Papers

In this work, we propose a data-driven divide and conquer strategy for the stability analysis of interconnected homogeneous nonlinear networks of degree one with unknown models and a fully unknown topology. The proposed scheme leverages joint dissipativity-type properties of subsystems described by storage functions, while providing a stability certificate over unknown interconnected networks. In our data-driven framework, we begin by formulating the required conditions for constructing storage functions as a robust convex program (RCP). Given that unknown models of subsystems are integrated into one of the constraints of the RCP, we collect data from trajectories of each unknown subsystem and provide a scenario convex program (SCP) that aligns with the original RCP. We solve the SCP as a linear program and construct a storage function for each subsystem with unknown dynamics. Under some newly developed data-driven compositionality conditions, we then construct a Lyapunov function for the fully unknown interconnected network utilizing storage functions derived from data of individual subsystems. We show that our data-driven {divide and conquer strategy} provides correctness guarantees (as opposed to probabilistic confidence) while significantly mitigating the sample complexity problem existing in data-driven approaches. To illustrate the effectiveness of our proposed results, we apply our approaches to three different case studies involving interconnected homogeneous (nonlinear) networks with unknown models. We collect data from trajectories of unknown subsystems and verify the global asymptotic stability (GAS) of the interconnected system with a guaranteed confidence.

Reconfigurable intelligent surfaces (RISs) have been extensively applied in integrated sensing and communication (ISAC) systems due to the capability of enhancing physical layer security (PLS). However, conventional static RIS architectures lack the flexibility required for adaptive beam control in multi-user and multifunctional scenarios. To address this issue without introducing additional hardware complexity and power consumption, in this paper, we exploit a movable RIS (MRIS) architecture, which consists of a large fixed sub-surface and a smaller movable sub-surface that slides on the fixed sub-surface to achieve dynamic beam reconfiguration with static phase shifts. This paper investigates an MRIS-assisted ISAC system under imperfect sensing estimation, where dedicated radar signals serve as artificial noise to enhance secure transmission against potential eavesdroppers (Eves). The transmit beamforming vectors, MRIS phase shifts, and relative positions of the two sub-surfaces are jointly optimized to maximize the minimum secrecy rate, ensuring robust secrecy performance for the weakest user under the uncertainty of the Eves' channels. To handle the non-convexity, a convex bound is derived for the Eve channel uncertainty, and the S-procedure is employed to reformulate semi-infinite constraints as linear matrix inequalities. An efficient alternating optimization and penalty dual decomposition-based algorithm is developed. Simulation results demonstrate that the proposed MRIS architecture substantially improves secrecy performance, especially when only a small number of elements are allocated to the movable sub-surface.

Recent advances in learned image codecs have extended from human perception toward machine perception However, progressive image compression with fine granular scalability (FGS)-which enables decoding a single bitstream at multiple quality levels-remains unexplored for machine-oriented codecs. In this work, we propose PICM-Net, a progressive learned image compression codec for machine perception built on trit-plane coding. Starting from a human-oriented codec, we adapt it for machine perception via a spatial-frequency modulation adapter (SFMA), a hyper-synthesis low-rank adapter (HSLoRA), and progressive decoding-aware training, and analyze how task-driven adaptation alters symbol prioritization for progressive transmission. To further support real-world deployment, we introduce an adaptive decoding controller that dynamically determines the necessary decoding level at inference time, requesting additional bits only when the current suitability is insufficient for the desired confidence level. Extensive experiments demonstrate that PICM-Net achieves efficient and adaptive progressive transmission while maintaining strong downstream classification performance. Our code is available at https://github.com/kjungwoo03/PICM-Net.

Controllable Markov chains describe the dynamics of sequential decision making tasks and are the central component in optimal control and reinforcement learning. In this work, we give the general form of an optimal policy for learning controllable dynamics in an unknown environment by exploring over a limited time horizon. This policy is simple to implement and efficient to compute, and allows an agent to “learn by exploring" as it maximizes its information gain in a greedy fashion by selecting controls from a constraint set that changes over time during exploration. We give a simple parameterization for the set of controls, and present an algorithm for finding an optimal policy. The reason for this policy is due to the existence of certain types of states that restrict control of the dynamics; such as transient states, absorbing states, and non-backtracking states. We show why the occurrence of these states makes a non-stationary policy essential for achieving optimal exploration. Six interesting examples of controllable dynamics are treated in detail. Policy optimality is demonstrated using counting arguments, comparing with suboptimal policies, and by making use of a sequential improvement property from dynamic programming.

An augmented system model provides an effective way to model non-Markovian quantum systems, which is useful in filtering and control for this class of systems. However, since a large number of ancillary quantum oscillators representing internal modes of a non-Markovian environment directly interact with the principal system in these models, the dimension of the augmented system may be very large causing significant computational burden in designing filters and controllers. In this context, this paper proposes an $\mathscr{H}_2$ model reduction method for the augmented model of linear non-Markovian quantum systems. We first establish necessary and sufficient conditions for the physical realizability of the augmented model of linear non-Markovian quantum systems, which are more stringent than those for Markovian quantum systems. However, these physical realizability conditions of augmented system model pose non-convex constrains in the optimization problem of model reduction, which makes the problem different from the corresponding classical model reduction problem. To solve the problem, we derive necessary conditions for determining the input matrix in the reduced model, with which a theorem for designing the system matrix of the ancillary system in the reduced system is proved. Building on this, we convert the nonlinear equality constraints into inequality constraints so that a semidefinite programming algorithm can be developed to solve the optimization problem for model reduction. A numerical example of a two-mode linear quantum system driven by three internal modes of a non-Markovian environment validates the effectiveness of our method.

We propose and demonstrate an elastic digital-analog radio-over-fiber (RoF) modulation and demodulation architecture, seamlessly bridging A-RoF and D-RoF solutions, achieving quasilinear SNR scaling with respect to 1/{\eta}, and evidenced by R^2=0.9908.

The Heyland circle diagram is a classical graphical tool for representing the steady-state behavior of induction machines using no-load and blocked-rotor test data. While widely used in alternating-current machinery texts, the diagram is typically presented as a hand-constructed aid and lacks a standardized computational formulation. This paper presents HeylandCircle, a computational framework that reconstructs the classical Heyland circle diagram directly from standard test parameters. The framework formalizes the traditional geometric construction as a deterministic, reproducible sequence of geometric operations, establishing a clear mapping between measured data, fixed geometric objects, and steady-state operating points. Quantities such as power factor, slip, output power, torque, and efficiency are obtained through explicit geometric relationships on the constructed diagram. Validation using a representative textbook example demonstrates close agreement with classical results. The framework provides a computational realization of the traditional Heyland diagram suitable for instruction, analysis, and systematic extension.

Large language models (LLMs) are emerging as key enablers of automation in domains such as telecommunications, assisting with tasks including troubleshooting, standards interpretation, and network optimization. However, their deployment in practice must balance inference cost, latency, and reliability. In this work, we study an edge-cloud-expert cascaded LLM-based knowledge system that supports decision-making through a question-and-answer pipeline. In it, an efficient edge model handles routine queries, a more capable cloud model addresses complex cases, and human experts are involved only when necessary. We define a misalignment-cost constrained optimization problem, aiming to minimize average processing cost, while guaranteeing alignment of automated answers with expert judgments. We propose a statistically rigorous threshold selection method based on multiple hypothesis testing (MHT) for a query processing mechanism based on knowledge and confidence tests. The approach provides finite-sample guarantees on misalignment risk. Experiments on the TeleQnA dataset – a telecom-specific benchmark – demonstrate that the proposed method achieves superior cost-efficiency compared to conventional cascaded baselines, while ensuring reliability at prescribed confidence levels.

We propose and demonstrate a power-fading-aware noise-shaping technique for C-band IMDD system with low resolution DAC, which shapes and concentrates quantization noise within the fading-induced notch areas, yielding 94% improvement in data-rate over traditional counterpart.

The exponential proliferation of mobile devices and data-intensive applications in future wireless networks imposes substantial computational burdens on resource-constrained devices, thereby fostering the emergence of over-the-air computation (AirComp) as a transformative paradigm for edge intelligence.} To enhance the efficiency and scalability of AirComp systems, this paper proposes a comprehensive dual-approach framework that systematically transitions from traditional mathematical optimization to deep reinforcement learning (DRL) for resource allocation under execution uncertainty. Specifically, we establish a rigorous system model capturing execution uncertainty via Gamma-distributed computational workloads, resulting in challenging nonlinear optimization problems involving complex Gamma functions. For single-user scenarios, we design advanced block coordinate descent (BCD) and majorization-maximization (MM) algorithms, which yield semi-closed-form solutions with provable performance guarantees. However, conventional optimization approaches become computationally intractable in dynamic multi-user environments due to inter-user interference and resource contention. To this end, we introduce a Deep Q-Network (DQN)-based DRL framework capable of adaptively learning optimal policies through environment interaction. Our dual methodology effectively bridges analytical tractability with adaptive intelligence, leveraging optimization for foundational insight and learning for real-time adaptability. Extensive numerical results corroborate the performance gains achieved via increased edge server density and validate the superiority of our optimization-to-learning paradigm in next-generation AirComp systems.

As integrated sensing and communication (ISAC) systems are deployed in next-generation wireless networks, a new security vulnerability emerges, particularly in terms of sensing privacy. Unauthorized sensing eavesdroppers (Eve) can potentially exploit the ISAC signal for their own independent passive sensing. However, solutions for sensing-secure ISAC remain largely unexplored to date. This work addresses sensing-security for OFDM- and OTFS-based ISAC waveforms from a target-detection perspective, aiming to prevent Eves from exploiting the ISAC signal for unauthorized passive sensing. We develop ISAC system models for the base station (BS), communication user equipment, and the sensing Eve, and define a Kullback-Leibler-divergence-based detection metric that accounts for mainlobe, sidelobe, and noise components in the ambiguity function and the resulting range-Doppler maps of the legitimate BS's and Eve's sensing. Building on this analysis, we formulate a sensing-secure ISAC signaling design problem that tunes a perturbation matrix to jointly control signal amplitude and phase in the time-frequency domain and solve it via simulated annealing. Simulation results show that the proposed scheme substantially degrades Eve's detection probability – from 79.4% to 37.4% for OTFS and from 94.3% to 33.0% for OFDM – while incurring only a small loss in BS sensing performance. In addition, it allows controllable trade-offs across sensing-security and communication performance.

This study introduces a novel data-driven framework and the first-ever county-scale application of Spatio-Temporal Graph Neural Networks (STGNN) to forecast composite sustainability indices from herd-level operational records. The methodology employs a novel, end-to-end pipeline utilizing a Variational Autoencoder (VAE) to augment Irish Cattle Breeding Federation (ICBF) datasets, preserving joint distributions while mitigating sparsity. A first-ever pillar-based scoring formulation is derived via Principal Component Analysis, identifying Reproductive Efficiency, Genetic Management, Herd Health, and Herd Management, to construct weighted composite indices. These indices are modelled using a novel STGNN architecture that explicitly encodes geographic dependencies and non-linear temporal dynamics to generate multi-year forecasts for 2026-2030.

This paper investigates a rate-splitting multiple access (RSMA) for uplink pinching antenna system (PASS). Our objective is to maximize the uplink sum rate by jointly optimizing a continuous antenna positioning and user's transmission power. The formulated problem is highly non-convex and difficult to solve directly; to address this challenge, we propose an alternating optimization (AO) framework which decomposes the original problem into two tractable sub-problems, namely (i) power allocation optimization sub-problem and (ii) antenna position optimization sub-problem. Both sub-problems are solved using successive convex approximation (SCA)-based algorithm and solved alternatively until convergence. The RSMA access for PASS is compared with conventional non-orthogonal multiple access (NOMA) and space-division multiple access (SDMA) techniques. The performance of discrete antenna activation with PASS strategy is also examined. Simulation results demonstrate that our proposed framework significantly enhances the achievable sum rate compared to other multiple access methods.

The principle of learning from errors is pedagogically powerful but often impractical in industrial settings due to risks to safety and equipment. This paper presents an integrated training approach specifically designed for tufting machine operators. It uses hybrid digital twins, augmented reality (AR), and Petri Net-based modelling to apply the learning from errors principle effectively. Operator actions and errors are simulated via experimentable digital twins (EDTs), and the consequences of errors are visualized in AR, enabling safe, experiential learning. A Petri Net model formally represents the process, including typical faults and recovery paths, and is implemented in VEROSIM using SOML++. This hybrid framework provides a scalable foundation for AR-guided training systems that reduce risk and accelerate skill acquisition.

Electroencephalography (EEG)-based brain-computer interfaces (BCIs) are strongly affected by non-stationary neural signals that vary across sessions and individuals, limiting the generalization of subject-agnostic models and motivating adaptive and personalized learning on resource-constrained platforms. Programmable memristive hardware offers a promising substrate for such post-deployment adaptation; however, practical realization is challenged by limited weight resolution, device variability, nonlinear programming dynamics, and finite device endurance. In this work, we show that spiking neural networks (SNNs) can be deployed on ferroelectric memristive synaptic devices for adaptive EEG-based motor imagery decoding under realistic device constraints, achieving classification performance comparable to software-based SNNs. We fabricate, characterize, and model the weight update in ferroelectric synapses. We then evaluate the deployment of convolutional-recurrent SNN architecture using two strategies. First, we adapt to SNNs a mixed precision strategy in which gradient-based updates are accumulated digitally and converted into discrete programming events only when a threshold is exceeded. Additionally, the weight update is device-aware and accounts for the nonlinear, state-dependent programming dynamics. During learning and adaptation, this scheme mitigates possible endurance and energy constraints. Second, we evaluate the transfer of software-trained weights followed by low-overhead on-device re-tuning. We show that, subject-specific transfer learning achieved by retraining only the final network layers improves classification accuracy. These results demonstrate that programmable ferroelectric hardware can support robust, low-overhead adaptation in spiking neural networks, opening a practical path toward personalized neuromorphic processing of neural signals.

Capsule endoscopy has enabled minimally invasive gastrointestinal imaging, but its clinical utility is limited by the inherently low resolution of captured images due to hardware, power, and transmission constraints. This limitation hampers the identification of fine-grained mucosal textures and subtle pathological features essential for early diagnosis. This work investigates a diffusion-based super-resolution framework to enhance capsule endoscopy images in a data-driven and anatomically consistent manner. We adopt the SR3 (Super-Resolution via Repeated Refinement) framework built upon Denoising Diffusion Probabilistic Models (DDPMs) to learn a probabilistic mapping from low-resolution to high-resolution images. Unlike GAN-based approaches that often suffer from training instability and hallucination artifacts, diffusion models provide stable likelihood-based training and improved structural fidelity. The HyperKvasir dataset, a large-scale publicly available gastrointestinal endoscopy dataset, is used for training and evaluation. Quantitative results demonstrate that the proposed method significantly outperforms bicubic interpolation and GAN-based super-resolution methods such as ESRGAN, achieving PSNR of 27.5 dB and SSIM of 0.65 for a baseline model, and improving to 29.3 dB and 0.71 with architectural enhancements including attention mechanisms. Qualitative results show improved preservation of anatomical boundaries, vascular patterns, and lesion structures. These findings indicate that diffusion-based super-resolution is a promising approach for enhancing non-invasive medical imaging, particularly in capsule endoscopy where image resolution is fundamentally constrained.

Drone applications continue to expand across various domains, with flocking offering enhanced cooperative capabilities but introducing significant challenges during initial formation. Existing flocking algorithms often struggle with efficiency and scalability, particularly when potential collisions force drones into suboptimal trajectories. This paper presents a time-efficient prioritised scheduling algorithm that improves the initial formation process of drone flocks. The method assigns each drone a priority based on its number of potential collisions and its likelihood of reaching its target position without permanently obstructing other drones. Using this hierarchy, each drone computes an appropriate delay to ensure a collision-free path. Simulation results show that the proposed algorithm successfully generates collision-free trajectories for flocks of up to 5000 drones and outperforms the coupling-degree-based heuristic prioritised planning method (CDH-PP) in both performance and computational efficiency.

We introduce a new class of attitude control laws for rotational systems; the proposed framework generalizes the use of the Euler \mbox{axis–angle} representation beyond quaternion-based formulations. Using basic Lyapunov stability theory and the notion of extended class $\mathcal{K}$ function, we developed a method for determining and enforcing the global asymptotic stability of the single fixed point of the resulting \mbox{closed-loop} (CL) scheme. In contrast with traditional \mbox{quaternion-based} methods, the introduced generalized \mbox{axis–angle} approach enables greater flexibility in the design of the control law, which is of great utility when employed in combination with a switching scheme whose transition state depends on the angular velocity of the controlled rotational system. Through simulation and \mbox{real-time} experimental results, we demonstrate the effectiveness of the developed formulation. According to the recorded data, in the execution of \mbox{high-speed} \mbox{tumble-recovery} maneuvers, the new method consistently achieves shorter stabilization times and requires lower control effort relative to those corresponding to the \mbox{quaternion-based} and \mbox{geometric-control} methods used as benchmarks.

This article presents an analysis of the practical challenges and implementation perspectives of point-to-point continuous-variable quantum key distribution (CV-QKD) systems over optical fiber. The study addresses the physical layer, including the design of transmitters, quantum channels, and receivers, with emphasis on impairments such as attenuation, chromatic dispersion, polarization fluctuations, and coexistence with classical channels. We further examine the role of digital signal processing (DSP) as the bridge between quantum state transmission and classical post-processing, highlighting its impact on excess noise mitigation, covariance matrix estimation, and reconciliation efficiency. The post-processing pipeline is detailed with a focus on parameter estimation in the finite-size regime, information reconciliation using LDPC-based codes optimized for low-SNR conditions, and privacy amplification employing large-block universal hashing. From a hardware perspective, we discuss modular digital architectures that integrate dedicated accelerators with programmable processors, supported by a reference software framework (CV-QKD-ModSim) for algorithm validation and hardware co-design. Finally, we outline perspectives for the deployment of CV-QKD in Brazil, starting from metropolitan testbeds and extending toward hybrid fiber/FSO and space-based infrastructures. The work establishes the foundations for the first point-to-point CV-QKD system in Brazil, while providing a roadmap for scalable and interoperable quantum communication networks.

The security of Automated Vehicles (AVs) is an important emerging area of research in traffic safety. Methods have been published and evaluated in experimental vehicles to secure safe AV control in the presence of attacks, but human motion comfort is rarely investigated in such studies. In this paper, we present an innovative optimal-coupling-observer-based framework that rejects the impact of bounded sensor attacks in a network of connected and automated vehicles from safety and comfort point of view. We demonstrate its performance in car following with cooperative adaptive cruise control for platoons with redundant distance and velocity sensors. The error dynamics are formulated as a Linear Time Variant (LTV) system, resulting in complex stability conditions that are investigated using a Linear Matrix Inequality (LMI) approach guaranteeing global asymptotic stability. We prove the capability of the framework to secure occupants' safety and comfort in the presence of bounded attacks. In the onset of attack, the framework rapidly detects attacked sensors and switches to the most reliable observer eliminating attacked sensors, even with modest attack magnitudes. Without our proposed method, severe (but bounded) attacks result in collisions and major discomfort. With our method, attacks had negligible effects on motion comfort evaluated using ISO-2631 Ride Comfort and Motion Sickness indexes. The results pave the path to bring comfort to the forefront of AVs security.

The emergence of commercial satellite communications networks, such as Starlink and OneWeb, has significantly transformed the communications landscape over the last years. As a complement to terrestrial cellular networks, non-terrestrial systems enable coverage extension and reliability enhancement beyond the limits of conventional infrastructure. Currently, the high reliance on terrestrial networks exposes communications to vulnerabilities in the event of terrestrial infrastructure failures, e.g., due to natural disasters. Therefore, this work proposes the joint evaluation of Key Performance Indicators (KPIs) for two non-terrestrial satellite networks (Starlink and OneWeb) and two terrestrial cellular networks to assess the current performance of these technologies across three different environments: (i) urban, (ii) suburban, and (iii) forest scenarios. Additionally, multi-connectivity techniques are explored to determine the benefits in connectivity when two technologies are used simultaneously. For instance, the outage probability of Starlink and OneWeb in urban areas is reduced from approximately 12-21% to 2% when both solutions are employed together. Finally, the joint analysis of KPIs in both terrestrial and non-terrestrial networks demonstrates that their integration enhances coverage, improves performance, and increases reliability, highlighting the benefits of combining satellite and terrestrial systems in the analyzed environments.

This paper introduces MauBERT, a multilingual extension of HuBERT that leverages articulatory features for robust cross-lingual phonetic representation learning. We continue HuBERT pre-training with supervision based on a phonetic-to-articulatory feature mapping in 55 languages. Our models learn from multilingual data to predict articulatory features or phones, resulting in language-independent representations that capture multilingual phonetic properties. Through comprehensive ABX discriminability testing, we show MauBERT models produce more context-invariant representations than state-of-the-art multilingual self-supervised learning models. Additionally, the models effectively adapt to unseen languages and casual speech with minimal self-supervised fine-tuning (10 hours of speech). This establishes an effective approach for instilling linguistic inductive biases in self-supervised speech models.

Dynamic PET enables the quantitative estimation of physiology-related parameters and is widely utilized in research and increasingly adopted in clinical settings. Parametric imaging in dynamic PET requires kinetic modeling to estimate voxel-wise physiological parameters based on specific kinetic models. However, parametric images estimated through kinetic model fitting often suffer from low image quality due to the inherently ill-posed nature of the fitting process and the limited counts resulting from non-continuous data acquisition across multiple bed positions in whole-body PET. In this work, we proposed a diffusion model-based kinetic modeling framework for parametric image estimation, using the Patlak model as an example. The score function of the diffusion model was pre-trained on static total-body PET images and served as a prior for both Patlak slope and intercept images by leveraging their patch-wise similarity. During inference, the kinetic model was incorporated as a data-consistency constraint to guide the parametric image estimation. The proposed framework was evaluated on total-body dynamic PET datasets with different dose levels, demonstrating the feasibility and promising performance of the proposed framework in improving parametric image quality.

Attitude control is essential for many satellite missions. Classical controllers, however, are time-consuming to design and sensitive to model uncertainties and variations in operational boundary conditions. Deep Reinforcement Learning (DRL) offers a promising alternative by learning adaptive control strategies through autonomous interaction with a simulation environment. Overcoming the Sim2Real gap, which involves deploying an agent trained in simulation onto the real physical satellite, remains a significant challenge. In this work, we present the first successful in-orbit demonstration of an AI-based attitude controller for inertial pointing maneuvers. The controller was trained entirely in simulation and deployed to the InnoCube 3U nanosatellite, which was developed by the Julius-Maximilians-Universität Würzburg in cooperation with the Technische Universität Berlin, and launched in January 2025. We present the AI agent design, the methodology of the training procedure, the discrepancies between the simulation and the observed behavior of the real satellite, and a comparison of the AI-based attitude controller with the classical PD controller of InnoCube. Steady-state metrics confirm the robust performance of the AI-based controller during repeated in-orbit maneuvers.

This work revisits the hyperspectral super-resolution (HSR) problem, i.e., fusing a pair of spatially co-registered hyperspectral (HSI) and multispectral (MSI) images to recover a super-resolution image (SRI) that enhances the spatial resolution of the HSI. Coupled tensor decomposition (CTD)-based methods have gained traction in this domain, offering recoverability guarantees under various assumptions. Existing models such as canonical polyadic decomposition (CPD) and Tucker decomposition provide strong expressive power but lack physical interpretability. The block-term decomposition model with rank-$(L_r, L_r, 1)$ terms (the LL1 model) yields interpretable factors under the linear mixture model (LMM) of spectral images, but LMM assumptions are often violated in practice – primarily due to nonlinear effects such as endmember variability (EV). To address this, we propose modeling spectral images using a more flexible block-term tensor decomposition with rank-$(L_r, M_r, N_r)$ terms (the LMN model). This modeling choice retains interpretability, subsumes CPD, Tucker, and LL1 as special cases, and robustly accounts for non-ideal effects such as EV, offering a balanced tradeoff between expressiveness and interpretability for HSR. Importantly, under the LMN model for HSI and MSI, recoverability of the SRI can still be established under proper conditions – providing strong theoretical support. Extensive experiments on synthetic and real datasets further validate the effectiveness and robustness of the proposed method compared with existing CTD-based approaches.

We study optimal charging of large electric vehicle populations that are exposed to a single real-time imbalance price. The problem is naturally cast as a multistage stochastic linear programme (MSLP), which can be solved by algorithms such as Stochastic Dual Dynamic Programming. However, these methods scale poorly with the number of devices and stages. This paper presents a novel approach to overcome this curse of dimensionality. Building prior work that characterises the aggregate flexibility sets of populations of EVs as a permutahdron, we reformulate the original problem in terms of aggregated quantities. The geometric structure of permutahedra lets us (i) construct an optimal disaggregation policy, (ii) derive an exact, lower-dimensional MSLP, and (iii) characterise the expected recourse function as piecewise affine with a finite, explicit partition. In particular, we provide closed-form expressions for the slopes and intercepts of each affine region via truncated expectations of future prices, yielding an exact form for the recourse function and first-stage policy. Comprehensive numerical studies validate our claims and demonstrate the practical utility of this work.

We examine unsourced random access in a fully asynchronous setup, where active users transmit their data without restriction on the start time over a fading channel. In the proposed scheme, the transmitted signal consists of a pilot sequence and a polar codeword, with the polar codeword distributed across the data part of the packet in an on-off pattern. The receiver uses a double sliding-window decoder, where the inner window employs iterative decoding with joint timing and pilot detection, channel estimation, single-user decoding, and successive interference cancellation to recover the message bits, while the outer window enhances interference cancellation. The numerical results indicate that the proposed scheme exhibits only a slight performance loss compared to the synchronous benchmark while being more applicable in practice.

Differentiable optimal control, particularly differentiable nonlinear model predictive control (NMPC), provides a powerful framework that enjoys the complementary benefits of machine learning and control theory. A key enabler of differentiable optimal control is the computation of derivatives of the optimal trajectory with respect to problem parameters, i.e., trajectory derivatives. Previous works compute trajectory derivatives by solving a differential Karush-Kuhn-Tucker (KKT) system, and achieve this efficiently by constructing an equivalent auxiliary system. However, we find that directly exploiting the matrix structures in the differential KKT system yields significant computation speed improvements. Motivated by this insight, we propose FastDOC, which applies a Gauss-Newton approximation of Hessian and takes advantage of the resulting block-sparsity and positive semidefinite properties of the matrices involved. These structural properties enable us to accelerate the computationally expensive matrix factorization steps, resulting in a factor-of-two speedup in theoretical computational complexity, and in a synthetic benchmark FastDOC achieves up to a 180% time reduction compared to the baseline method. Finally, we validate the method on an imitation learning task for human-like autonomous driving, where the results demonstrate the effectiveness of the proposed FastDOC in practical applications.

Diffusion-based generative models have greatly impacted the speech processing field in recent years, exhibiting high speech naturalness and spawning a new research direction. Their application in real-time communication is, however, still lagging behind due to their computation-heavy nature involving multiple calls of large DNNs. Here, we present Stream$.$FM, a frame-causal flow-based generative model with an algorithmic latency of 32 milliseconds (ms) and a total latency of 48 ms, paving the way for generative speech processing in real-time communication. We propose a buffered streaming inference scheme and an optimized DNN architecture, show how learned few-step numerical solvers can boost output quality at a fixed compute budget, explore model weight compression to find favorable points along a compute/quality tradeoff, and contribute a model variant with 24 ms total latency for the speech enhancement task. Our work looks beyond theoretical latencies, showing that high-quality streaming generative speech processing can be realized on consumer GPUs available today. Stream$.$FM can solve a variety of speech processing tasks in a streaming fashion: speech enhancement, dereverberation, codec post-filtering, bandwidth extension, STFT phase retrieval, and Mel vocoding. As we verify through comprehensive evaluations and a MUSHRA listening test, Stream$.$FM establishes a state-of-the-art for generative streaming speech restoration, exhibits only a reasonable reduction in quality compared to a non-streaming variant, and outperforms our recent work (Diffusion Buffer) on generative streaming speech enhancement while operating at a lower latency.

Cockcroft-Walton (CW) voltage multipliers suffer from output ripple that classical analytical models underestimate due to neglected non-idealities like diode drops and capacitor ESR, particularly in high-stage, low-frequency and heavy-load regimes. This paper proposes a hybrid framework that generates a comprehensive 324-case MATLAB/Simulink dataset varying stages (2-8), input voltage (5-25 kV), capacitance (1-10 {\mu}F), frequency (50-500 Hz) and load (6-60 M{\Omega}), then trains a Random Forest model to predict residuals between simulated and theoretical peak-to-peak ripple. The approach achieves 70.6% RMSE reduction (131 V vs. 448 V) globally and 66.7% in critical regimes, with near-zero bias, enabling physically interpretable design optimization while outperforming pure ML in extrapolation reliability.