SSPS), a bootstrapped technique for sampling appropriate and diverse positives in SSL frameworks for SV. SSPS samples positives close to their anchor in the representation space, assuming that these pseudo-positives belong to the same speaker identity but correspond to different recording conditions. This method consistently demonstrates improvements in SV performance on VoxCeleb benchmarks when applied to major SSL frameworks, including SimCLR, SwAV, VICReg, and DINO. Using SSPS, SimCLR and DINO achieve 2.57% and 2.53% EER on VoxCeleb1-O, respectively. SimCLR yields a 58% relative reduction in EER, getting comparable performance to DINO with a simpler training framework. Furthermore, SSPS lowers intra-class variance and reduces channel information in speaker representations while exhibiting greater robustness without data-augmentation.

Integrated sensing and communication (ISAC) has garnered significant attention in recent years. In this paper, we delve into the topic of sensing-assisted communication within ISAC systems. More specifically, a novel sensing-assisted channel estimation scheme is proposed for bistatic orthogonal-frequency-division-multiplexing (OFDM) ISAC systems. A framework of sensing-assisted channel estimator is first developed, integrating a tailored low-complexity sensing algorithm to facilitate real-time channel estimation and decoding. To address the potential sensing errors caused by low-complexity sensing algorithms, a sensing-assisted linear minimum mean square error (LMMSE) estimation algorithm is then developed. This algorithm incorporates tolerance factors designed to account for deviations between estimated and true channel parameters, enabling the construction of robust correlation matrices for LMMSE estimation. Additionally, we establish a systematic mechanism for determining these tolerance factors. A comprehensive analysis of the normalized mean square error (NMSE) performance and computational complexity is finally conducted, providing valuable insights into the selection of the estimator's parameters. The effectiveness of our proposed scheme is validated by extensive simulations. Compared to existing methods, our proposed scheme demonstrates superior performance, particularly in high signal-to-noise ratio (SNR) regions or with large bandwidths, while maintaining low computational complexity.

Illumination estimation remains a pivotal challenge in image processing, particularly for robotics, where robust environmental perception is essential under varying lighting conditions. Traditional approaches, such as RGB histograms and GIST descriptors, often fail in complex scenarios due to their sensitivity to illumination changes. This study introduces a novel method utilizing the Wasserstein distance, rooted in optimal transport theory, to estimate illuminant and light direction in images. Experiments on diverse images indoor scenes, black-and-white photographs, and night images demonstrate the method's efficacy in detecting dominant light sources and estimating their directions, outperforming traditional statistical methods in complex lighting environments. The approach shows promise for applications in light source localization, image quality assessment, and object detection enhancement. Future research may explore adaptive thresholding and integrate gradient analysis to enhance accuracy, offering a scalable solution for real-world illumination challenges in robotics and beyond.

The pinching-antenna system (PASS) introduces new degrees of freedom (DoFs) for physical layer security (PLS) through pinching beamforming. In this paper, a couple of scenarios for secure beamforming for PASS are studied. 1) For the case with a single legitimate user (Bob) and a single eavesdropper (Eve), a closed-form expression for the optimal baseband beamformer is derived. On this basis, a gradient-based method is proposed to optimize the activated positions of pinching antennas (PAs). 2) For the case with multiple Bobs and multiple Eves, a fractional programming (FP)-based block coordinate descent (BCD) algorithm, termed FP-BCD, is proposed for optimizing the weighted secrecy sum-rate (WSSR). Specifically, a closed-form baseband beamformer is obtained via Lagrange multiplier method. Furthermore, owing to the non-convex objective function exhibiting numerous stationary points, a low-complexity one-dimensional search is used to find a high-quality solution of the PAs' activated locations. Numerical results are provided to demonstrate that: i) All proposed algorithms achieve stable convergence within a few iterations, ii) across all considered power ranges, the FP-BCD algorithm outperforms baseline methods using zero-forcing (ZF) and maximal-ratio transmission (MRT) beamforming in terms of the WSSR, and iii) PASS achieves a significantly higher secrecy rate than traditional fixed-antenna systems.

Artificial intelligence (AI) is increasingly being used for medical imaging tasks. However, there can be biases in AI models, particularly when they are trained using imbalanced training datasets. One such example has been the strong ethnicity bias effect in cardiac magnetic resonance (CMR) image segmentation models. Although this phenomenon has been reported in a number of publications, little is known about the effectiveness of bias mitigation algorithms in this domain. We aim to investigate the impact of common bias mitigation methods to address bias between Black and White subjects in AI-based CMR segmentation models. Specifically, we use oversampling, importance reweighing and Group DRO as well as combinations of these techniques to mitigate the ethnicity bias. Second, motivated by recent findings on the root causes of AI-based CMR segmentation bias, we evaluate the same methods using models trained and evaluated on cropped CMR images. We find that bias can be mitigated using oversampling, significantly improving performance for the underrepresented Black subjects whilst not significantly reducing the majority White subjects' performance. Using cropped images increases performance for both ethnicities and reduces the bias, whilst adding oversampling as a bias mitigation technique with cropped images reduces the bias further. When testing the models on an external clinical validation set, we find high segmentation performance and no statistically significant bias.

Approximating nonlinear systems as linear ones is a common workaround to apply control tools tailored for linear systems. This motivates our present work where we developed a data-driven model predictive controller (MPC) based on the Koopman operator framework, allowing the embedding of nonlinear dynamics in a higher dimensional, but linear function space. The controller, termed adaptive Koopman model predictive control (KMPC), uses online closed-loop feedback to learn and incrementally update a linear representation of nonlinear system dynamics, without the prior knowledge of a model. Adaptive KMPC differs from most other Koopman-based control frameworks that aim to identify high-validity-range models in advance and then enter closed-loop control without further model adaptations. To validate the controller, trajectory tracking experiments are conducted with 1R and 2R robots under force disturbances and changing model parameters. We compare the controller to classical linearization MPC and Koopman-based MPC without model updates, denoted static KMPC. The results show that adaptive KMPC can, opposed to static KMPC, generalize over unforeseen force disturbances and can, opposed to linearization MPC, handle varying dynamic parameters, while using a small set of basis functions to approximate the Koopman operator.

To address the computational challenges of Model Predictive Control (MPC), recent research has studied using imitation learning to approximate the MPC to a computationally efficient Deep Neural Network (DNN). However, this introduces a common issue in learning-based control, the simulation-to-reality (sim-to-real) gap, and Domain Randomization (DR) has been widely used to mitigate this gap by introducing perturbations in the source domain. However, DR inevitably leads to low data collection efficiency and an overly conservative control strategy. This study proposes a new control framework that addresses this issue from a control perspective inspired by Robust Tube MPC. The framework ensures the DNN operates in the same environment as the source domain, handling the sim-to-real gap with great data collection efficiency. Moreover, a parameter governor is introduced to address the DNN's inability to adapt to variations in model parameters, enabling the system to satisfy MPC constraints more robustly under changing conditions. The proposed framework was validated through two case studies: cart-pole control and vehicle collision avoidance control, which analyzed the principles of the proposed framework in detail and demonstrated its application to a vehicle control case.

With the exponential growth of the unmanned aerial vehicle (UAV) industry and a broad range of applications expected to appear in the coming years, the employment of traditional radar systems is becoming increasingly cumbersome for UAV supervision. Motivated by this emerging challenge, this paper investigates the feasibility of employing integrated sensing and communication (ISAC) systems implemented over current and future wireless networks to perform this task. We propose a sensing mechanism based on the synchronization signal block (SSB) in the fifth-generation (5G) standard that performs sensing in a passive bistatic setting. By assuming planar arrays at the sensing nodes and according to the 5G standard, we consider that the SSB signal is sent in a grid of orthogonal beams that are multiplexed in time, with some of them pointing toward a surveillance region where low-altitude drones can be flying. The Cramer-Rao Bound (CRB) is derived as the theoretical bound for range and velocity estimation. Our results demonstrate the potential of employing SSB signals for UAV-like target localization at low SNR.

Magnetic Resonance Imaging (MRI) is highly susceptible to motion artifacts due to the extended acquisition times required for k-space sampling. These artifacts can compromise diagnostic utility, particularly for dynamic imaging. We propose a novel alternating minimization framework that leverages a bespoke diffusion model to jointly reconstruct and correct non-rigid motion-corrupted k-space data. The diffusion model uses a coarse-to-fine denoising strategy to capture large overall motion and reconstruct the lower frequencies of the image first, providing a better inductive bias for motion estimation than that of standard diffusion models. We demonstrate the performance of our approach on both real-world cine cardiac MRI datasets and complex simulated rigid and non-rigid deformations, even when each motion state is undersampled by a factor of 64x. Additionally, our method is agnostic to sampling patterns, anatomical variations, and MRI scanning protocols, as long as some low frequency components are sampled during each motion state.

Low-altitude economy (LAE) represents an emerging economic paradigm that redefines commercial and social aerial activities. Large artificial intelligence models (LAIMs) offer transformative potential to further enhance the intelligence of LAE services. However, deploying LAIMs in LAE poses several challenges, including the significant gap between their computational/storage demands and the limited onboard resources of LAE entities, the mismatch between lab-trained LAIMs and dynamic physical environments, and the inefficiencies of traditional decoupled designs for sensing, communication, and computation. To address these issues, we first propose a hierarchical system architecture tailored for LAIM deployment and present representative LAE application scenarios. Next, we explore key enabling techniques that facilitate the mutual co-evolution of LAIMs and low-altitude systems, and introduce a task-oriented execution pipeline for scalable and adaptive service delivery. Then, the proposed framework is validated through real-world case studies. Finally, we outline open challenges to inspire future research.

In practical systems, recorded analog signals must be digitized for processing, introducing quantization as a critical aspect of data acquisition. While prior studies have examined quantization effects in various signal processing contexts, its impact on differential microphone arrays (DMAs), particularly in one-dimensional (1D) first-order configurations, remains unexplored. This paper investigates the influence of signal quantization on the performance of first-order 1D-DMAs across various beampatterns. An analytical expression for quantized beamformed output for a first-order 1D-DMA has been formulated. The effect of signal quantization has been studied on array performance measures such as the Beampattern, Directivity Factor (DF), Front-to-Back Ratio (FBR), and suppression depth at null points (SDN). Simulation results reveal that the beampattern shape remains structurally invariant across quantization bit depths, with quantization primarily affecting SDN. DF and FBR remain constant with the varying number of quantization bits. Additionally, SDN is shown to be frequency-independent, however, it increases with increasing quantization bit depths, enhancing interference suppression. The study also examines the effect of steering nulls across the azimuthal range, showing that SDN degrades as the null moves closer to the source look direction (0 degree), indicating reduced interference suppression.

Bias field artifacts in magnetic resonance imaging (MRI) scans introduce spatially smooth intensity inhomogeneities that degrade image quality and hinder downstream analysis. To address this challenge, we propose a novel variational Hadamard U-Net (VHU-Net) for effective body MRI bias field correction. The encoder comprises multiple convolutional Hadamard transform blocks (ConvHTBlocks), each integrating convolutional layers with a Hadamard transform (HT) layer. Specifically, the HT layer performs channel-wise frequency decomposition to isolate low-frequency components, while a subsequent scaling layer and semi-soft thresholding mechanism suppress redundant high-frequency noise. To compensate for the HT layer's inability to model inter-channel dependencies, the decoder incorporates an inverse HT-reconstructed transformer block, enabling global, frequency-aware attention for the recovery of spatially consistent bias fields. The stacked decoder ConvHTBlocks further enhance the capacity to reconstruct the underlying ground-truth bias field. Building on the principles of variational inference, we formulate a new evidence lower bound (ELBO) as the training objective, promoting sparsity in the latent space while ensuring accurate bias field estimation. Comprehensive experiments on abdominal and prostate MRI datasets demonstrate the superiority of VHU-Net over existing state-of-the-art methods in terms of intensity uniformity, signal fidelity, and tissue contrast. Moreover, the corrected images yield substantial downstream improvements in segmentation accuracy. Our framework offers computational efficiency, interpretability, and robust performance across multi-center datasets, making it suitable for clinical deployment.

Recent developments in the Internet of Bio-Nano Things (IoBNT) are laying the groundwork for innovative applications across the healthcare sector. Nanodevices designed to operate within the body, managed remotely via the internet, are envisioned to promptly detect and actuate on potential diseases. In this vision, an inherent challenge arises due to the limited capabilities of individual nanosensors; specifically, nanosensors must communicate with one another to collaborate as a cluster. Aiming to research the boundaries of the clustering capabilities, this survey emphasizes data-driven communication strategies in molecular communication (MC) channels as a means of linking nanosensors. Relying on the flexibility and robustness of machine learning (ML) methods to tackle the dynamic nature of MC channels, the MC research community frequently refers to neural network (NN) architectures. This interdisciplinary research field encompasses various aspects, including the use of NNs to facilitate communication in MC environments, their implementation at the nanoscale, explainable approaches for NNs, and dataset generation for training. Within this survey, we provide a comprehensive analysis of fundamental perspectives on recent trends in NN architectures for MC, the feasibility of their implementation at the nanoscale, applied explainable artificial intelligence (XAI) techniques, and the accessibility of datasets along with best practices for their generation. Additionally, we offer open-source code repositories that illustrate NN-based methods to support reproducible research for key MC scenarios. Finally, we identify emerging research challenges, such as robust NN architectures, biologically integrated NN modules, and scalable training strategies.

Mitral regurgitation is one of the most prevalent cardiac disorders. Four-dimensional (4D) ultrasound has emerged as the primary imaging modality for assessing dynamic valvular morphology. However, 4D mitral valve (MV) analysis remains challenging due to limited phase annotations, severe motion artifacts, and poor imaging quality. Yet, the absence of inter-phase dependency in existing methods hinders 4D MV analysis. To bridge this gap, we propose a Motion-Topology guided consistency network (MTCNet) for accurate 4D MV ultrasound segmentation in semi-supervised learning (SSL). MTCNet requires only sparse end-diastolic and end-systolic annotations. First, we design a cross-phase motion-guided consistency learning strategy, utilizing a bi-directional attention memory bank to propagate spatio-temporal features. This enables MTCNet to achieve excellent performance both per- and inter-phase. Second, we devise a novel topology-guided correlation regularization that explores physical prior knowledge to maintain anatomically plausible. Therefore, MTCNet can effectively leverage structural correspondence between labeled and unlabeled phases. Extensive evaluations on the first largest 4D MV dataset, with 1408 phases from 160 patients, show that MTCNet performs superior cross-phase consistency compared to other advanced methods (Dice: 87.30%, HD: 1.75mm). Both the code and the dataset are available at this https URL.

Vertical Take-Off and Landing (VTOL) vehicles are gaining traction in both the delivery drone market and passenger transportation, driving the development of Urban Air Mobility (UAM) systems. UAM seeks to alleviate road congestion in dense urban areas by leveraging urban airspace. To handle UAM traffic, vertiport terminals (vertiminals) play a critical role in supporting VTOL vehicle operations such as take-offs, landings, taxiing, passenger boarding, refueling or charging, and maintenance. Efficient scheduling algorithms are essential to manage these operations and optimize vertiminal throughput while ensuring safety protocols. Unlike fixed-wing aircraft, which rely on runways for take-off and climbing in fixed directions, VTOL vehicles can utilize multiple surface directions for climbing and approach. This flexibility necessitates specialized scheduling methods. We propose a Mixed Integer Linear Program (MILP) formulation to holistically optimize vertiminal operations, including taxiing, climbing (or approach) using multiple directions, and turnaround at gates. The proposed MILP reduces delays by up to 50%. Additionally, we derive equations to compute upper bounds of the throughput capacity of vertiminals, considering its core elements: the TLOF pad system, taxiway system, and gate system. Our results demonstrate that the MILP achieves throughput levels consistent with the theoretical maximum derived from these equations. We also validate our framework through a case study using a well-established vertiminal topology from the literature. Our MILP can be used to find the optimal configuration of vertiminal. This dual approach, MILP and throughput analysis, allows for comprehensive capacity analysis without requiring simulations while enabling efficient scheduling through the MILP formulation.

Modern speaker verification systems primarily rely on speaker embeddings, followed by verification based on cosine similarity between the embedding vectors of the enrollment and test utterances. While effective, these methods struggle with multi-talker speech due to the unidentifiability of embedding vectors. In this paper, we propose Neural Scoring (NS), a refreshed end-to-end framework that directly estimates verification posterior probabilities without relying on test-side embeddings, making it more robust to complex conditions, e.g., with multiple talkers. To make the training of such an end-to-end model more efficient, we introduce a large-scale trial e2e training (LtE2E) strategy, where each test utterance pairs with a set of enrolled speakers, thus enabling the processing of large-scale verification trials per batch. Experiments on the VoxCeleb dataset demonstrate that NS consistently outperforms both the baseline and competitive methods across various conditions, achieving an overall 70.36% reduction in EER compared to the baseline.

The compression of real-world scanned 3D human dynamic meshes is an emerging research area, driven by applications such as telepresence, virtual reality, and 3D digital streaming. Unlike synthesized dynamic meshes with fixed topology, scanned dynamic meshes often not only have varying topology across frames but also scan defects such as holes and outliers, increasing the complexity of prediction and compression. Additionally, human meshes often combine rigid and non-rigid motions, making accurate prediction and encoding significantly more difficult compared to objects that exhibit purely rigid motion. To address these challenges, we propose a compression method designed for real-world scanned human dynamic meshes, leveraging embedded key nodes. The temporal motion of each vertex is formulated as a distance-weighted combination of transformations from neighboring key nodes, requiring the transmission of solely the key nodes' transformations. To enhance the quality of the KeyNode-driven prediction, we introduce an octree-based residual coding scheme and a Dual-direction prediction mode, which uses I-frames from both directions. Extensive experiments demonstrate that our method achieves significant improvements over the state-of-the-art, with an average bitrate savings of 58.43% across the evaluated sequences, particularly excelling at low bitrates.

Movable antenna (MA) and intelligent reflecting surface (IRS) are considered promising technologies for the next-generation wireless communication systems due to their shared channel reconfiguration capabilities. This, however, raises a fundamental question: Does the performance gain of MAs over conventional fixed-position antennas (FPAs) still exist in the presence of the IRS? To answer this question, we investigate in this paper an IRS-assisted multi-user multiple-input single-output (MISO) MA system, where a multi-MA base station (BS) transmits to multiple single-FPA users. We formulate a sum-rate maximization problem by jointly optimizing the active/passive beamforming of the BS/IRS and the MA positions within a one-dimensional transmit region, which is challenging to be optimally solved. To drive essential insights, we first study a simplified case with a single user. Then, we analyze the performance gain of MAs over FPAs in the light-of-sight (LoS) BS-IRS channel and derive the conditions under which this gain becomes more or less significant. In addition, we propose an alternating optimization (AO) algorithm to solve the signal-to-noise ratio (SNR) maximization problem in the single-user case by combining the block coordinate descent (BCD) method and the graph-based method. For the general multi-user case, our performance analysis unveils that the performance gain of MAs over FPAs diminishes with typical transmit precoding strategies at the BS under certain conditions. We also propose a high-quality suboptimal solution to the sum-rate maximization problem by applying the AO algorithm that combines the weighted minimum mean square error (WMMSE) algorithm, manifold optimization method and discrete sampling method. Numerical results validate our theoretical analyses and demonstrate that the performance gain of MAs over FPAs may be reduced if the IRS passive beamforming is optimized.

The fusion of Synthetic Aperture Radar (SAR) and RGB imagery for land cover classification remains challenging due to modality heterogeneity and underutilized spectral complementarity. Existing methods often fail to decouple shared structural features from modality-complementary radiometric attributes, causing feature conflicts and information loss. To address this, we propose Phase-Amplitude Decoupling (PAD), a frequency-aware framework that separates phase (modality-shared) and amplitude (modality-complementary) components in the Fourier domain, thus reinforcing shared structures while preserving complementary characteristics to improve fusion quality. Unlike prior approaches that overlook the distinct physical properties encoded in frequency spectra, PAD is the first to introduce explicit amplitude-phase decoupling for multi-modal fusion. Specifically, PAD comprises two key components: 1) Phase Spectrum Correction (PSC), which aligns cross-modal phase features via convolution-guided scaling to enhance geometric consistency; and 2) Amplitude Spectrum Fusion (ASF), which dynamically integrates high-frequency and low-frequency patterns using frequency-adaptive multilayer perceptrons, leveraging SAR's morphological sensitivity and RGB's spectral richness. Extensive experiments on WHU-OPT-SAR and DDHR-SK datasets demonstrate state-of-the-art performance. Our work establishes a new paradigm for physics-aware multi-modal fusion in remote sensing. The code will be available at this https URL.

Pioneered by the foundational information theory by Claude Shannon and the visionary framework of machine intelligence by Alan Turing, the convergent evolution of information and communication technologies (IT/CT) has created an unbroken wave of connectivity and computation. This synergy has sparked a technological revolution, now reaching its peak with large artificial intelligence (AI) models that are reshaping industries and redefining human-machine collaboration. However, the realization of ubiquitous intelligence faces considerable challenges due to substantial resource consumption in large models and high communication bandwidth demands. To address these challenges, AI Flow has been introduced as a multidisciplinary framework that integrates cutting-edge IT and CT advancements, with a particular emphasis on the following three key points. First, device-edge-cloud framework serves as the foundation, which integrates end devices, edge servers, and cloud clusters to optimize scalability and efficiency for low-latency model inference. Second, we introduce the concept of familial models, which refers to a series of different-sized models with aligned hidden features, enabling effective collaboration and the flexibility to adapt to varying resource constraints and dynamic scenarios. Third, connectivity- and interaction-based intelligence emergence is a novel paradigm of AI Flow. By leveraging communication networks to enhance connectivity, the collaboration among AI models across heterogeneous nodes achieves emergent intelligence that surpasses the capability of any single model. The innovations of AI Flow provide enhanced intelligence, timely responsiveness, and ubiquitous accessibility to AI services, paving the way for the tighter fusion of AI techniques and communication systems.

Music mastering style transfer aims to model and apply the mastering characteristics of a reference track to a target track, simulating the professional mastering process. However, existing methods apply fixed processing based on a reference track, limiting users' ability to fine-tune the results to match their artistic intent. In this paper, we introduce the ITO-Master framework, a reference-based mastering style transfer system that integrates Inference-Time Optimization (ITO) to enable finer user control over the mastering process. By optimizing the reference embedding during inference, our approach allows users to refine the output dynamically, making micro-level adjustments to achieve more precise mastering results. We explore both black-box and white-box methods for modeling mastering processors and demonstrate that ITO improves mastering performance across different styles. Through objective evaluation, subjective listening tests, and qualitative analysis using text-based conditioning with CLAP embeddings, we validate that ITO enhances mastering style similarity while offering increased adaptability. Our framework provides an effective and user-controllable solution for mastering style transfer, allowing users to refine their results beyond the initial style transfer.

Turn-taking prediction models are essential components in spoken dialogue systems and conversational robots. Recent approaches leverage transformer-based architectures to predict speech activity continuously and in real-time. In this study, we propose a novel model that enables turn-taking prediction to be dynamically controlled via textual prompts. This approach allows intuitive and explicit control through instructions such as "faster" or "calmer" adapting dynamically to conversational partners and contexts. The proposed model builds upon a transformer-based voice activity projection (VAP) model, incorporating textual prompt embeddings into both channel-wise transformers and a cross-channel transformer. We evaluated the feasibility of our approach using over 950 hours of human-human spoken dialogue data. Since textual prompt data for the proposed approach was not available in existing datasets, we utilized a large language model (LLM) to generate synthetic prompt sentences. Experimental results demonstrated that the proposed model improved prediction accuracy and effectively varied turn-taking timing behaviors according to the textual prompts.

Region of Interest (ROI)-based image compression optimizes bit allocation by prioritizing ROI for higher-quality reconstruction. However, as the users (including human clients and downstream machine tasks) become more diverse, ROI-based image compression needs to be customizable to support various preferences. For example, different users may define distinct ROI or require different quality trade-offs between ROI and non-ROI. Existing ROI-based image compression schemes predefine the ROI, making it unchangeable, and lack effective mechanisms to balance reconstruction quality between ROI and non-ROI. This work proposes a paradigm for customizable ROI-based deep image compression. First, we develop a Text-controlled Mask Acquisition (TMA) module, which allows users to easily customize their ROI for compression by just inputting the corresponding semantic \emph{text}. It makes the encoder controlled by text. Second, we design a Customizable Value Assign (CVA) mechanism, which masks the non-ROI with a changeable extent decided by users instead of a constant one to manage the reconstruction quality trade-off between ROI and non-ROI. Finally, we present a Latent Mask Attention (LMA) module, where the latent spatial prior of the mask and the latent Rate-Distortion Optimization (RDO) prior of the image are extracted and fused in the latent space, and further used to optimize the latent representation of the source image. Experimental results demonstrate that our proposed customizable ROI-based deep image compression paradigm effectively addresses the needs of customization for ROI definition and mask acquisition as well as the reconstruction quality trade-off management between the ROI and non-ROI.

Federated learning (FL) has emerged as a promising paradigm for distributed machine learning, enabling collaborative training of a global model across multiple local devices without requiring them to share raw data. Despite its advancements, FL is limited by factors such as: (i) privacy risks arising from the unprotected transmission of local model updates to the fusion center (FC) and (ii) decreased learning utility caused by heterogeneity in model quantization resolution across participating devices. Prior work typically addresses only one of these challenges because maintaining learning utility under both privacy risks and quantization heterogeneity is a non-trivial task. In this paper, our aim is therefore to improve the learning utility of a privacy-preserving FL that allows clusters of devices with different quantization resolutions to participate in each FL round. Specifically, we introduce a novel stochastic quantizer (SQ) that is designed to simultaneously achieve differential privacy (DP) and minimum quantization error. Notably, the proposed SQ guarantees bounded distortion, unlike other DP approaches. To address quantization heterogeneity, we introduce a cluster size optimization technique combined with a linear fusion approach to enhance model aggregation accuracy. Numerical simulations validate the benefits of our approach in terms of privacy protection and learning utility compared to the conventional LaplaceSQ-FL algorithm.