Modern retrieval agents expose many configuration choices -- LLM, retriever, number of documents, number of hops, and synthesis strategy -- each shaping both answer quality and serving cost. Today, these pipelines are typically hand-tuned once per workload, leaving substantial per-query optimization untapped. We formulate the problem: given a natural-language query and either an accuracy or a budget target, select from a predefined pipeline catalog the configuration that minimizes cost or maximizes accuracy at inference time. We propose **BRANE**, which uses an LLM to convert each query into workload-specific characteristics, then trains a lightweight per-configuration predictor that estimates whether the pipeline will answer the query correctly. At inference time, **BRANE** selects the configuration that maximizes predicted correctness penalized by cost, exposing a tunable cost-quality tradeoff without retraining. Across MuSiQue, BrowseComp-Plus, and FinanceBench, **BRANE** consistently pushes the cost-quality Pareto frontier, matches the best fixed configuration's accuracy at up to 89% lower cost, and outperforms LLM-routing, rule-based, and fine-tuned Qwen3-4B baselines. These results show that per-query configuration of the full retrieval pipeline is a practical alternative to static workload-level tuning.

Reactive control is often considered insufficient for multi-objective tasks because conflicting objectives give rise to local minima. We argue this limitation is not inherent but arises from static encodings that fail to reflect how objectives currently interact. We exploit the interaction structure encoded in a graph-based world model by extending it with nullspace projections: conflicts are resolved where they arise by projecting lower-priority gradients into the nullspace of higher-priority ones, with priorities determined continuously from the current state. We demonstrate this in two domains where conflicts between objectives are central: navigation around non-convex obstacles, where static potential fields fundamentally fail, and planar pushing of non-convex objects, where our method achieves $100\%$ success across one-hundred configurations versus $0\%$ for the steepest-descent baseline and ${\sim}55\%$ for diffusion policy, without demonstrations or retraining. The same formulation transfers directly to a real robot with additional perceptual and kinematic constraints, accommodating them through the same mechanism.

Modern intrusion detection systems generate thousands of alerts daily, but alert fatigue severely limits security operations effectiveness due to too many false positives or low-impact events. We address this by proposing a principled framework for alert prioritization based on subnormal Gaussian fuzzy numbers, explicitly modeling three sources of uncertainty: threat severity, detection confidence, and organizational risk attitude. Each alert is represented as a fuzzy number with the core indicating severity, spread indicating uncertainty, and height reflecting detection reliability. We apply ranking indices to prioritize alerts, allowing organizations to tune security posture through a risk-attitude parameter. Experimental validation on CIC-IDS2017 and NSL-KDD demonstrates greater robustness than baselines under detector degradation (0.9963 vs 0.8215 NDCGrel@100), with distinct differentiation in mid-confidence alerts and near-parity with baselines under robust detectors. The framework is theoretically grounded, computationally efficient, provides interpretable reasoning, and remains robust across detector families and miscalibration scenarios.

Wireless digital twins require repeated synchronization between a time-evolving physical scene and its digital counterpart under limited and time-varying communication resources. For perception-centric twins, pixel-domain transmission or uniformly protected bitstreams can be mismatched to the semantic state consumed by twin-side applications. This paper proposes TWIST, a closed-loop token synchronization framework for application-aware wireless digital twins. TWIST represents each physical observation as a token and synchronizes this state over a wireless link, rather than optimizing visual reconstruction. Token positions are grouped by task relevance and protected through mode-conditioned unequal error protection under low-, medium-, and high-synchronization modes. At the twin side, decoding confidence converts unreliable hard token decisions into erasures, which are restored by a completion model before updating the semantic twin state. The recovered state supports traffic-state inference and generates compact feedback statistics, including channel quality, receiver uncertainty, semantic drift, and application priority, for subsequent mode adaptation. Experiments on a dynamic road-scene digital-twin scenario show that TWIST improves traffic-state inference and semantic twin-state synchronization compared with fixed-mode and channel-only adaptation strategies, while reducing the average synchronization cost relative to always-high transmission.

Quantum state tomography (QST) is a fundamental task in quantum information science that aims to reconstruct unknown quantum states from measurement data. However, the exponential growth of Hilbert-space dimension with system size makes full tomography of general quantum states statistically and computationally prohibitive. This challenge has motivated extensive research on structured quantum state tomography, where prior structure, such as low-rankness, tensor-network representations, shallow quantum circuits, and neural quantum states, can substantially reduce the effective degrees of freedom and enable scalable recovery. In this review, we provide a unified perspective on QST for structured quantum states through three closely related themes: compact state representations, measurement design, and computational algorithms. After reviewing common models for structured quantum states, we survey existing work on geometric preservation properties of measurement frameworks, ranging from informationally complete POVMs to randomized measurements, and their implications for sample complexity. On the algorithmic side, we review optimization methods for reconstructing structured quantum states from empirical measurements. By connecting QST with broader principles from compressive sensing, matrix sensing, and structured inverse problems, this survey highlights common theoretical foundations underlying sample complexity, measurement efficiency, and scalable recovery.

This study examines the relationship between speech representations and the hierarchical structure of cognitive assessment in mild cognitive impairment. Utilizing 5,754 German neuropsychological assessment recordings, we evaluate six cognitive tasks across three score levels: task, domain, and global levels. We compare hand-crafted acoustic features with self-supervised learning (SSL) embeddings. Results show that although SSL representations generally outperform hand-crafted features at lower levels, this trend reverses for MCI classification. Furthermore, task-specific constraints influence performance: tasks with greater response freedom exhibit performance dilution as hierarchical levels increase, suggesting ``specialist'' representations, whereas the performance of highly structured tasks increases toward higher levels, suggesting ``generalist'' representations. These findings show links between task constraints and assessment hierarchy in automated clinical speech analysis.

Unloading containers in the courier, express and parcel industry is a physically demanding and labor-intensive work. Automatizing this process is an important step towards increasing the efficiency of parcel-handling systems. This work investigates the potential of reinforcement learning to learn a policy for item selection in container unloading scenarios. For that, a simulation environment is created and a masked deep Q-learning with a specially designed neural network architecture is implemented. The results indicate that the agent can learn to select items with an average success rate of 60 %, which is significantly better than a random policy at a random chance of 20 %. The findings suggest that RL could be a promising approach for automatizing item unloading tasks in the future.

Electron tomography (ET) plays an important role in the three-dimensional (3D) characterization of nanomaterials. However, under limited-angle and sparse-view conditions, conventional algorithms produce degraded reconstructions, which compromise the quality and interpretability of resulting 3D data. In this paper, we present deep image prior (DIP), an unsupervised deep learning (DL) approach, for highly degraded tomography acquisitions and demonstrate, using simulated data, that its performance is comparable to that of supervised approaches requiring training datasets, even for tilt ranges as limited as 60° and tilt increments of 10°. We then apply it to experimental data and show that it enables reliable 3D quantification under both sparse-view and limited-angle conditions, highlighting its potential for a wide range of materials and acquisition modalities.

The electrification of long-haul freight transport introduces several new challenges, such as the limited capacity and congestion at en-route charging infrastructure. To reduce waiting times during peak periods, this paper proposes a framework for coordinated charging scheduling. The approach employs a mixed-integer formulation to optimize charging-related costs across charging, operation, battery degradation, and congestion delay, considering a range of scenarios. The results demonstrate that coordinated scheduling yields substantial cost savings up to 36% compared to uncoordinated scheduling, particularly by reducing battery degradation and delay costs.

Large audio language models (LALMs) process both speech and environmental acoustic cues, yet struggle to retain non-speech information across multi-turn interactions. The performance gap between semantic (speech) and acoustic (non-speech) understanding remains poorly understood, and the underlying mechanisms of representation and retrieval are still unclear. This work introduces EnvMem, a controlled multi-turn benchmark designed to study this gap and identify the root causes of failures at the representation (i.e., latent embeddings) and retrieval levels (i.e., attention allocation). We further conduct post-hoc interventions to probe representational structure and attention dynamics. Our results reveal representational trajectory drift as the key failure mode, while showing that attention allocation plays a limited role in explaining the observed degradation. Overall, we provide a systematic framework for analyzing and improving non-linguistic memory in long-context LALMs, shedding light on future data and training design for robust acoustic memory modeling.

This paper studies how to balance onboard and ground computation under intermittent LEO connectivity for optimized inference freshness. As connectivity varies in time, the system switches among the actions of onboard computation, cached semantic transmission, raw-data offloading, and waiting. We define Age of Inference (AoInf) as the performance metric, where the age resets only upon successful task-valid updates. We formulate long-run average AoInf minimization as a finite-state average-cost semi-Markov decision process whose state captures the ground AoInf, orbital contact phase, cache occupancy, and cache age. We then transform the SMDP into an equivalent average-cost MDP and compute the solution via normalized relative value iteration (RVI). Numerical results indicate that the resulting hybrid policy reduces average AoInf relative to onboard-only and offload-only baselines, while requiring less computational resources on the satellite than the former, and fewer communication resources than the latter.

Modern energy systems in vehicles and built infrastructure are governed by high-dimensional dynamics spanning multiple physical domains (e.g., electrical, thermal, mechanical) and timescales. This tutorial paper presents a graph-based modeling approach created to facilitate the modeling, analysis, control, estimation, optimization, and design of these systems. Matured and validated through more than a decade of research spanning multiple academic institutions and companies, the graph-based approach combines transient energy conservation with an explicit mathematical representation of the network by which energy is stored and transferred within a system. Following a mathematical overview of graph-based models, examples of multi-domain component and system models from the recent literature are presented, including single-phase thermal systems, two-phase thermal systems, and electro-mechanical systems. This is followed by a survey of recent applications for decentralized and hierarchical model predictive control, design optimization, and control co-design. Lastly, the paper describes an open-source toolbox created to facilitate the generation and analysis of graph-based models.

An over-the-air (OTA) experimental evaluation of concurrent 5G New Radio (5G NR) and Wi-Fi transmission using successive interference cancellation (SIC) in a shielded-box environment is presented. A USRP is used as the receiver, which captures the composite waveform containing both air-interface signals and applies sample-domain SIC to suppress the dominant 5G-NR signal and recover Wi-Fi signal from the residual waveform. The framework reports error vector magnitude (EVM), bit error rate (BER), sample-domain cancellation depth, and channel-estimate suppression, and, at the representative \(18\) dB attenuation point, measures \(11.88\) dB cancellation depth and \(26.96\) dB 5G channel suppression. The proposed methodology provides a practical basis for assessing cross-technology coexistence and receiver-side interference suppression under controlled OTA conditions.

This paper presents a controlled over-the-air (OTA) characterization of dual-user IEEE 802.11be Extremely High Throughput Multi-User (EHT-MU) transmission under transmit-chain imbalance. The objective is to determine whether attenuation applied to one access-point transmit chain produces packet-global degradation or appears primarily as stream-dependent payload degradation after receiver processing. Measurements are performed in a shielded RF enclosure using two NI USRP-2953R and NI USRP-2942R software-defined radios, with one USRP generating a dual-user non-OFDMA EHT-MU waveform and the other implementing synchronized dual-branch packet recovery. A calibrated attenuation sweep is applied to the second AP transmit chain (TX2), and performance is evaluated using bit error rate (BER), EHT-Data error vector magnitude (EVM), control-field success probability, payload-success probability, and subcarrier-level EVM distributions. The results show that the stream decoded as User~1 remains at the BER floor over the tested range, while the stream decoded as User~2 exhibits progressive EVM degradation followed by threshold-like BER and payload-success collapse. Common signaling fields remain recoverable, indicating that the dominant observed failure mode is stream-local at the receiver output than the packet-global. Replacing User~2 binary convolutional coding (BCC) with low density parity check (LDPC) coding delays the BER and payload-success collapse by approximately \(5\)~dB of TX2 attenuation, demonstrating a measurable coding-dependent robustness margin for the more sensitive stream.

Submarine eruptions, accounting for over 80% of Earth's volcanic activity, primarily occur along mid-ocean ridges, where shallow magmatic systems are accessible to high-resolution imaging. Yet, their remoteness often leaves them undetected. Recent seismic studies at the East Pacific Rise (EPR) 9°50'N-one of the most dynamic ridge segments, imaged the detailed architecture of the shallowest magma lens, but no data-constrained model yet explains how magma accumulates, migrates, or triggers eruptions. Similarly, the formation of oceanic crust remains poorly understood. While 2-D seismic data reveal only a few vertically stacked, transient magma lenses, our study applies matrix imaging, a novel technique in controlled-source seismology, to map the inner structure of on- and off-axis magma reservoirs. We uncover a conical on-axis reservoir and interconnected magma-rich zones throughout the crust. Combined with ophiolite evidence, these findings reveal that magma channels dominate the first 3 km for lower crust formation, while in situ crystallization prevails in the final 1 km, resolving a long-standing debate.

To support the adoption of electric transport systems, public charging opportunities are becoming increasingly important. In this dynamic environment, a central challenge for route planning and charging scheduling is forecasting charging-station availability under fluctuating demand. In this work, we propose a fluid-based forecasting method that accounts for uncertainty in both known and unforeseen electric vehicle arrival patterns while respecting station capacity constraints. We further evaluate the congestion forecasting method by applying it to an electric vehicle scheduling problem. Compared to scheduling frameworks that rely on standard baselines, charging schedules based on the fluid congestion forecasting model reduce waiting-related downtime by up to 14%. Finally, we quantify how increased knowledge of vehicle arrivals and different levels of station congestion affect overall system performance.

In recent years, progress in adaptive graph signal processing algorithms has provided effective solutions for processing signals defined on graph structures. As a classical strategy in information theory, the Generalized Maximum Correntropy Criterion (GMCC) exhibits good resistance to non-Gaussian noises. When non-Gaussian noise interferes with the graph signal, the graph signal processing algorithm based on GMCC (GSP GMCC) algorithm shows better performance. However, the GSP GMCC algorithm itself has three parameters that need to be manually tuned, and the process of manually tuning the parameters is complex and tedious. Meanwhile, the non-concave and non-convex nature of the GMCC function itself limits its own convergence rate and adaptive estimation accuracy. To solve the above problems, based on the strongly convex function half-quadratic criterion (HQC), the GSP HQC algorithm is proposed in this paper. The performance analysis of the GSP HQC algorithm is implemented in this paper. Simulation experiments demonstrate that the GSP HQC algorithm achieves superior performance in terms of convergence rate and adaptive estimation accuracy while maintaining computational complexity comparable to existing algorithms

Low Earth Orbit (LEO) satellite constellations are emerging as a key component of non-terrestrial networks due to their low-latency and high-capacity communication capabilities. However, satellites in these orbits are characterized by a small coverage footprint and high orbital velocity compared to those in higher orbits. This results in constantly changing and dynamic constellations that require smart design of orbital parameters to ensure continuous coverage. Existing constellation deployments are typically optimized either for low- and mid-latitude regions or for full polar coverage, leaving high-latitude regional scenarios such as the North Atlantic insufficiently explored. This work provides insights into the key characteristics associated with the deployment of satellites in LEO for North Atlantic coverage. Therefore, we investigate how constellation inclination, minimum elevation angle, altitude, and satellite footprint jointly affect visibility probability, revisit time, path loss, and coverage continuity. Results show that the minimum elevation angle is a critical design parameter since a Walker Delta constellation with 64 satellites at 1000 km altitude can provide continuous coverage above 55°N for elevations below 20°, whereas coverage probability degrades drastically for larger elevation angles. Similarly, inclinations above approximately 70° are required to achieve robust North Atlantic coverage with medium-size constellations. Thus, these results provide practical guidelines on how a satellite constellation should be designed to achieve an efficient deployment with a focus on coverage over the North Atlantic, targeting maritime, aviation, and Arctic connectivity scenarios.

User localization and beam management are tightly linked in extremely large-scale multiple-input multiple-output (XL-MIMO) systems, especially in dense low-altitude economy (LAE) scenarios. However, the near-field propagation in XL-MIMO introduces strong distance sensitivity and complex spatial coupling, which makes joint trajectory and beam prediction challenging. Meanwhile, large language models (LLMs) have attracted attention in physical-layer transmission for modeling long-range dependencies. In this paper, we propose NF-TrackLLM, a multi-modal semantic-aware framework for near-field unmanned aerial vehicles (UAVs) positioning and beam prediction in XL-MIMO systems. By incorporating visual and LiDAR sensing into a Sionna-based channel generation pipeline, environmental semantics and GPS are utilized to guide trajectory and beam prediction. Built upon the aligned multi-modal representation, a GPT-2-based spatiotemporal reasoning backbone, and a cascaded prediction strategy are employed, where future trajectories are first inferred and then used to guide beam prediction as geometric priors. Simulation results demonstrate that NF-TrackLLM achieves accurate beam prediction and reliable UAV trajectory tracking in dense urban low-altitude scenarios.

This paper introduces a Gaussian process (GP)-based method for extended object estimation (EOE) in integrated sensing and communication (ISAC) scenarios, representing a promising approach to enhance environmental awareness beyond the conventional point-scatterer assumption. The suitability of the proposed GP-based method for EOE is investigated through a practical measurement setup compliant with the fifth-generation (5G) New Radio (NR) standard and employing bistatic sensing, with results evaluated for both mapping and simultaneous localization and mapping (SLAM ) cases at millimeter-wave (mmWave) frequencies. The findings reveal that the enhanced capabilities of communication networks, when combined with bistatic sensing and GP-based EOE, enable improved environmental awareness in future wireless systems. Importantly, the results demonstrate that, under practical conditions, GP effectively performs EOE in both mmWave mapping and SLAM scenarios.

The growing adoption of electric vehicles (EVs) is increasing peak demand in distribution systems, which can threaten grid stability and reduce operational efficiency. Dynamic electricity pricing is a promising means of mitigating these peaks by shifting flexible demand. However, most existing approaches rely on detailed user-level consumption data and behavioral models, which are often difficult to obtain in practice and may raise privacy concerns. This paper proposes an Online Feedback Optimization (OFO) algorithm for day-ahead price design with limited data, where only aggregate loads are observed. OFO updates prices iteratively using aggregate load measurements, enabling effective peak reduction without access to individual user data. The formulation also includes a term that penalizes deviations in total electricity cost relative to a reference tariff. Although relying only on aggregate load measurements, the OFO price updates efficiently converge to the optimal price. In finite-horizon simulations, OFO achieves peak reduction close to that of the Stackelberg benchmark with full model information. Meanwhile, its computational effort is substantially lower. Additional tests under multiple initial conditions and delayed charging-window mismatch further confirm the robustness of the proposed method. Overall, these results show that OFO is a scalable and computationally efficient approach for peak-demand management in distribution systems with limited observability.

This paper develops a robust fixed time optimization framework for constrained problems that guarantees exact constraint satisfaction and convergence to KKT points within fixed time , independent of initial conditions. The approach treats the Lagrange multipliers as control inputs, composed of an equivalent control and a switching control, with the system states representing the decision variables. An equivalent control steers the gradient flow to a local KKT point asymptotically for nonconvex objectives and to unique global optimum in fixed time for convex objectives. Constraint enforcement is achieved by embedding the equality constraints directly as a sliding manifold, with a fixed time switching control ensuring rapid and reliable feasibility. The framework further accounts for the matched disturbances, providing robustness guarantees that are theoretically characterized and illustrated using spherical constraints. Numerical studies on a 3-bus AC optimal power flow problem and distributed consensus=based parameter estimation problem demonstrate the effectiveness, scalability and robustness of proposed approach.

Gaussian Splatting (GS) has emerged as an efficient representation for high-quality 3D reconstruction and novel view synthesis. However, its large model size poses challenges for storage and transmission. While several GS compression solutions have been proposed, their perceptual impact remains poorly understood due to the lack of dedicated evaluation datasets. To address this gap, this paper introduces GScomp-QA, a subjective quality assessment dataset for evaluating synthesis quality from compressed GS models. The dataset comprises 331 video stimuli from 13 real-world scenes, covering 9 state-of-the-art GS compression solutions. By using videos synthesized from uncompressed models as reference, GScomp-QA isolates compression-induced distortions from synthesis artifacts. A subjective study with 20 participants was conducted, providing reliable perceptual scores. Based on these data, GS compression solutions are evaluated through perceptual rate-distortion analysis. In addition, 18 objective quality metrics are evaluated, showing that they do not fully capture GS-specific distortions. GScomp-QA will be publicly available and provide a benchmark for evaluating GS compression solutions and supporting the development of quality metrics tailored to GS compression.

This article presents a closed-loop, uncertainty-aware framework for defending a protected zone against coordinated incursions by swarms of small uncrewed aircraft systems (UAS). The interaction structure of the attackers is modeled as time-varying, while defenders operate under imperfect sensing. The proposed criticality-driven defender-to-attacker assignment strategy integrates three components: a probabilistic graph-based representation of the attacking swarm inferred from uncertain observations; a risk-aware attacker criticality model combining time-to-breach urgency with uncertainty; an online defender allocation mechanism that assigns and selectively reassigns defenders while limiting switching-induced instability through robust execution constraints. Analytical guarantees are established within a filtration-based first-hitting-time framework. In particular, finite-time triggering of the first capture event following detection is proven, and explicit mixed linear-geometric upper bounds are derived for the expected neutralization time. Monte Carlo simulations demonstrate the effectiveness of the proposed framework, achieving 85.6% neutralization efficiency under probabilistic sensing and 99.9% under deterministic sensing. Systematic ablation and sensitivity studies further quantify how detection thresholds and coordination parameters influence reliability and time-to-first-capture.

Wi-Fi sensing has emerged as a versatile tool for tasks such as localization, gesture recognition, and vital-sign monitoring, enabling applications from smart environments to personalized healthcare. However, sensing accuracy often significantly degrades when pretrained models are deployed across different commodity receivers. We present the first systematic comparison of Channel State Information (CSI) across diverse Commercial Off-The-Shelf Wi-Fi sensing platforms. Using a unified experimental setup delivering precisely precoded signals simultaneously to multiple receivers, we isolate receiver-specific variability. We find that dominant cross-device differences arise from Automatic Gain Control and consistent subcarrier nonlinearities. We propose a simple gain-alignment preprocessing step, recovering most of the lost accuracy (up to 75%) in cross-device Human Activity Recognition model deployments. Without preprocessing, model accuracy sharply drops-effectively breaking practical deployments. Additional analyses reveal measurable inherent differences in receiver faithfulness, sensitivity and noise. While these receiver-induced differences do not significantly affect robust sensing tasks such as Human Activity Recognition, they become relevant in scenarios demanding high precision (e.g., single-shot time of flight). Our findings demonstrate that cross-device variability in CSI is real but manageable, and we provide tools and guidelines for robust, hardware-agnostic Wi-Fi sensing.

High-quality speech coding at low bitrates is crucial for bandwidth-constrained applications, yet remains challenging due to the severe loss of quality-critical information in highly compressed representations. To overcome this challenge, we propose CFMDCTCodec, a low-bitrate neural speech codec that operates entirely in the modified discrete cosine transform (MDCT) domain. CFMDCTCodec integrates a lightweight encoder-quantizer-decoder-style MDCT-spectral codec with a noise-prior-aware, conditional-flow-matching (CFM)-based MDCT-spectral enhancer. Within this framework, the codec serves as a base module that compactly discretizes the MDCT spectrum extracted from speech and produces an initial coarse reconstruction, while the enhancer further restores fine-grained spectral details. The enhancer improves the decoded MDCT spectrum by integrating a conditional MDCT velocity-field filter with an ordinary differential equation (ODE) solver, under the guidance of an MDCT-derived magnitude-adaptive noise prior, aiming to emphasize perceptually significant high-energy regions while stabilizing low-energy and silent regions. Finally, the enhanced MDCT spectrum is reconstructed into the decoded speech using the inverse MDCT. When optimizing CFMDCTCodec, we adopt a unified non-adversarial training strategy that jointly combines reconstruction, quantization and CFM objectives. Both objective and subjective evaluations show that CFMDCTCodec outperforms competitive baselines in low-bitrate regimes, e.g., 0.65 kbps, while approaching the perceptual quality of large-scale codecs with significantly fewer parameters and computations.

Incentive-based load curtailment unlocks critical demand-side flexibility but is hindered by the limited knowledge of private user parameters and the inherent nonsmoothness of responses due to physical device constraints. We address this via a constrained bilevel optimization framework and propose the Bi-ZOL (Bilevel Zeroth-Order Learning) algorithm. Unlike conventional black-box methods, Bi-ZOL exploits the bilevel structure to decompose the hypergradient, integrating the exact analytical information of the SO's objective with a zeroth-order estimate of the unknown response sensitivity. This structural decomposition-based learning method mathematically smoothes the nonsmooth response landscape and reduces hypergradient estimation error. We provide theoretical convergence guarantees to an approximate stationary point and demonstrate through simulations that Bi-ZOL achieves near-optimal performance.

In this work, we introduce a real-time capable algorithm for considering monotonicity assumptions for recursive Gaussian Process regression (RGP). Therefore, we present how to efficiently calculate the RGP gradients online. Then, we utilize an extended Kalman filter and pseudo-measurements in combination with a ReLU pseudo-measurement function to enforce soft inequality constraints. This work builds upon a previously published conference paper with the same goal and a similar fundamental approach. Opposite to our previous work, however, we now use an exact covariance calculation for the RGP gradients. Furthermore, we also present a real-time optimized version of this algorithm with less simplifications compared to the previously published version. These and several other algorithmic innovations lead to an algorithm with greatly improved numerical robustness. The algorithm is validated and compared to its previously published version for a 2D numerical example. The paper is concluded with a successful experimental validation of the developed algorithm for the monotonicity-preserving learning of pneumatic valve characteristics for the control of a pneumatic system, leveraging a partial input - output linearization.

Multimodal signals on sensor networks are commonly modeled under the twofold graph assumption (TGA), which represents spatial structure and inter-modality relations as two separate graphs. Existing TGA-based signal restoration methods, however, either assume the graphs are known or restrict edge weights to be non-negative, preventing them from capturing negative inter-modal correlations. We address both limitations by formulating joint signal restoration and twofold graph learning as MAP estimation under a matrix normal prior, where the spatial and modality graph Laplacians appear directly as precision matrices. The resulting non-convex objective is solved by alternating minimization: The signal is updated via conjugate gradient applied to the arising Sylvester-type linear system; the graphs are updated via primal-dual hybrid gradient (PDHG). We further propose a method to estimate the signed structure of the modality graph from the dominant eigenspace of a complementary kernel matrix, which is then used in PDHG to update edge magnitudes. These iterative solvers are then unrolled into a feedforward network, with regularization weights and step sizes as layer-wise trainable parameters. Experiments on synthetic multimodal graph signals and a real Japan meteorological dataset confirm that the proposed method outperforms existing baselines across a range of noise levels and missing-data patterns.

Ultrasound Localization Microscopy (ULM) has presented great potential in functional imaging, benefiting from its ability to reconstruct deep microvasculature. However, the hemodynamic reconstruction is compromised by sparsity in the ULM data, as a limited number of MB tracks cannot sample the complete speed profile in one vessel. Here, we propose to reconstruct hemodynamics using sparse ULM velocity maps by solving a laminar flow model through stochastic variational inference. In addition to vascular geometry and flow velocity maps, the proposed method generates two new ULM maps - a pressure gradient map and a map describing uncertainty of the estimation. By investigating the effect of sparsity in ULM maps on the quantification and visualization of hemodynamics, we demonstrate the effectiveness of the proposed method in dealing with sparse ULM maps via simulations and 3D rat brain imaging. Accurately reconstructing a broad range of hemodynamic parameters and associate uncertanties using sparse ULM data may help detect subtle and dynamic brain activity.