Papers

Non-reference metrics (NRMs) can assess the visual quality of images and videos without a reference, making them well-suited for the evaluation of user-generated content. Nonetheless, rate-distortion optimization (RDO) in video coding is still mainly driven by full-reference metrics, such as the sum of squared errors, which treat the input as an ideal target. A way to incorporate NRMs into RDO is through linearization (LNRM), where the gradient of the NRM with respect to the input guides bit allocation. While this strategy improves the quality predicted by some metrics, we show that it can yield limited gains or degradations when evaluated with other NRMs. We argue that NRMs are highly non-linear predictors with locally unstable gradients that can compromise the quality of the linearization; furthermore, optimizing a single metric may exploit model-specific biases that do not generalize across quality estimators. Motivated by this observation, we extend the LNRM framework to optimize ensembles of NRMs and, to further improve robustness, we introduce a smoothing-based formulation that stabilizes NRM gradients prior to linearization. Our framework is well-suited to hybrid codecs, and we advocate for its use with overfitted codecs, where it avoids iterative evaluations and backpropagation of neural network-based NRMs, reducing encoder complexity relative to direct NRM optimization. We validate the proposed approach on AVC and Cool-chic, using the YouTube UGC dataset. Experiments demonstrate consistent bitrate savings across multiple NRMs with no decoder complexity overhead and, for Cool-chic, a substantial reduction in encoding runtime compared to direct NRM optimization.

Neural autoencoders underpin generative models. Practical, large-scale use of neural autoencoders for generative modeling necessitates fast encoding, low latent rates, and a single model across representations. Existing approaches are reconstruction-first: they incur high latent rates, slow encoding, and separate architectures for discrete vs. continuous latents and for different audio channel formats, hindering workflows from preprocessing to inference conditioning. We introduce a generative-first architecture for audio autoencoding that increases temporal downsampling from 2048x to 3360x and supports continuous and discrete representations and common audio channel formats in one model. By balancing compression, quality, and speed, it delivers 10x faster encoding, 1.6x lower rates, and eliminates channel-format-specific variants while maintaining competitive reconstruction quality. This enables applications previously constrained by processing costs: a 60-second mono signal compresses to 788 tokens, making generative modeling more tractable.

Integrating artificial intelligence (AI) into wireless channel modeling requires large, accurate, and physically consistent datasets derived from real measurements. Such datasets are essential for training and validating models that learn spatio-temporal channel behavior across frequencies and environments. NYUSIM, introduced by NYU WIRELESS in 2016, generates realistic spatio-temporal channel data using extensive outdoor and indoor measurements between 28 and 142 GHz. To improve scalability and support 6G research, we migrated the complete NYUSIM framework from MATLAB to Python, and are incorporating new statistical model generation capabilities from extensive field measurements in the new 6G upper mid-band spectrum at 6.75 GHz (FR1(C)) and 16.95 GHz (FR3) [1]. The NYUSIM Python also incorporates a 3D antenna data format, referred to as Ant3D, which is a standardized, full-sphere format for defining canonical, commercial, or measured antenna patterns for any statistical or site-specific ray tracing modeling tool. Migration from MATLAB to Python was rigorously validated through Kolmogorov-Smirnov (K-S) tests, moment analysis, and end-to-end testing with unified randomness control, confirming statistical consistency and reproduction of spatio-temporal channel statistics, including spatial consistency with the open-source MATLAB NYUSIM v4.0 implementation. The NYUSIM Python version is designed to integrate with modern AI workflows and enable large-scale parallel data generation, establishing a robust, verified, and extensible foundation for future AI-enabled channel modeling.

Visual analogy learning enables image editing via demonstration rather than textual description, allowing users to specify complex transformations difficult to articulate in words. Given a triplet $\{\mathbf{a}$, $\mathbf{a}'$, $\mathbf{b}\}$, the goal is to generate $\mathbf{b}'$ such that $\mathbf{a} : \mathbf{a}' :: \mathbf{b} : \mathbf{b}'$. Recent methods adapt text-to-image models with a single Low-Rank Adaptation (LoRA) module, but they face a fundamental limitation: attempting to capture the diverse space of visual transformations within a fixed module constrains generalization. Inspired by recent work showing that LoRAs in constrained domains span meaningful, interpolatable semantic spaces, we propose LoRWeB, which specializes the model for each analogy task in a single inference pass. LoRWeB dynamically composes learned transformation primitives, informally, choosing a point in a "space of LoRAs". We introduce two key components: (1) a learnable basis of LoRAs to span the space of different visual transformations, and (2) a lightweight encoder that dynamically weighs these basis LoRAs given the input analogy pair. Comprehensive evaluations demonstrate state-of-the-art performance and significantly improved generalization to unseen transformations. Our findings suggest LoRA basis decompositions are a promising direction for flexible visual manipulation tasks. See https://research.nvidia.com/labs/par/lorweb for code.

Node embeddings map graph vertices into low-dimensional Euclidean spaces while preserving structural information. They are central to tasks such as node classification, link prediction, and signal reconstruction. A key goal is to design node embeddings whose dot products capture meaningful notions of node similarity induced by the graph. Graph kernels offer a principled way to define such similarities, but their direct computation is often prohibitive for large networks. Inspired by random feature methods for kernel approximation in Euclidean spaces, we introduce randomized spectral node embeddings whose dot products estimate a low-rank approximation of any specific graph kernel. We provide theoretical and empirical results showing that our embeddings achieve more accurate kernel approximations than existing methods, particularly for spectrally localized kernels. These results demonstrate the effectiveness of randomized spectral constructions for scalable and principled graph representation learning.

Learning dynamical systems through purely data-driven methods is challenging as they do not learn the underlying conservation laws that enable them to correctly generalize. Existing port-Hamiltonian neural network methods have recently been successfully applied for modeling mechanical systems. However, even though these methods are designed on power-balance principles, they usually do not consider power-preserving discretizations and often rely on Runge-Kutta numerical methods. In this work, we propose to use a second-order discrete gradient method embedded in the learning of dynamical systems with port-Hamiltonian neural networks. Numerical results are provided for three systems deliberately selected to span different ranges of dynamical behavior under control: a baseline harmonic oscillator with quadratic energy storage; a Duffing oscillator, with a non-quadratic Hamiltonian offering amplitude-dependent effects; and a self-sustained oscillator, which can stabilize in a controlled limit cycle through the incorporation of a nonlinear dissipation. We show how the use of this discrete gradient method outperforms the performance of a Runge-Kutta method of the same order. Experiments are also carried out to compare two theoretically equivalent port-Hamiltonian systems formulations and to analyze the impact of regularizing the Jacobian of port-Hamiltonian neural networks during training.

Assessing human muscle fatigue is critical for optimizing performance and safety in physical human-robot interaction(pHRI). This work presents a data-driven framework to estimate fatigue in dynamic, cyclic pHRI using arm-mounted surface electromyography(sEMG). Subject-specific machine-learning regression models(Random Forest, XGBoost, and Linear Regression predict the fraction of cycles to fatigue(FCF) from three frequency-domain and one time-domain EMG features, and are benchmarked against a convolutional neural network(CNN) that ingests spectrograms of filtered EMG. Framing fatigue estimation as regression (rather than classification) captures continuous progression toward fatigue, supporting earlier detection, timely intervention, and adaptive robot control. In experiments with ten participants, a collaborative robot under admittance control guided repetitive lateral (left-right) end-effector motions until muscular fatigue. Average FCF RMSE across participants was 20.8+/-4.3% for the CNN, 23.3+/-3.8% for Random Forest, 24.8+/-4.5% for XGBoost, and 26.9+/-6.1% for Linear Regression. To probe cross-task generalization, one participant additionally performed unseen vertical (up-down) and circular repetitions; models trained only on lateral data were tested directly and largely retained accuracy, indicating robustness to changes in movement direction, arm kinematics, and muscle recruitment, while Linear Regression deteriorated. Overall, the study shows that both feature-based ML and spectrogram-based DL can estimate remaining work capacity during repetitive pHRI, with the CNN delivering the lowest error and the tree-based models close behind. The reported transfer to new motion patterns suggests potential for practical fatigue monitoring without retraining for every task, improving operator protection and enabling fatigue-aware shared autonomy, for safer fatigue-adaptive pHRI control.

This work considers merging two independent models, TTS and A2F, into a unified model to enable internal feature transfer, thereby improving the consistency between audio and facial expressions generated from text. We also discuss the extension of the emotion control mechanism from TTS to the joint model. This work does not aim to showcase generation quality; instead, from a system design perspective, it validates the feasibility of reusing intermediate representations from TTS for joint modeling of speech and facial expressions, and provides engineering practice references for subsequent speech expression co-design. The project code has been open source at: https://github.com/GoldenFishes/UniTAF

Semantic communication promises task-aligned transmission but must reconcile semantic fidelity with stringent latency guarantees in immersive and safety-critical services. This paper introduces a time-constrained human-in-the-loop reinforcement learning (TC-HITL-RL) framework that embeds human feedback, semantic utility, and latency control within a semantic-aware Open radio access network (RAN) architecture. We formulate semantic adaptation driven by human feedback as a constrained Markov decision process (CMDP) whose state captures semantic quality, human preferences, queue slack, and channel dynamics, and solve it via a primal–dual proximal policy optimization algorithm with action shielding and latency-aware reward shaping. The resulting policy preserves PPO-level semantic rewards while tightening the variability of both air-interface and near-real-time RAN intelligent controller processing budgets. Simulations over point-to-multipoint links with heterogeneous deadlines show that TC-HITL-RL consistently meets per-user timing constraints, outperforms baseline schedulers in reward, and stabilizes resource consumption, providing a practical blueprint for latency-aware semantic adaptation.

It is known that waves generated by ambient noise sources and recorded by passive receivers can be used to image the reflectivities of an unknown medium. However, reconstructing the reflectivity of the medium from partial boundary measurements remains a challenging problem, particularly when the background wave speed is unknown. In this paper, we investigate passive correlation-based imaging in the daylight configuration, where uncontrolled noise sources illuminate the medium and only ambient fields are recorded by a sensor array. We first analyze daylight migration for a point reflector embedded in a homogeneous background. By introducing a searching wave speed into the migration functional, we derive an explicit characterization of the deterministic shift and defocusing effects induced by wave-speed mismatch. We show that the maximum of the envelope of the resulting functional provides a reliable estimator of the true wave speed. We then extend the analysis to a random medium with correlation length smaller than the wavelength. Leveraging the shift formula obtained in the homogeneous case, we introduce a virtual guide star that remains fixed under migration with different searching speeds. This property enables an effective wave-speed estimation strategy based on spatial averaging around the virtual guide star. For both homogeneous and random media, we establish resolution analyses for the proposed wave-speed estimators. Numerical experiments are conducted to validate the theoretical result.

In this paper we consider physics-informed detection of terrain material change in radar imagery (e.g., shifts in permittivity, roughness or moisture). We propose a lightweight electromagnetic (EM) forward model to simulate bi-temporal single-look complex (SLC) images from labelled material maps. On these data, we derive physics-aware feature stacks that include interferometric coherence, and evaluate unsupervised detectors: Reed-Xiaoli (RX)/Local-RX with robust scatter (Tyler's M-estimator), Coherent Change Detection (CCD), and a compact convolutional auto-encoder. Monte Carlo experiments sweep dielectric/roughness/moisture changes, number of looks and clutter regimes (gamma vs K-family) at fixed probability of false alarm. Results on synthetic but physically grounded scenes show that coherence and robust covariance markedly improve anomaly detection of material changes; a simple score-level fusion achieves the best F1 in heavy-tailed clutter.

Certifying and accelerating execution times of nonlinear model predictive control (NMPC) implementations are two core requirements. Execution-time certificate guarantees that the NMPC controller returns a solution before the next sampling time, and achieving faster worst-case and average execution times further enables its use in a wider set of applications. However, NMPC produces a nonlinear program (NLP) for which it is challenging to derive its execution time certificates. Our previous works, \citep{wu2025direct,wu2025time} provide data-independent execution time certificates (certified number of iterations) for box-constrained quadratic programs (BoxQP). To apply the time-certified BoxQP algorithm \citep{wu2025time} for state-input constrained NMPC, this paper i) learns a linear model via Koopman operator; ii) proposes a dynamic-relaxation construction approach yields a structured BoxQP rather than a general QP; iii) exploits the structure of BoxQP, where the dimension of the linear system solved in each iteration is reduced from $5N(n_u+n_x)$ to $Nn_u$ (where $n_u, n_x, N$ denote the number of inputs, states, and length of prediction horizon), yielding substantial speedups (when $n_x \gg n_u$, as in PDE control).

The scenario approach is an established data-driven design framework that comes equipped with a powerful theory linking design complexity to generalization properties. In this approach, data are simultaneously used both for design and for certifying the design's reliability, without resorting to a separate test dataset. This paper takes a step further by guaranteeing additional properties, useful in post-design usage but not considered during the design phase. To this end, we introduce a two-level framework of appropriateness: baseline appropriateness, which guides the design process, and post-design appropriateness, which serves as a criterion for a posteriori evaluation. We provide distribution-free upper bounds on the risk of failing to meet the post-design appropriateness; these bounds are computable without using any additional test data. Under additional assumptions, lower bounds are also derived. As part of an effort to demonstrate the usefulness of the proposed methodology, the paper presents two practical examples in H2 and pole-placement problems. Moreover, a method is provided to infer comprehensive distributional knowledge of relevant performance indexes from the available dataset.

Reliable detection and tracking in active sonar require accurate and efficient learning of the acoustic multipath background environment. Conventionally, background learning is performed after transforming measurements into the range-Doppler domain, a step that is computationally expensive and can obscure phase-coherent structure useful for monitoring and tracking. This paper proposes a framework for learning and tracking the multipath background directly in the raw measurement domain. Starting from a wideband Doppler linearization of the impulse response of a time-varying multipath channel, a state-space model with a heteroscedastic measurement equation is derived. This model enables channel tracking using an extended Kalman filter (EKF), and unknown model parameters are learned from the marginalized likelihood. The statistical adequacy of the proposed models is assessed via a p-value significance test. Finally, this paper integrates the learned channel model into a sequential likelihood-ratio test for target detection. BELLHOP-based simulations show that the proposed model better captures channel dynamics induced by sea-surface fluctuations and transmitter and receiver drift, yielding more reliable detection in time-varying shallow-water environments

We study joint transmit-waveform and receive-filter design for a multi-user downlink integrated sensing and communication (ISAC) system under practical constant-modulus and similarity constraints. We cast the design as a unified multi-objective program that balances communication sum rate and sensing signal-to-interference-plus-noise ratio (SINR). To address this, we introduce an efficient algorithm that use consensus alternating direction method of multipliers (ADMM) framework to alternately update the transmit waveform and radar filter. The proposed method effectively handles the non-convex fractional sensing's SINR formulation and ensures fast convergence. Simulation results demonstrate that the proposed approach achieves better trade-offs between communication sum rate and sensing's SINR compared to existing benchmark schemes.

Pure speech language models aim to learn language directly from raw audio without textual resources. A key challenge is that discrete tokens from self-supervised speech encoders result in excessively long sequences, motivating recent work on syllable-like units. However, methods like Sylber and SyllableLM rely on intricate multi-stage training pipelines. We propose ZeroSyl, a simple training-free method to extract syllable boundaries and embeddings directly from a frozen WavLM model. Using L2 norms of features in WavLM's intermediate layers, ZeroSyl achieves competitive syllable segmentation performance. The resulting segments are mean-pooled, discretized using K-means, and used to train a language model. ZeroSyl outperforms prior syllabic tokenizers across lexical, syntactic, and narrative benchmarks. Scaling experiments show that while finer-grained units are beneficial for lexical tasks, our discovered syllabic units exhibit better scaling behavior for syntactic modeling.

This paper addresses the problem of adaptive codebook (CB) selection for downlink (DL) precoder quantization in channel state information (CSI) reporting. The accuracy of precoder quantization depends on propagation conditions, requiring independent parameter adaptation for each user equipment (UE). To enable optimal CB selection, this paper proposes UE-assisted CB selection at the base station (BS) using reported by the UE statistical channel properties across time, frequency, and spatial domains. The reported assistance information serves as input to a neural network (NN), which predicts the quantization accuracy of various CB types for each served user. The predicted accuracy is then used to select the optimal CB while considering the associated CSI reporting overhead and precoding performance. System-level simulations demonstrate that the proposed approach reduces total CSI overhead while maintaining the target system throughput performance.

We derive a small-signal transfer function for a system comprising a virtual synchronous generator (VSG), a synchronous generator (SG), and a load, capturing voltage and frequency dynamics. Using this model, we analyze the sensitivity of SG dynamics to VSG parameters, highlighting trade-offs in choosing virtual inertia and governor lag, the limited effect of damper-winding emulation, and several others.

Target speech extraction (TSE) typically relies on pre-recorded high-quality enrollment speech, which disrupts user experience and limits feasibility in spontaneous interaction. In this paper, we propose Enroll-on-Wakeup (EoW), a novel framework where the wake-word segment, captured naturally during human-machine interaction, is automatically utilized as the enrollment reference. This eliminates the need for pre-collected speech to enable a seamless experience. We perform the first systematic study of EoW-TSE, evaluating advanced discriminative and generative models under real diverse acoustic conditions. Given the short and noisy nature of wake-word segments, we investigate enrollment augmentation using LLM-based TTS. Results show that while current TSE models face performance degradation in EoW-TSE, TTS-based assistance significantly enhances the listening experience, though gaps remain in speech recognition accuracy.

In this study, we have presented a novel approach to predict the Short-Time Objective Intelligibility (STOI) metric using a bottleneck transformer architecture. Traditional methods for calculating STOI typically requires clean reference speech, which limits their applicability in the real world. To address this, numerous deep learning-based nonintrusive speech assessment models have garnered significant interest. Many studies have achieved commendable performance, but there is room for further improvement. We propose the use of bottleneck transformer, incorporating convolution blocks for learning frame-level features and a multi-head self-attention (MHSA) layer to aggregate the information. These components enable the transformer to focus on the key aspects of the input data. Our model has shown higher correlation and lower mean squared error for both seen and unseen scenarios compared to the state-of-the-art model using self-supervised learning (SSL) and spectral features as inputs.

Today we design wireless networks using mathematical models that govern communication in different propagation environments. We rely on measurement campaigns to deliver parametrized propagation models, and on the 3GPP standards process to optimize model-based performance, but as wireless networks become more complex this model-based approach is losing ground. Mobile Network Operators (MNOs) are counting on Artificial Intelligence (AI) to transform wireless by increasing spectral efficiency, reducing signaling overhead, and enabling continuous network innovation through software upgrades. They may also be interested in new use cases like integrated sensing and communications (ISAC). All we need is an AI-native physical layer, so why not simply tailor the offline AI algorithms that have revolutionized image and natural language processing to the wireless domain? We argue that these algorithms rely on off-line training that is precluded by the sub-millisecond speeds at which the wireless interference environment changes. We present an alternative architecture, a universal neural receiver based on convolution, which governs transmit and receive signal processing of any signal in any part of the wireless spectrum. Our neural receiver is designed to invert convolution, and we separate the question of which convolution to invert from the actual deconvolution. The neural network that performs deconvolution is very simple, and we configure this network by setting weights based on domain knowledge. By telling our neural network what we know, we avoid extensive offline training. By developing a universal receiver, we hope to simplify discussions about the proper choice of waveform for different use cases in the international standards. Since the receiver architecture is largely independent of technologies introduced at the base station, we hope to increase the rate of innovation in wireless.

This paper introduces an input-output bilinear Koopman realization with an optimization algorithm of lifting functions. For nonlinear systems with inputs, Koopman-based modeling is effective because the Koopman operator enables a high-dimensional linear representation of nonlinear dynamics. However, traditional approaches face significant challenges in industrial applications. Measuring all system states is often impractical due to constraints on sensor installation. Moreover, the predictive performance of a Koopman model strongly depends on the choice of lifting functions, and their design typically requires substantial manual effort. In addition, although a linear time-invariant (LTI) Koopman model is the most commonly used model structure in the Koopman framework, such model exhibit limited predictive accuracy. To address these limitations, we propose an input-output bilinear Koopman modeling in which the design parameters of radial basis function (RBF)-based lifting functions are optimized using a global metaheuristic algorithm to improve long-term prediction performance. Consideration of the long-term prediction performance enhances the reliability of the resulting model. The proposed methodology is validated in simulations and experimental tests, with the airpath control system of a diesel engine as the plant to be modeled. This plant represents a challenging industrial application because it exhibits strong nonlinearities and coupled multi-input multi-output (MIMO) dynamics. These results demonstrate that the proposed input-output bilinear Koopman model significantly outperforms traditional linear Koopman models in predictive accuracy.

The paper presents an approach to the construction of stabilizing feedback for strongly nonlinear systems. The class of systems of interest includes systems with drift which are affine in control and which cannot be stabilized by continuous state feedback. The approach is independent of the selection of a Lyapunov type function, but requires the solution of a nonlinear programming 'satisficing problem' stated in terms of the logarithmic coordinates of flows. As opposed to other approaches, point-to-point steering is not required to achieve asymptotic stability. Instead, the flow of the controlled system is required to intersect periodically a certain reachable set in the space of the logarithmic coordinates.

Generative models have made it possible to synthesize highly realistic images, potentially providing an abundant data source for training machine learning models. Despite the advantages of these synthesizable data sources, the indiscriminate use of generated images as real images for training can even cause mode collapse due to modality discrepancies between real and synthetic domains. In this paper, we propose a novel framework for discriminative use of generated images, coined GMAIL, that explicitly treats generated images as a separate modality from real images. Instead of indiscriminately replacing real images with generated ones in the pixel space, our approach bridges the two distinct modalities in the same latent space through a multi-modal learning approach. To be specific, we first fine-tune a model exclusively on generated images using a cross-modality alignment loss and then employ this aligned model to further train various vision-language models with generated images. By aligning the two modalities, our approach effectively leverages the benefits of recent advances in generative models, thereby boosting the effectiveness of generated image learning across a range of vision-language tasks. Our framework can be easily incorporated with various vision-language models, and we demonstrate its efficacy throughout extensive experiments. For example, our framework significantly improves performance on image captioning, zero-shot image retrieval, zero-shot image classification, and long caption retrieval tasks. It also shows positive generated data scaling trends and notable enhancements in the captioning performance of the large multimodal model, LLaVA.

This paper presents a performance comparison of different estimation and prediction techniques applied to the problem of tracking multiple robots. The main performance criteria are the magnitude of the estimation or prediction error, the computational effort and the robustness of each method to non-Gaussian noise. Among the different techniques compared are the well known Kalman filters and their different variants (e.g. extended and unscented), and the more recent techniques relying on Sequential Monte Carlo Sampling methods, such as particle filters and Gaussian Mixture Sigma Point Particle Filter.

Public Safety Power Shutoffs (PSPS) force rapid topology changes that can render standard operating points infeasible, requiring operators to quickly identify corrective transmission switching actions that reduce load shedding while maintaining acceptable voltage behavior. We present a verifiable, multi-stage adaptation pipeline that fine-tunes an instruction-tuned large language model (LLM) to generate open-only corrective switching plans from compact PSPS scenario summaries under an explicit switching budget. First, supervised fine-tuning distills a DC-OPF MILP oracle into a constrained action grammar that enables reliable parsing and feasibility checks. Second, direct preference optimization refines the policy using AC-evaluated preference pairs ranked by a voltage-penalty metric, injecting voltage-awareness beyond DC imitation. Finally, best-of-$N$ selection provides an inference-time addition by choosing the best feasible candidate under the target metric. On IEEE 118-bus PSPS scenarios, fine-tuning substantially improves DC objective values versus zero-shot generation, reduces AC power-flow failure from 50% to single digits, and improves voltage-penalty outcomes on the common-success set. Code and data-generation scripts are released to support reproducibility.

Reliable interpretation of cardiac ultrasound images is essential for accurate clinical diagnosis and assessment. Self-supervised learning has shown promise in medical imaging by leveraging large unlabelled datasets to learn meaningful representations. In this study, we evaluate and compare two self-supervised learning frameworks, USF-MAE, developed by our team, and MoCo v3, on the recently introduced CACTUS dataset (37,736 images) for automated simulated cardiac view (A4C, PL, PSAV, PSMV, Random, and SC) classification. Both models used 5-fold cross-validation, enabling robust assessment of generalization performance across multiple random splits. The CACTUS dataset provides expert-annotated cardiac ultrasound images with diverse views. We adopt an identical training protocol for both models to ensure a fair comparison. Both models are configured with a learning rate of 0.0001 and a weight decay of 0.01. For each fold, we record performance metrics including ROC-AUC, accuracy, F1-score, and recall. Our results indicate that USF-MAE consistently outperforms MoCo v3 across metrics. The average testing AUC for USF-MAE is 99.99% (+/-0.01% 95% CI), compared to 99.97% (+/-0.01%) for MoCo v3. USF-MAE achieves a mean testing accuracy of 99.33% (+/-0.18%), higher than the 98.99% (+/-0.28%) reported for MoCo v3. Similar trends are observed for the F1-score and recall, with improvements statistically significant across folds (paired t-test, p=0.0048 < 0.01). This proof-of-concept analysis suggests that USF-MAE learns more discriminative features for cardiac view classification than MoCo v3 when applied to this dataset. The enhanced performance across multiple metrics highlights the potential of USF-MAE for improving automated cardiac ultrasound classification.

Pulsatile drift makes molecular arrival times depend on the release phase. We study a one-dimensional absorbing timing channel with constant diffusivity and positive-mean sinusoidal drift. A moving-boundary formulation yields an exact Volterra characterization of the phase-conditioned first-hitting law, with regularity, uniqueness, properness, and moment properties. Within this framework, we integrate and assess the corrected inverse-Gaussian (C-IG) approximation introduced in our conference work. Comparisons with a numerically validated Volterra reference distinguish raw-kernel mass discrepancy from normalized shape errors and show that close cumulative agreement can coexist with density and peak-count discrepancies, particularly under transient backflow. Phase-family analysis further shows that a small between-phase contribution to delay variance can coexist with substantial differences between conditional cumulative distributions. We connect these conditional laws and protocol-induced mixtures to timing likelihoods and absorption probabilities within a prescribed reception window, clarifying the arrival-time structures preserved by C-IG and the phase information concealed by averaging.

Decentralized air traffic management requires coordination among self-interested stakeholders operating under shared safety and capacity constraints, where conventional centralized or implicitly cooperative models do not adequately capture this setting. We develop a unified perspective on noncooperative coordination, in which system-level outcomes emerge by designing incentives and assigning signals that reshape individual optimality rather than imposing cooperation or enforcement. We advance this framework along three directions: scalable equilibrium engineering via reduced-rank and uncertainty-aware correlated equilibria, decentralized mechanism design for equilibrium selection without enforcement, and structured noncooperative dynamics with convergence guarantees. Beyond these technical contributions, we discuss core design principles that govern incentive-compatible coordination in decentralized systems. Together, these results establish a foundation for scalable, robust coordination in safety-critical air traffic systems.

We propose SCENE (Self-Centering Noncoherent Estimator), a pilot-free and phase-invariant aggregation primitive for over-the-air federated distillation (OTA-FD). Each device maps its soft-label (class-probability) vector to nonnegative transmit energies under constant per-round power and constant-envelope signaling (PAPR near 1). At the server, a self-centering energy estimator removes the noise-energy offset and yields an unbiased estimate of the weighted soft-label average, with variance decaying on the order of 1/(SM) in the number of receive antennas M and repetition factor S. We also develop a pilot-free ratio-normalized variant that cancels unknown large-scale gains, provide a convergence bound consistent with coherent OTA-FD analyses, and present an overhead-based crossover comparison. SCENE targets short-coherence and hardware-constrained regimes, where avoiding per-round CSI is essential: it trades a modest noncoherent variance constant for zero uplink pilots, unbiased aggregation, and hardware-friendly transmission, and can outperform coherent designs when pilot overhead is non-negligible.