Papers

The advent of artificial intelligence (AI)-native wireless communication is fundamentally reshaping the design paradigm of next-generation (NextG) systems, where intelligent air interfaces are expected to operate adaptively and efficiently in highly dynamic environments. Conventional orthogonal frequency division multiplexing (OFDM) systems rely heavily on pilots and the cyclic prefix (CP), resulting in significant overhead and reduced spectral efficiency. To address these limitations, we propose an adaptive end-to-end (E2E) transceiver architecture tailored for pilot-free and CP-free wireless systems. The architecture combines AI-driven constellation shaping and a neural receiver through joint training. To enhance robustness against mismatched or time-varying channel conditions, we introduce a lightweight channel adapter (CA) module, which enables rapid adaptation with minimal computational overhead by updating only the CA parameters. Additionally, we present a framework that is scalable to multiple modulation orders within a unified model, significantly reducing model storage requirements. Moreover, to tackle the high peak-to-average power ratio (PAPR) inherent to OFDM, we incorporate constrained E2E training, achieving compliance with PAPR targets without additional transmission overhead. Extensive simulations demonstrate that the proposed framework delivers superior bit error rate (BER), throughput, and resilience across diverse channel scenarios, highlighting its potential for AI-native NextG.

The rapid deployment of unmanned aerial vehicle (UAV) corridors in sixth-generation (6G) networks requires safe, intelligence-driven integrated sensing and communications (ISAC). Reconfigurable intelligent surfaces (RIS) enhance spectrum efficiency, localisation accuracy, and situational awareness, while introducing new vulnerabilities. The rise of quantum computing increases the risks associated with harvest-now-decrypt-later strategies and quantum-enhanced spoofing. We propose a Quantum-Resilient Threat Modelling (QRTM) framework for RIS-assisted ISAC in UAV corridors to address these challenges. QRTM integrates classical, quantum-ready, and quantum-aided adversaries, countered using post-quantum cryptographic (PQC) primitives: ML-KEM for key establishment and Falcon for authentication, both embedded within RIS control signalling and UAV coordination. To strengthen security sensing, the framework introduces RIS-coded scene watermarking validated through a generalised likelihood ratio test (GLRT), with its detection probability characterised by the Marcum Q function. Furthermore, a Secure ISAC Utility (SIU) jointly optimises secrecy rate, spoofing detection, and throughput under RIS constraints, enabled by a scheduler with computational complexity of O(n^2). Monte Carlo evaluations using 3GPP Release 19 mid-band urban-canyon models (7-15 GHz) demonstrate a spoof-detection probability approaching 0.99 at a false-alarm rate of 1e-3, secrecy-rate retention exceeding 90 percent against quantum-capable adversaries, and signal-interference utilisation improvements of about 25 percent compared with baselines. These results show a standards-compliant path towards reliable, quantum-resilient ISAC for UAV corridors in smart cities and non-terrestrial networks.

Satellite-based communications are expected to be a substantial future market in 6G networks. As satellite constellations grow denser and transmission resources remain limited, frequency reuse plays an increasingly important role in managing inter-user interference. In the multi-user downlink, precoding enables the reuse of frequencies across spatially separated users, greatly improving spectral efficiency. The analytical calculation of suitable precodings for perfect channel information is well studied, however, their performance can quickly deteriorate when faced with, e.g., outdated channel state information or, as is particularly relevant for satellite channels, when position estimates are erroneous. Deriving robust precoders under imperfect channel state information is not only analytically intractable in general but often requires substantial relaxations of the optimization problem or heuristic constraints to obtain feasible solutions. Instead, in this paper we flexibly derive robust precoding algorithms from given data using reinforcement learning. We describe how we adapt the applied Soft Actor-Critic learning algorithm to the problem of downlink satellite beamforming and show numerically that the resulting precoding algorithm adjusts to all investigated scenarios. The considered scenarios cover both single satellite and cooperative multi-satellite beamforming, using either global or local channel state information, and two error models that represent increasing levels of uncertainty. We show that the learned algorithms match or markedly outperform two analytical baselines in sum rate performance, adapting to the required level of robustness. We also analyze the mechanisms that the learned algorithms leverage to achieve robustness. The implementation is publicly available for use and reproduction of the results.

Accurate channel state information (CSI) is critical for current and next-generation multi-antenna systems. Yet conventional pilot-based estimators incur prohibitive overhead as antenna counts grow. In this paper, we address this challenge by developing a novel framework based on Gaussian process regression (GPR) that predicts full CSI from only a few observed entries, thereby reducing pilot overhead. The correlation between data points in GPR is defined by the covariance function, known as kernel. In the proposed GPR-based CSI estimation framework, we incorporate three kernels, i.e., radial basis function, Mat'ern, and rational quadratic, to model smooth and multi-scale spatial correlations derived from the antenna array geometry. The proposed approach is evaluated across two channel models with three distinct pilot probing schemes. Results show that the proposed GPR with 50% pilot saving achieves the lowest prediction error, the highest empirical 95% credible-interval coverage, and the best preservation of spectral efficiency relative to the benchmarks.

This paper introduces AirCNN, a novel paradigm for implementing convolutional neural networks (CNNs) via over-the-air (OTA) analog computation. By leveraging multiple reconfigurable intelligent surfaces (RISs) and transceiver designs, we engineer the ambient wireless propagation environment to emulate the operations of a CNN layer. To comprehensively evaluate AirCNN, we consider two types of CNNs, namely classic two-dimensional (2D) convolution (Conv2d) and light-weight convolution, i.e., depthwise separable convolution (ConvSD). For Conv2d realization via OTA computation, we propose and analyze two RIS-aided transmission architectures: multiple-input multiple-output (MIMO) and multiple-input single-output (MISO), balancing transmission overhead and emulation performance. We jointly optimize all parameters, including the transmitter precoder, receiver combiner, and RIS phase shifts, under practical constraints such as transmit power budget and unit-modulus phase shift requirements. We further extend the framework to ConvSD, which requires distinct transmission strategies for depthwise and pointwise convolutions. Simulation results demonstrate that the proposed AirCNN architectures can achieve satisfactory classification performance. Notably, Conv2d MISO consistently outperforms Conv2d MIMO across various settings, while for ConvSD, MISO is superior only under poor channel conditions. Moreover, employing multiple RISs significantly enhances performance compared to a single RIS, especially in line-of-sight (LoS)-dominated wireless environments.

This survey provides an analysis of current methodologies integrating legal and logical specifications into the perception, prediction, and planning modules of automated driving systems. We systematically explore techniques ranging from logic-based frameworks to computational legal reasoning approaches, emphasizing their capability to ensure regulatory compliance and interpretability in dynamic and uncertain driving environments. A central finding is that significant challenges arise at the intersection of perceptual reliability, legal compliance, and decision-making justifiability. To systematically analyze these challenges, we introduce a taxonomy categorizing existing approaches by their theoretical foundations, architectural implementations, and validation strategies. We particularly focus on methods that address perceptual uncertainty and incorporate explicit legal norms, facilitating decisions that are both technically robust and legally defensible. The review covers neural-symbolic integration methods for perception, logic-driven rule representation, and norm-aware prediction strategies, all contributing toward transparent and accountable autonomous vehicle operation. We highlight critical open questions and practical trade-offs that must be addressed, offering multidisciplinary insights from engineering, logic, and law to guide future developments in legally compliant autonomous driving systems.

This paper investigates the energy implications of remote rendering for Virtual Reality (VR) applications within a real 5G testbed. Remote rendering enables lightweight devices to access high-performance graphical content by offloading computationally intensive tasks to Cloud-native Network Functions (CNFs) running on remote servers. However, this approach raises concerns regarding energy consumption across the various network components involved, including the remote computing node, the 5G Core, the Radio Access Network (RAN), and the User Equipment (UE). This work proposes and evaluates two complementary energy monitoring solutions, one hardware-based and one software-based, to measure energy consumption at different system levels. A VR remote renderer, deployed as CNF and leveraging the Media over QUIC (MoQ) protocol, is used as test case for assessing its energy footprint under different multimedia and network configurations. The results provide critical insights into the trade-off between energy consumption and performance of a real-world VR application running in a 5G environment.

Federated learning enables collaborative machine learning while preserving data privacy, but high communication and computation costs, exacerbated by statistical and device heterogeneity, limit its practicality in mobile edge computing. Existing compression methods like sparsification and pruning reduce per-round costs but may increase training rounds and thus the total training cost, especially under heterogeneous environments. We propose a lightweight personalized FL framework built on parameter decoupling, which separates the model into shared and private subspaces, enabling us to uniquely apply gradient sparsification to the shared component and model pruning to the private one. This structural separation confines communication compression to global knowledge exchange and computation reduction to local personalization, protecting personalization quality while adapting to heterogeneous client resources. We theoretically analyze convergence under the combined effects of sparsification and pruning, revealing a sparsity-pruning trade-off that links to the iteration complexity. Guided by this analysis, we formulate a joint optimization that selects per-client sparsity and pruning rates and wireless bandwidth to reduce end-to-end training time. Simulation results demonstrate faster convergence and substantial reductions in overall communication and computation costs with negligible accuracy loss, validating the benefits of coordinated and resource-aware personalization in resource-constrained heterogeneous environments.

Trajectory planning in dense, interactive traffic scenarios presents significant challenges for autonomous vehicles, primarily due to the uncertainty of human driver behavior and the non-convex nature of collision avoidance constraints. This paper introduces a stochastic optimal control framework to address these issues simultaneously, without excessively conservative approximations. We opt to model human driver decisions as a Markov Decision Process and propose a method for handling collision avoidance between non-convex vehicle shapes by imposing a positive distance constraint between compact sets. In this framework, we investigate three alternative chance constraint formulations. To ensure computational tractability, we introduce tight, continuously differentiable reformulations of both the non-convex distance constraints and the chance constraints. The efficacy of our approach is demonstrated through simulation studies of two challenging interactive scenarios: an unregulated intersection crossing and a highway lane change in dense traffic.

High-fidelity large-field-of-view (LFOV) 3D sensing, essential for autonomous platforms, is hindered by the coupling of anisotropic off-axis aberrations and the ill-posed nature of monocular depth estimation. To address this fundamental physical bottleneck, we propose the Bio-inspired Monocentric Imaging (BMI) framework, a holistic co-design integrating a monocentric optical topology with a physics-aware reconstruction network. By pairing a concentric spherical lens with a hemispherical sensor, our system structurally eliminates coma, astigmatism, and field curvature. Instead of relying on external modulators or hardware redundancy, we exploit the intrinsic aberration as a deterministic, radially symmetric carrier for depth encoding. Theoretical analysis via the Cramér-Rao Lower Bound confirms that this topology maintains superior depth sensitivity across the entire FOV. To decode these optical cues, our framework utilizes a dual-head network to jointly recover high-fidelity All-in-Focus images and dense metric depth maps from single-shot captures. The efficacy of this co-design is evaluated through physically-based simulations of field-dependent PSFs across the full optical FOV. On NYU Depth V2, our system achieves a superior balance of image fidelity (31.15dB PSNR) and depth precision (0.161m RMSE). It further maintains consistent reconstruction under spatially varying PSFs sampled across a 120{\deg} optical FOV, mitigating the peripheral degradation observed in conventional wide-field designs. The source code is publicly available at https://github.com/ZongxiYu-ZJU/BMI.

This paper presents a fully data-driven 3-D path-following framework for autonomous underwater vehicles (AUVs), a representative class of underwater field robotics, based on Data-Enabled Predictive Control (DeePC). The approach eliminates explicit hydrodynamic modeling by exploiting measured input-output trajectories to predict and optimize future system behavior. Classic DeePC is employed for heading control, while a cascaded DeePC architecture with loop-frequency separation is proposed for depth regulation, extending DeePC to plants whose dominant output evolves significantly slower than the actuator bandwidth. For 3-D waypoint path following, the Adaptive Line-of-Sight (ALOS) guidance law is extended to a predictive multistep formulation (PALOS) that supplies the horizon-consistent reference required by receding-horizon predictive controllers. All methods are validated in high-fidelity 6 degrees of freedom simulation on the REMUS~100 AUV under nominal operation, ocean-current disturbances, operation beyond the data regime, and 3-D waypoint path following, consistently outperforming the corresponding state-of-the-art benchmarks. In 3-D waypoint path following, the framework reduces cross-track error by approximately 28% relative to the ALOS-PI/PID baseline.

This study demonstrates the feasibility of radar-based non-contact respiratory monitoring for wombats. Two measurement experiments were conducted in June and December 2024 using 79-GHz millimeter-wave radar systems to monitor the respiration of two wombats. To estimate the respiratory interval, we used a method based on summing harmonic components in the autocorrelation function, capturing the quasi-periodic displacement of the body surface caused by respiration. Estimation accuracy was evaluated through simultaneous measurements from different angles using two radar units. The respiratory interval and respiratory rate were measured with errors of 47.4 ms (2.44%) and 0.81 bpm (2.21%), respectively. We also discuss the differences in respiratory rates between the two wombats, as well as seasonal variations between June and December. The results support the potential application of this method to non-contact health monitoring of wombats.

This paper explores a multi-cell multiple-input single-output (MISO) downlink communication system enabled by a unique transmissive reconfigurable intelligent surface (TRIS) transceiver configuration. Within this system framework, we formulate an optimization problem for the purpose of maximizing the minimum rate of users for each cell via designing the transmit beamforming of the TRIS transceiver, subject to the power constraints of each TRIS transceiver unit. Since the objective function is non-differentiable, the max-min rate problem is difficult to solve. In order to tackle this challenging optimization problem, an efficient low-complexity optimization algorithm is developed. Specifically, the log-form rate function is transformed into a tractable form by employing the fractional programming (FP) methodology. Next, the max-min objective function can be approximated using a differentiable function derived from smooth approximation theory. Moreover, by applying the majorization-minimization (MM) technique and examining the optimality conditions, a solution is proposed that updates all variables analytically without relying on any numerical solvers. Numerical results are presented to demonstrate the convergence and effectiveness of the proposed low-complexity algorithm. Additionally, the algorithm can significantly reduce the computational complexity without performance loss. Furthermore, the simulation results illustrate the clear superiority of the deployment of the TRIS transceiver over the benchmark schemes.

Lane changes are common yet challenging driving maneuvers that require continuous decision-making and dynamic interaction with surrounding vehicles. Relying solely on human drivers for lane-changing can lead to traffic disturbances due to the stochastic nature of human behavior and its variability under different task demands. Such uncertainties may significantly degrade traffic string stability, which is critical for suppressing disturbance propagation and ensuring smooth merging of the lane-changing vehicles. This paper presents a human-automation shared lane-changing control framework that preserves driver authority while allowing automated assistance to achieve stable maneuvers in the presence of driver's behavioral uncertainty. Human driving behavior is modeled as a Markov jump process with transitions driven by task difficulty, providing a tractable representation of stochastic state switching. Based on this model, we first design a nominal stabilizing controller that guarantees stochastic ${L}_2$ string stability under imperfect mode estimation. To further balance performance and automated effort, we then develop a Minimal Intervention Controller (MIC) that retains acceptable stability while limiting automation. Simulations using lane-changing data from the NGSIM dataset verify that the nominal controller reduces speed perturbations and shorten lane-changing time, while the MIC further reduces automated effort and enhances comfort but with moderate stability and efficiency loss. Validations on the TGSIM dataset with SAE Level 2 vehicles show that the MIC enables earlier lane changes than Level 2 control while preserving driver authority with a slight stability compromise. These findings highlight the potential of shared control strategies to balance stability, efficiency, and driver acceptance.

Several centimeters below the skin lie multiple biomarkers, such as glucose, oxygenation, and blood flow. Monitoring these biomarkers regularly and in a non-invasive manner would enable early insight into metabolic status and vascular health. Currently, there are only a handful of non-invasive monitoring systems. Optical methods offer molecular specificity (i.e., multi-biomarker monitoring) but have shallow reach (a few millimeters); ultrasound penetrates deeper but lacks specificity; and MRI is large, slow, and costly. Photoacoustic (PA) sensing combines the best of optical and ultrasound methods. A laser transmitter emits pulses that are absorbed by different molecules, providing specificity. These light pulses generate pressure changes that are captured by an ultrasound receiver, providing depth. Photoacoustic sensing is promising, but the current platforms are bulky, complex, and costly. We propose the first embedded PA platform. Our contributions are fourfold. First, inspired by LiDAR technology, we propose a novel transmitter that emits pulses similar to those in the state-of-the-art (SoA), but instead of using high-voltage sources and complex electronic interfaces, we use a simple low-power microcontroller (MCU). Second, we carry out a thorough analysis of our custom transmitter and a commercial system. Third, we build a basic ultrasound receiver that is able to process the faint signal generated by our transmitter. Lastly, we compare the performance of our platform against a SoA commercial system, and show that we can detect glucose and (de)oxygenated hemoglobin in two controlled solution studies. The resulting signal characteristics indicate a plausible path toward noninvasive, real-time, at-home sensing relevant to diabetes care. More broadly, this platform lays the groundwork for translating the promise of PA sensing into a broader practical reality.

Integrated sensing and communication (ISAC) is a promising solution for the future sixth-generation (6G) system. However, classical fixed-position antenna (FPA) ISAC systems fail to fully utilize spatial degrees of freedom (DoFs), resulting in limited gains for both radar sensing and communication functionalities. This challenge can be addressed by the emerging novel fluid antenna (FA) technology, which can pursue better channel conditions and improve sensing and communication performances. In this paper, we aim to minimize the Cramér-Rao bound (CRB) for estimating the target's angle while guaranteeing communication performance. This involves jointly optimizing active beamforming, power allocation, receiving filters, and FA position configurations, which is a highly non-convex problem. To tackle this difficulty, we propose an efficient iterative solution that analytically optimizes all variables without relying on numerical solvers, i.e., CVX. Specifically, by leveraging cutting-edge majorization-minimization (MM) and penalty-dual-decomposition (PDD) methods, we develop a low-complexity algorithm to solve the beamformer configuration problem containing the fractional and quartic terms. Numerical simulation results demonstrate the effectiveness and efficiency of our proposed algorithm, highlighting significant performance improvements achieved by employing FA in the ISAC system.

A problem of achieving minimum time consensus for a set of $N$ second-order LTI system agents with bounded inputs and fuel constraints is considered. Unlike our other works, here the damping effect in agent dynamics is included. First, the attainable set for each agent with fuel budget constraints is characterized, and its boundary equations are derived. Then, using the convexity property, the minimum time at which attainable sets of all agents have a non-empty intersection is computed. By applying Helly's theorem, the computation reduces to finding the minimum time to consensus and the corresponding consensus point for each of the triplets separately.

Age-related hearing loss (HL) reduces speech intelligibility (SI) in older adults (OAs). However, deficits in central and cognitive processing also substantially impact SI. Understanding these contributions is essential for explaining individual differences and developing effective assistive hearing strategies. This study presents a framework that distinguishes peripheral HL from central and cognitive influences on SI. This framework uses the Wakayama University Hearing Impairment Simulator (WHIS), and the Gammachirp Envelope Similarity Index (GESI), an objective measure of intelligibility. First, speech-in-noise tests were conducted with young, normal-hearing listeners (YNHs) using WHIS to simulate the audiogram of a target OA. The target OA achieved SI scores comparable to or higher than those of YNHs with simulated HL, suggesting contributions beyond peripheral hearing function. Then, GESI was used to predict SI scores for YNHs and OAs across different hearing levels. The prediction accuracy was comparable for both groups. Interestingly, many OAs' subjective SI scores were higher than those predicted using parameters derived from YNHs' experiments. This finding is inconsistent with previous research indicating that speech perception ability declines with age. This issue will be discussed. There was no significant correlation between the average hearing levels and the residual differences between the subjective and predicted SI scores. This suggests that GESI effectively absorbed the effects of peripheral HL. Thus, the proposed framework may facilitate systematic examination and comparison of central and cognitive factors beyond peripheral HL among individual YNHs and OAs with and without HL.

The continuous miniaturisation of metal-oxide-semiconductor field-effect transistors (MOSFETs) from long- to short-channel architectures has advanced beyond the predictions of Moore's Law. Continued advances in semiconductor electronics, even near current scaling and performance boundaries under cryogenic conditions, are driving the development of innovative device paradigms that enable ultra-low-power and high-speed functionality. Among emerging candidates, the Josephson Junction Field-Effect Transistor (JJFET or JoFET) provides an alternative by integrating superconducting source and drain electrodes for efficient, phase-coherent operation at ultra-low temperatures. These hybrid devices have the potential to bridge conventional semiconductor electronics with cryogenic logic and quantum circuits, enabling energy-efficient and high-coherence signal processing across temperature domains. This review traces the evolution from Josephson junctions to field-effect transistors, emphasising the structural and functional innovations that underpin modern device scalability. The performance and material compatibility of JJFETs fabricated on Si, GaAs, and InGaAs substrates are analysed, alongside an assessment of their switching dynamics and material compatibility. Particular attention is given to superconductor-silicon-superconductor Josephson junctions as the active core of JJFET architectures. By unfolding more than four decades of experimental progress, this work highlights the promise of JJFETs as foundational building blocks for next-generation cryogenic logic and quantum electronic systems.

Accurate direction-of-arrival (DOA) estimation for sound sources is challenging due to the continuous changes in acoustic characteristics across time and frequency. In such scenarios, accurate localization relies on the ability to aggregate relevant features and model temporal dependencies effectively. In time series modeling, achieving a balance between model performance and computational efficiency remains a significant challenge. To address this, we propose FA-Stateformer, a state space and self-attention collaborative network with feature aggregation. The proposed network first employs a feature aggregation module to enhance informative features across both temporal and spectral dimensions. This is followed by a lightweight Conformer architecture inspired by the squeeze-and-excitation mechanism, where the feedforward layers are compressed to reduce redundancy and parameter overhead. Additionally, a temporal shift mechanism is incorporated to expand the receptive field of convolutional layers while maintaining a compact kernel size. To further enhance sequence modeling capabilities, a bidirectional Mamba module is introduced, enabling efficient state-space-based representation of temporal dependencies in both forward and backward directions. The remaining self-attention layers are combined with the Mamba blocks, forming a collaborative modeling framework that achieves a balance between representation capacity and computational efficiency. Extensive experiments demonstrate that FA-Stateformer achieves superior performance and efficiency compared to conventional architectures.

The joint transmit and pinching beamforming design for spectral efficiency (SE) and energy efficiency (EE) tradeoff in pinching-antenna systems (PASS) is proposed, under practical channel and energy consumption models. In the single-user scenario, it is proved that the optimal pinching antenna (PA) positions are independent of the transmit beamforming. Based on this insight, a two-stage joint beamforming design is proposed. Specifically, in the first stage, a general PA placement framework is proposed for multi-waveguide systems. In the second stage, the closed-form solution for the optimal transmit beamformer is derived given the optimized PA positions. In the multi-user scenario, an alternating optimization (AO)-based joint beamforming design is proposed to balance the SE-EE performance while taking the quality-of-service (QoS) requirements into account. It is proved that the proposed AO-based algorithm is guaranteed to converge when no constraints are violated in PA placement subproblem. Numerical results demonstrate that: 1) the proposed algorithms effectively improve joint SE-EE performance; 2) PASS exhibits strong robustness against variations in the service area along the waveguide direction.

Anomalous sound detection (ASD) in the wild requires robustness to distribution shifts such as unseen low-SNR input mixtures of machine and noise types. State-of-the-art systems extract embeddings from an adapted audio encoder and detect anomalies via nearest-neighbor search, but fine tuning on noisy machine sounds often acts like a denoising objective, suppressing noise and reducing generalization under mismatched mixtures or inconsistent labeling. Training-free systems with frozen self-supervised learning (SSL) encoders avoid this issue and show strong first-shot generalization, yet their performance drops when mixture embeddings deviate from clean-source embeddings. We propose to improve SSL backbones with a retain-not-denoise strategy that better preserves information from mixed sound sources. The approach combines a multi-label audio tagging loss with a mixture alignment loss that aligns student mixture embeddings to convex teacher embeddings of clean and noise inputs. Controlled experiments on stationary, non-stationary, and mismatched noise subsets demonstrate improved robustness under distribution shifts, narrowing the gap toward oracle mixture representations.

In the rapidly evolving research on artificial intelligence (AI) the demand for fast, computationally efficient, and scalable solutions has increased in recent years. The problem of optimizing the computing resources for distributed machine learning (ML) and optimization is considered in this paper. Given a set of data distributed over a network of computing-nodes/servers, the idea is to optimally assign the CPU (central processing unit) usage while simultaneously training each computing node locally via its own share of data. This formulates the problem as a co-optimization setup to (i) optimize the data processing and (ii) optimally allocate the computing resources. The information-sharing network among the nodes might be time-varying, but with balanced weights to ensure consensus-type convergence of the algorithm. The algorithm is all-time feasible, which implies that the computing resource-demand balance constraint holds at all iterations of the proposed solution. Moreover, the solution allows addressing possible log-scale quantization over the information-sharing channels to exchange log-quantized data. For some example applications, distributed support-vector-machine (SVM) and regression are considered as the ML training models. Results from perturbation theory, along with Lyapunov stability and eigen-spectrum analysis, are used to prove the convergence towards the optimal case. As compared to existing CPU scheduling solutions, the proposed algorithm improves the cost optimality gap by more than $50\%$.

We present a transformer-based multimodal framework for generating clinically relevant captions for MRI scans. Our system combines a DEiT-Small vision transformer as an image encoder, MediCareBERT for caption embedding, and a custom LSTM-based decoder. The architecture is designed to semantically align image and textual embeddings, using hybrid cosine-MSE loss and contrastive inference via vector similarity. We benchmark our method on the MultiCaRe dataset, comparing performance on filtered brain-only MRIs versus general MRI images against state-of-the-art medical image captioning methods including BLIP, R2GenGPT, and recent transformer-based approaches. Results show that focusing on domain-specific data improves caption accuracy and semantic alignment. Our work proposes a scalable, interpretable solution for automated medical image reporting.

In linear time-invariant systems, the sensitivity function to disturbances is designed under a sensitivity tradeoff known as the waterbed effect. To compensate for a quasiperiodic disturbance, a quasiperiodic disturbance observer using time delays was proposed. Its sensitivity function avoids the sensitivity tradeoff, achieving wideband harmonic suppression without amplifying aperiodic disturbances or shifting harmonic suppression frequencies. However, its open-loop transfer function is not rational and does not satisfy the assumptions of existing Bode sensitivity integrals due to its time delays. This paper provides Bode-like sensitivity integrals for the quasiperiodic disturbance observer in both continuous-time and discrete-time representations and clarifies the avoided sensitivity tradeoff with time delays.

With the rapid growth of artificial intelligence (AI) and cloud services, data centers have become critical infrastructures driving digital economies, with increasing energy demand heightening concerns over electricity use and carbon emissions, emphasizing the need for carbon-aware infrastructure planning. Most studies assume static power systems, focus only on operational emissions, and overlook co-optimization. This paper proposes a dynamic joint planning framework that co-optimizes long-term data center and power system development over 15 years. The model determines siting, capacity, and type of data centers alongside power generation expansion, storage deployment, and retirements, accounting for both operational and embodied emissions. To handle multi-scale uncertainty, a large-scale two-stage stochastic program is formulated and solved via an enhanced Benders decomposition. Applied to the PJM Interconnection, with curated datasets released on GitHub, results show the system can support up to 55 GW peak data center demand, with Virginia (DOM) and Northern Illinois (ComEd) as optimal hosts. Compared to non-joint planning, the framework cuts investment cost by 12.6%, operational cost by 8.25%, and emissions by 5.63%. Including lifecycle emissions further raises renewable deployment by 25.5%, highlighting embodied carbon's role in deeper decarbonization.

In this work, we introduce Neighborhood Feature Pooling (NFP), a novel pooling layer designed to enhance texture-aware representation learning for remote sensing image classification. The proposed NFP layer captures relationships between neighboring spatial features by aggregating local similarity patterns across feature dimensions. Implemented using standard convolutional operations, NFP can be seamlessly integrated into existing neural network architectures with minimal additional parameters. Extensive experiments across multiple benchmark datasets and backbone models demonstrate that NFP consistently improves classification performance compared to conventional pooling strategies, while maintaining computational efficiency. These results highlight the effectiveness of neighborhood-based feature aggregation for capturing discriminative texture information in remote sensing imagery.

Annotating time boundaries of sound events is labor-intensive, limiting the scalability of strongly supervised learning in audio detection. To reduce annotation costs, weakly-supervised learning with only clip-level labels has been widely adopted. As an alternative, partial label learning offers a cost-effective approach, where a set of possible labels is provided instead of exact weak annotations. However, partial label learning for audio analysis remains largely unexplored. Motivated by the observation that acoustic scenes provide contextual information for constructing a set of possible sound events, we utilize acoustic scene information to construct partial labels of sound events. On the basis of this idea, in this paper, we propose a multitask learning framework that jointly performs acoustic scene classification and sound event detection with partial labels of sound events. While reducing annotation costs, weakly-supervised and partial label learning often suffer from decreased detection performance due to lacking the precise event set and their temporal annotations. To better balance between annotation cost and detection performance, we also explore a semi-supervised framework that leverages both strong and partial labels. Moreover, to refine partial labels and achieve better model training, we propose a label refinement method based on self-distillation for the proposed approach with partial labels.

Controllers designed with reinforcement learning can be sensitive to model mismatch. We demonstrate that designing such controllers in a virtual simulation environment with an inaccurate model is not suitable for deployment in a physical setup. Controllers designed using an accurate model is robust against disturbance and small mismatch between the physical setup and the mathematical model derived from first principles; while a poor model results in a controller that performs well in simulation but fails in physical experiments. Sensitivity analysis is used to justify these discrepancies and an empirical region of attraction estimation help us visualize their robustness.

Advancements in spoken language processing have driven the development of spoken language models (SLMs), designed to achieve universal audio understanding by jointly learning text and audio representations for a wide range of tasks. Although promising results have been achieved, there is growing discussion regarding these models' generalization capabilities and the extent to which they truly integrate audio and text modalities in their internal representations. In this work, we evaluate four SLMs on the task of speech emotion recognition using a dataset of emotionally incongruent speech samples, a condition under which the semantic content of the spoken utterance conveys one emotion while speech expressiveness conveys another. Our results indicate that SLMs rely predominantly on textual semantics rather than speech emotion to perform the task, indicating that text-related representations largely dominate over acoustic representations. We release both the code and the Emotionally Incongruent Synthetic Speech dataset (EMIS) to the community.