Papers

Neural audio codecs (NACs) provide compact latent speech representations in the form of sequences of continuous vectors or discrete tokens. In this work, we investigate how these two types of speech representations compare when used as training targets for supervised speech enhancement. We consider both autoregressive and non-autoregressive speech enhancement models based on the Conformer architecture, as well as a simple baseline where the NAC encoder is simply fine-tuned for speech enhancement. Our experiments reveal three key findings: predicting continuous latent representations consistently outperforms discrete token prediction; autoregressive models achieve higher quality but at the expense of intelligibility and efficiency, making non-autoregressive models more attractive in practice; and adding encoder fine-tuning yields the strongest enhancement metrics overall, though at the cost of degraded codec reconstruction. The code and audio samples are available online.

Driven by the growing demand for higher spectral efficiency in wireless communications, intelligent reflecting surfaces (IRS) have attracted considerable attention for their ability to dynamically reconfigure the propagation environment. This work addresses the spectral efficiency maximization problem in IRS-assisted multiple-input multiple-output (MIMO) systems, which involves the joint optimization of the transmit precoding matrix and the IRS phase shift configuration. This problem is inherently challenging due to its non-convex nature. To tackle it effectively, we introduce a computationally efficient algorithm, termed ADMM-APG, which integrates the alternating direction method of multipliers (ADMM) with the accelerated projected gradient (APG) method. The proposed framework decomposes the original problem into tractable subproblems, each admitting a closed-form solution while maintaining low computational complexity. Simulation results demonstrate that the ADMM-APG algorithm consistently surpasses existing benchmark methods in terms of spectral efficiency and computational complexity, achieving significant performance gains across a range of system configurations.

This paper deals with the optimal synthesis of aperture fields for (radiating) near-field communications in obstructed environments. A physically consistent model based on knife-edge diffraction is used to formulate the problem as a maximization in Hilbert space. The optimal solution is obtained as a matched filter that “matches" the shape of a diffraction-induced kernel, thus linking wave propagation with signal processing methods. The framework supports hardware implementation using continuous apertures such as metasurfaces or lens antennas. This approach bridges physically grounded modeling, signal processing, and hardware design for efficient energy focusing in near-field obstructed channels.

This study analyzes the performance of positioning techniques based on configuration changes of 5G New Radio signals. In 5G networks, a terminal position is determined from the Time of Arrival of Positioning Reference Signals transmitted by base stations. We propose an algorithm that improves TOA accuracy under low sampling rate constraints and implement 5G PRS for positioning in a software defined modem. We also examine how flexible time frequency resource allocation of PRS affects TOA estimation accuracy and discuss optimal PRS configurations for a given signal environment.

Most existing semantic communication systems based on joint source-channel coding (JSCC) employ analog modulation and are thus inherently incompatible with modern digital communication systems and impose stringent hardware design challenges. Although several digital transmission approaches have been proposed to address this issue, they often suffer from high sensitivity to bit errors, limited adaptability to varying source distributions, or re-training overhead under different modulation schemes. This letter proposes BitSemCom, a novel end-to-end bit-level JSCC framework that is robust to channel noise and modulation-agnostic. The core component is a learnable bit mapper that establishes a probabilistic mapping between continuous semantic features and discrete bit sequences. By leveraging a sampling-based bit generation method based on the Gumbel-Softmax trick, the framework enables differentiable bit-level optimization while maintaining robustness to channel errors. Simulation results on image transmission demonstrate that BitSemCom achieves consistent peak signal-to-noise ratio (PSNR) gains of 2-3 dB over codebook-based digital semantic transmission methods and competitive performance with stronger robustness compared to separate source-channel coding (SSCC) benchmarks. Ablation studies further validate the effectiveness of the learnable bit mapper.

In this work we extend the results developed in 2022 for a sequential change detection algorithm making use of Page's CUSUM statistic, the empirical distribution as an estimate of the pre-change distribution, and a universal code as a tool for estimating the post-change distribution, from the i.i.d. case to the Markov setup.

The evaluation of intelligibility for TTS has reached a bottleneck, as existing assessments heavily rely on word-by-word accuracy metrics such as WER, which fail to capture the complexity of real-world speech or reflect human comprehension needs. To address this, we propose Spoken-Passage Multiple-Choice Question Answering, a novel subjective approach evaluating the accuracy of key information in synthesized speech, and release SP-MCQA-Eval, an 8.76-hour news-style benchmark dataset for SP-MCQA evaluation. Our experiments reveal that low WER does not necessarily guarantee high key-information accuracy, exposing a gap between traditional metrics and practical intelligibility. SP-MCQA shows that even state-of-the-art (SOTA) models still lack robust text normalization and phonetic accuracy. This work underscores the urgent need for high-level, more life-like evaluation criteria now that many systems already excel at WER yet may fall short on real-world intelligibility.

Edge computing alleviates the computation burden of data-driven control in cyber-physical systems (CPSs) by offloading complex processing to edge servers. However, the increasing sophistication of cyberattacks underscores the need for security measures that go beyond conventional IT protections and address the unique vulnerabilities of CPSs. This study proposes a confidential data-driven gain-tuning framework using homomorphic encryption, such as ElGamal and CKKS encryption schemes, to enhance cybersecurity in gain-tuning processes outsourced to external servers. The idea for realizing confidential FRIT is to replace the matrix inversion operation with a vector summation form, allowing homomorphic operations to be applied. Numerical examples under 128-bit security confirm performance comparable to conventional methods while providing guidelines for selecting suitable encryption schemes for secure CPS.

This paper investigates the construction of channel knowledge map (CKM) from sparse channel measurements. Dif ferent from conventional two-/three-dimensional (2D/3D) CKM approaches assuming fixed base station configurations, we present a six-dimensional (6D) CKM framework named bidirectional wireless Gaussian splatting (BiWGS), which is capable of mod eling wireless channels across dynamic transmitter (Tx) and receiver (Rx) positions in 3D space. BiWGS uses Gaussian el lipsoids to represent virtual scatterer clusters and environmental obstacles in the wireless environment. By properly learning the bidirectional scattering patterns and complex attenuation profiles based on channel measurements, these ellipsoids inherently cap ture the electromagnetic transmission characteristics of wireless environments, thereby accurately modeling signal transmission under varying transceiver configurations. Experiment results show that BiWGS significantly outperforms classic multi-layer perception (MLP) for the construction of 6D channel power gain map with varying Tx-Rx positions, and achieves spatial spectrum prediction accuracy comparable to the state-of-the art wireless radiation field Gaussian splatting (WRF-GS) for 3D CKM construction. This validates the capability of the proposed BiWGS in accomplishing dimensional expansion of 6D CKM construction, without compromising fidelity.

This paper investigates a novel computation and communication co-design framework for large-scale split learning in intelligent reflecting surface (IRS)-assisted internet of things (IoT) networks integrated with digital twin (DT) technique. The considered system consists of a multi-antenna access point (AP), multiple heterogeneous user devices (UDs), and an deployed IRS to enhance both uplink and downlink transmission. The training process of a deep neural network is partitioned between devices and the AP, where a DT replica is activated to replace UDs with insufficient local computation capabilities. We formulate a delay-optimal split learning problem, which optimizes five key variables: layer partitioning points, DT assignment decisions, IRS phase shift matrix, AP downlink power allocation, and DT frequency adjustment, aiming to minimize the overall end-to-end delay under communication and computation. The proposed optimization problem is a highly coupled non-convex mixed-integer problem. Therefore, we solve using an alternating optimization approach combining closed-form updates, semidefinite relaxation (SDR), and low-complexity heuristics. Extensive simulations demonstrate that the proposed scheme significantly reduces training delay compared to conventional baselines and achieves up to 35% delay improvement, especially under high UD density and stringent power constraints.

In this paper, we investigate the beamforming design problem in an integrated sensing and communication (ISAC) system, where a multi-antenna base station simultaneously serves multiple communication users while performing radar sensing. We formulate the problem as the minimization of the total transmit power, subject to signal-to-interference-plus-noise ratio (SINR) constraints for communication users and mean-squared-error (MSE) constraints for radar sensing. The core challenge arises from the complex coupling between communication SINR requirements and sensing performance metrics. To efficiently address this challenge, we first establish the equivalence between the original ISAC beamforming problem and its semidefinite relaxation (SDR), derive its Lagrangian dual formulation, and further reformulate it as a generalized downlink beamforming (GDB) problem with potentially indefinite weighting matrices. Compared to the classical DB problem, the presence of indefinite weighting matrices in the GDB problem introduces substantial analytical and computational challenges. Our key technical contributions include (i) a necessary and sufficient condition for the boundedness of the GDB problem, and (ii) a tailored efficient fixed point iteration (FPI) algorithm with a provable convergence guarantee for solving the GDB problem. Building upon these results, we develop a duality-based fixed point iteration (Dual-FPI) algorithm, which integrates an outer subgradient ascent loop with an inner FPI loop. Simulation results demonstrate that the proposed Dual-FPI algorithm achieves globally optimal solutions while significantly reducing computational complexity compared with existing baseline approaches.

The imminent emergence of sixth-generation (6G) networks faces critical challenges from spatially heterogeneous traffic and escalating energy consumption, necessitating sustainable scaling strategies for network infrastructure such as base stations (BSs) and reconfigurable intelligent surfaces (RISs). This paper presents a systematic scaling analysis of the Integrated Relative Energy Efficiency (IREE) metric under joint multi-BS and multi-RIS deployment in traffic-mismatched scenarios. Specifically, we propose an Alternating Directional Dual Radial Basis Function (ADD-RBF) framework that models the spatial capacity contributions of BSs and RISs through two separately parameterized RBF-type branches and maximizes IREE through accepted alternating optimization, with established representation expressiveness and stage-wise convergence properties. Theoretical analysis reveals distinct scaling behaviors: BS proliferation drives logarithmic capacity growth $\mathcal{O}(\log N^{BS})$ and polynomial large-scale mismatch reduction $\mathcal{O}\big((N^{BS})^{-t_g/2}\big)$, whereas RIS deployment provides a bounded passive capacity-gain correction and residual-structure-dependent mismatch reduction. Specifically, the RIS-side mismatch decreases polynomially as $\mathcal{O}\big((N^R)^{-t_\ell/2}\big)$ for spatially diffuse residuals and follows the stretched-exponential order $\mathcal{O}\big(\exp[-c_R\sqrt{N^R}]\big)$ for hotspot-dominated residuals. Simulation results show that RISs are effective in refining spatial traffic-capacity mismatch and alleviating hotspots, making them particularly attractive when mismatch dominates, while BSs are generally preferable under capacity shortages. These findings offer practical guidelines for green 6G network design.

Advances in data analysis and machine learning have revolutionized the study of brain signatures using fMRI, enabling non-invasive exploration of cognition and behavior through individual neural patterns. Functional connectivity (FC), which quantifies statistical relationships between brain regions, has emerged as a key metric for studying individual variability and developing biomarkers for personalized medicine in neurological and psychiatric disorders. The concept of subject fingerprinting, introduced by Finn et al. (2015), leverages neural connectivity variability to identify individuals based on their unique patterns. While traditional FC methods perform well on small datasets, machine learning techniques are more effective with larger datasets, isolating individual-specific features and maximizing inter-subject differences. In this study, we propose a framework combining convolutional autoencoders and sparse dictionary learning to enhance fingerprint accuracy. Autoencoders capture shared connectivity patterns while isolating subject-specific features in residual FC matrices, which are analyzed using sparse coding to identify distinctive features. Tested on the Human Connectome Project dataset, this approach achieved a 10% improvement over baseline group-averaged FC models. Our results highlight the potential of integrating deep learning and sparse coding techniques for scalable and robust functional connectome fingerprinting, advancing personalized neuroscience applications and biomarker development.

Accurate channel state information (CSI) is essential for reliable multiuser MIMO operation. In 5G NR, reciprocity-based beamforming via uplink Sounding Reference Signals (SRS) face resource and coverage constraints, motivating sparse non-uniform SRS allocation. Prior masked-autoencoder (MAE) approaches improve coverage but overfit to training masks and degrade under unseen distortions (e.g., additional masking, interference, clipping, non-Gaussian noise). We propose a diffusion-based channel inpainting framework that integrates system-model knowledge at inference via a likelihood-gradient term, enabling a single trained model to adapt across mismatched conditions. On standardized CDL channels, the score-based diffusion variant consistently outperforms a UNet score-model baseline and the one-step MAE under distribution shift, with improvements up to 14 dB NMSE in challenging settings (e.g., Laplace noise, user interference), while retaining competitive accuracy under matched conditions. These results demonstrate that diffusion-guided inpainting is a robust and generalizable approach for super-capacity SRS design in 5G NR systems.

Lightweight online detection of series arc faults is critically needed in residential and industrial power systems to prevent electrical fires. Existing diagnostic methods struggle to achieve both rapid response and robust accuracy under resource-constrained conditions. To overcome the challenge, this work suggests leveraging a multi-frequency neural network named MFNN, embedding prior physical knowledge into the network. Inspired by arcing current curve and the Fourier decomposition analysis, we create an adaptive activation function with super-expressiveness, termed EAS, and a novel network architecture with branch networks to help MFNN extract features with multiple frequencies. In our experiments, eight advanced arc fault diagnosis models across an experimental dataset with multiple sampling times and multi-level noise are used to demonstrate the superiority of MFNN. The corresponding experiments show: 1) The MFNN outperforms other models in arc fault location, befitting from signal decomposition of branch networks. 2) The noise immunity of MFNN is much better than that of other models, achieving 14.51% over LCNN and 16.3% over BLS in test accuracy when SNR=-9. 3) EAS and the network architecture contribute to the excellent performance of MFNN.

This paper presents a novel approach to stochastic economic model predictive control (SEMPC) that minimizes average economic cost while satisfying an empirical expected shortfall (EES) constraint to manage risk. A new scenario-based problem formulation ensuring controlled risk with high confidence while minimizing the average cost is introduced. The probabilistic guarantees is dependent on the number of support elements over the entire input domain, which is difficult to find for high-dimensional systems. A heuristic algorithm is proposed to find the number of support elements. Finally, an efficient method is presented to reduce the computational complexity of the SEMPC problem with an EES constraint. The approach is validated on a water distribution network, showing its effectiveness in balancing performance and risk.

Electric vehicle charging and geo-distributed datacenters introduce spatially flexible loads (FLs) that couple power, transportation, and datacenter networks. These couplings create a closed-loop feedback between locational marginal prices (LMPs) and decisions of the FL systems, challenging the foundations of conventional competitive equilibrium (CE) in electricity markets. This paper studies a notion of generalized competitive equilibrium (GCE) that aims to capture such price-demand interactions across the interconnected infrastructures. We establish structural conditions under which the GCE preserves key properties of the conventional CE, including existence, uniqueness, and efficiency, without requiring detailed knowledge of decision processes within individual FL systems. The framework generalizes to settings where the grid is coupled with multiple FL systems. Stylized examples and case studies on the New York ISO grid, coupled with the Sioux Falls transportation and distributed datacenter networks, demonstrate the use of our theoretical framework and illustrate the mutual influence among the grid and the studied FL systems.

Depression is a major cause of global mental illness and significantly influences suicide rates. Timely and accurate diagnosis is essential for effective intervention. Electroencephalography (EEG) provides a non-invasive and accessible method for examining cerebral activity and identifying disease-associated patterns. We propose a novel graph-based deep learning framework, named Edge-gated, axis-mixed Pooling Attention Network (ExPANet), for differentiating major depressive disorder (MDD) patients from healthy controls (HC). EEG recordings undergo preprocessing to eliminate artifacts and are segmented into short periods of activity. We extract 14 features from each segment, which include time, frequency, fractal, and complexity domains. Electrodes are represented as nodes, whereas edges are determined by the phase-locking value (PLV) to represent functional connectivity. The generated brain graphs are examined utilizing an adapted graph attention network. This architecture acquires both localized electrode characteristics and comprehensive functional connectivity patterns. The proposed framework attains superior performance relative to current EEG-based approaches across two different datasets. A fundamental advantage of our methodology is its explainability. We evaluated the significance of features, channels, and edges, in addition to intrinsic attention weights. These studies highlight features, cerebral areas, and connectivity associations that are especially relevant to MDD, many of which correspond with clinical data. Our findings demonstrate a reliable and transparent method for EEG-based screening of MDD, using deep learning with clinically relevant results.

Earth s gravity fundamentally shapes human behaviour. The brain encodes this force as an internal model of gravity, enabling the prediction and interpretation of gravitational effects during perception and action. Understanding how this model adapts to altered gravity is critical for predicting human performance in spaceflight. We present a computational framework for modelling neurophysiological adaptation across diverse gravitational environments. The framework has two components trained on open-access data from altered-gravity studies, particularly parabolic flights. The first component (CorticalG) employs a lightweight multilayer perceptron neural network to predict gravity-dependent changes in EEG frequency bands, estimating cortical state under different gravitational loads. The second component (PhysioG) uses independent Gaussian process models to capture broader physiological responses, including heart rate variability, electrodermal activity, and motor control. To complement the quantitative modelling, we simulated subjective experience across gravitational environments using the Large Language Model (LLM) Claude 3.5 Sonnet. Physiological outputs prompted the model to generate narratives describing alertness, bodily awareness, and cognitive state across zero gravity, partial gravity of the Moon and Mars, and hypergravity. This framework provides a novel approach for investigating human adaptation to spaceflight. It offers a predictive tool to assess performance and resilience, supporting the design of future space exploration missions.

Brain disorders are an umbrella term for a group of neurological and psychiatric conditions that have a major effect on thinking, feeling, and acting. These conditions encompass a wide range of conditions. The illnesses in question pose significant difficulties not only for individuals, but also for healthcare systems all across the world. In this study, we explore the capability of explainable machine learning for classification of people who suffer from brain disorders. This is accomplished by the utilization of brain connection map, also referred as connectome, derived from functional magnetic resonance imaging (fMRI) data. In order to analyze features that are based on the connectome, we investigated several different feature selection procedures. These strategies included the Least Absolute Shrinkage and Selection Operator (LASSO), Relief, and Analysis of Variance (ANOVA), in addition to a logistic regression (LR) classifier. First and foremost, the purpose was to evaluate and contrast the classification accuracy of different feature selection methods in terms of distinguishing healthy controls from diseased individuals. The evaluation of the stability of the traits that were chosen was the second objective. The identification of the regions of the brain that have an effect on the classification was the third main objective. When applied to the UCLA dataset, the LASSO approach, which is our most effective strategy, produced a classification accuracy of 91.85% and a stability index of 0.74, which is greater than the results obtained by other approaches: Relief and ANOVA. These methods are effective in locating trustworthy biomarkers, which adds to the development of connectome-based classification in the context of issues that impact the brain.

Stability analysis of the Kalman filter under randomly lost measurements has been widely studied. We revisit this problem in a general continuous-time framework, where both the measurement matrix and noise covariance evolve as random processes, capturing variability in sensing locations. Within this setting, we derive a closed-form upper bound on the expected estimation covariance for continuous-time Kalman filtering. We then apply this framework to spatiotemporal field estimation, where the field is modeled as a Gaussian process observed by randomly located, noisy sensors. Using clarity, introduced in our earlier work as a rescaled form of the differential entropy of a random variable, we establish a lower bound on the expected spatially averaged clarity and show that its steady-state value converges to a grid-independent limit as the grid is refined. This result exposes fundamental performance limits through a composite sensing parameter that jointly captures the effects of the number of sensors, noise level, and measurement frequency. Simulations show close agreement between the continuous-time analytical solution and the discrete-time covariance, with the agreement improving as the sampling interval decreases. The analytical solution also avoids the recursive computations required in the discrete-time formulation. Most importantly, the derived limits provide principled and efficient guidelines for sensor network design prior to deployment.

Deploying advanced cardiac artificial intelligence for daily cardiac monitoring is hindered by its reliance on extensive medical data and high computational resources. Low-cost cardiac intelligence (LCCI) offers a promising alternative by using wearable device data, such as 1-lead electrocardiogram (ECG), but it suffers from a significant diagnostic performance gap compared to high-cost cardiac intelligence (HCCI). To bridge this gap, we propose LiteHeart, a semi-supervised knowledge distillation framework. LiteHeart introduces a region-aware distillation module to mimic how cardiologists focus on diagnostically relevant ECG regions and a cross-layer mutual information module to align the decision processes of LCCI and HCCI systems. Using a semi-supervised training strategy, LiteHeart further improves model robustness under limited supervision. Evaluated on five datasets covering over 38 cardiovascular diseases, LiteHeart substantially reduces the performance gap between LCCI and HCCI, outperforming existing methods by 4.27% to 7.10% in macro F1 score. These results demonstrate that LiteHeart significantly enhances the diagnostic capabilities of low-cost cardiac intelligence systems, paving the way for scalable, affordable, and accurate daily cardiac healthcare using wearable technologies.

Parity is exactly determined by the least significant bit (LSB), so machine learning is unnecessary for solving the task. This work instead uses parity as a controlled diagnostic for studying how a classical signal-processing pipeline preserves, amplifies, or suppresses access to symbolic information embedded in a numerical representation. Integers from 0 to 10,000 are encoded as fixed-width 32-bit binary signals, decomposed with a level-3 Daubechies-2 (db2) discrete wavelet transform, summarized by mean absolute coefficient magnitude, and clustered independently per subband with k-means. Clustering is unsupervised, while cluster-to-parity calibration uses training labels only under a stratified 60/20/20 train/validation/test protocol. The frozen configuration reaches 84.26% accuracy on the held-out test set (95% Wilson CI: 82.60-85.79%), and 84.20% +/- 0.57% across 20 stratified resplits. Raw binary summary statistics reach 61.30%, while masking the LSB reduces validation accuracy to 48.15%, consistent with chance. The A3 approximation band alone retains 83.20%, whereas detail bands remain near chance. Performance is strongly representation-dependent: moving the parity bit changes validation accuracy up to 98.60%, and changing the wavelet boundary mode ranges from 54.45% to 83.20%. Cross-magnitude extrapolation degrades from 79.98% on 10,001-20,000 to 59.69% on 100,001-1,000,000. These results do not show that wavelets discover the arithmetic rule of parity. Rather, they show that information already present in the binary encoding becomes more or less recoverable depending on multiscale filtering, spatial alignment, and boundary handling.

Automated monitoring of marine mammals in the St. Lawrence Estuary faces extreme challenges: calls span low-frequency moans to ultrasonic clicks, often overlap, and are embedded in variable anthropogenic and environmental noise. We introduce a multi-step, attention-guided framework that first segments spectrograms to generate soft masks of biologically relevant energy and then fuses these masks with the raw inputs for multi-band, denoised classification. Image and mask embeddings are integrated via mid-level fusion, enabling the model to focus on salient spectrogram regions while preserving global context. Using real-world recordings from the Saguenay St. Lawrence Marine Park Research Station in Canada, we demonstrate that segmentation-driven attention and mid-level fusion improve signal discrimination, reduce false positive detections, and produce reliable representations for operational marine mammal monitoring across diverse environmental conditions and signal-to-noise ratios. Beyond in-distribution evaluation, we further assess the generalization of Mask-Guided Classification (MGC) under distributional shifts by testing on spectrograms generated with alternative acoustic transformations. While high-capacity baseline models lose accuracy in this Out-of-distribution (OOD) setting, MGC maintains stable performance, with even simple fusion mechanisms (gated, concat) achieving comparable results across distributions. This robustness highlights the capacity of MGC to learn transferable representations rather than overfitting to a specific transformation, thereby reinforcing its suitability for large-scale, real-world biodiversity monitoring. We show that in all experimental settings, the MGC framework consistently outperforms baseline architectures, yielding substantial gains in accuracy on both in-distribution and OOD data.

This work introduces MedForj, a suite of 3D foundational generative priors based on diffusion models. The MedForj models were trained on $72{,}659$ 1~mm isotropic 3D $T_1$-weighted MRI human brain image volumes from $38{,}174$ subjects, drawn from a curated corpus of $80{,}675$ volumes from $42{,}506$ subjects spanning $38$ publicly available datasets. These training images were manually inspected to exclude those with poor quality and excessive pathology, and otherwise were minimally processed. The models include six different diffusion training strategies: rectified flow, latent diffusion rectified flow, flow matching, velocity prediction, clean prediction, and noise prediction. Image samples produced by each of these models were compared to each other and against real, ground truth data under downstream segmentation distributions, FID, five inverse problems, and blind human inspection in an observer study. Flow matching was the strongest strategy overall, achieving the best inverse problem solving results at $28.80$~dB PSNR and $0.874$ SSIM averaged over the five forward problems, the highest rate of reconstructions judged real by blind human raters at $72.6\%$, and the closest per-structure match to real segmented anatomy in a permutation test. It was not best everywhere: rectified flow produced the most convincing unconditional samples in the observer study and the best FID, and the latent rectified-flow model achieved the smallest joint distributional distance to real anatomy. No other strategy, however, performed consistently well across all four evaluations. We therefore recommend flow matching as the default MedForj prior, while releasing every strategy so that the choice can be revisited per application. All model weights and corresponding code are publicly available at https://github.com/piksl-research/medforj.

Medical image datasets frequently exhibit significant class imbalance, a challenge that is further amplified by the inherently limited sample sizes that characterize clinical imaging data. Using human embryo time-lapse imaging (TLI) as a case study, this work investigates how generative adversarial networks (GANs) can be optimized for small datasets to generate realistic and diagnostically meaningful images. Based on systematic experiments with R3GAN, we established effective training strategies and designed an optimized configuration for 256x256-resolution datasets, featuring a full burn-in phase and a low, gradually increasing gamma range (5 to 40). The generated samples were used to balance an imbalanced embryo dataset, leading to substantial improvement in classification performance. The recall and F1-score of the three-cell (t3) class increased from 0.06 to 0.69 and from 0.11 to 0.60, respectively, without compromising the performance of other classes. These results demonstrate that tailored R3GAN training strategies can effectively alleviate data scarcity and improve model robustness in small-scale medical imaging tasks.

Medical image segmentation models face severe performance drops under domain shifts, especially when data sharing constraints prevent access to source images. We present a novel Uncertainty-aware Progressive Pseudo-label Denoising (UP2D) framework for source-free domain adaptation (SFDA), designed to mitigate noisy pseudo-labels and class imbalance during adaptation. UP2D integrates three key components: (i) a Refined Prototype Filtering module that suppresses uninformative regions and constructs reliable class prototypes to denoise pseudo-labels, (ii) an Uncertainty-Guided EMA (UG-EMA) strategy that selectively updates the teacher model based on spatially weighted boundary uncertainty, and (iii) a quantile-based entropy minimization scheme that focuses learning on ambiguous regions while avoiding overconfidence on easy pixels. This single-stage student-teacher framework progressively improves pseudo-label quality and reduces confirmation bias. Extensive experiments on three challenging retinal fundus benchmarks demonstrate that UP2D achieves state-of-the-art performance across both standard and open-domain settings, outperforming prior UDA and SFDA approaches while maintaining superior boundary precision.

Audio-visual speech enhancement (AVSE) is a task that uses visual auxiliary information to extract a target speaker's speech from mixed audio. In real-world scenarios, there often exist complex acoustic environments, accompanied by various interfering sounds and reverberation. Most previous methods struggle to cope with such complex conditions, resulting in poor perceptual quality of the extracted speech. In this paper, we propose an effective AVSE system that performs well in complex acoustic environments. Specifically, we design a "separation before dereverberation" pipeline that can be extended to other AVSE networks. The 4th COGMHEAR Audio-Visual Speech Enhancement Challenge (AVSEC) aims to explore new approaches to speech processing in multimodal complex environments. We validated the performance of our system in AVSEC-4: we achieved excellent results in the three objective metrics on the competition leaderboard, and ultimately secured first place in the human subjective listening test.

Contemporary macro energy systems modelling is characterized by the need to represent strategic and operational decisions with high temporal and spatial resolution and represent discrete investment and retirement decisions. This drive towards greater fidelity, however, conflicts with a simultaneous push towards greater model representation of inherent complexity in decision making, including methods like Modelling to Generate Alternatives (MGA). MGA aims to map the feasible space of a model within a cost slack by varying investment parameters without changing the operational constraints, a process which frequently requires hundreds of solutions. For large, detailed energy system models this is impossible with traditional methods, leading researchers to reduce complexity with linearized investments and zonal or temporal aggregation. This research presents a new solution method for MGA type problems using cutting-plane methods based on a tailored reformulation of Benders Decomposition. We accelerate the algorithm by sharing cuts between MGA master problems and grouping MGA objectives. We find that our new solution method consistently solves MGA problems times faster and requires less memory than existing monolithic Modelling to Generate Alternatives solution methods on linear problems, enabling rapid computation of a greater number of solutions to highly resolved models. We also show that our novel cutting-plane algorithm enables the solution of very large MGA problems with integer investment decisions.

Multiparametric mapping MRI has become a viable tool for myocardial tissue characterization. However, misalignment between multiparametric maps makes pixel-wise analysis challenging. To address this challenge, we developed a generalizable physics-informed deep-learning model using test-time adaptation to enable group image registration across contrast weighted images acquired from multiple physical models (e.g., a T1 mapping model and T2 mapping model). The physics-informed adaptation utilized the synthetic images from specific physics model as registration reference, allows for transductive learning for various tissue contrast. We validated the model in healthy volunteers with various MRI sequences, demonstrating its improvement for multi-modal registration with a wide range of image contrast variability.