Papers

Ensuring that Text-to-Speech (TTS) systems deliver human-perceived quality at scale is a central challenge for modern speech technologies. Human subjective evaluation protocols such as Mean Opinion Score (MOS) and Side-by-Side (SBS) comparisons remain the de facto gold standards, yet they are expensive, slow, and sensitive to pervasive assessor biases. This study addresses these barriers by formulating, and implementing a suite of novel neural models designed to approximate expert judgments in both relative (SBS) and absolute (MOS) settings. For relative assessment, we propose NeuralSBS, a HuBERT-backed model achieving 73.7% accuracy (on SOMOS dataset). For absolute assessment, we introduce enhancements to MOSNet using custom sequence-length batching, as well as WhisperBert, a multimodal stacking ensemble that combines Whisper audio features and BERT textual embeddings via weak learners. Our best MOS models achieve a Root Mean Square Error (RMSE) of ~0.40, significantly outperforming the human inter-rater RMSE baseline of 0.62. Furthermore, our ablation studies reveal that naively fusing text via cross-attention can degrade performance, highlighting the effectiveness of ensemble-based stacking over direct latent fusion. We additionally report negative results with SpeechLM-based architectures and zero-shot LLM evaluators (Qwen2-Audio, Gemini 2.5 flash preview), reinforcing the necessity of dedicated metric learning frameworks.

Robotic manipulation tasks that require repeated tool motion along curved surfaces frequently arise in surface finishing, inspection, and guided interaction. In practice, nominal motion primitives are often designed independently of the deployment surface and later reused across varying geometries. Directly tiling such primitives onto nonplanar surfaces introduces geometric inconsistencies, leading to interpenetration, orientation discontinuities, and cumulative drift over repeated cycles. We present a two-stage framework that separates geometric embedding from execution-level regulation. An offline surface-constrained warping operator embeds a nominal periodic primitive onto curved surfaces through asymmetric diffeomorphic deformation of dual-track waypoints and axis-consistent orientation completion, producing a surface-adapted reference trajectory. An online contact-aware projection operator then enforces bounded deviation relative to this reference using FSR-driven disturbance adaptation and a conic orientation safety constraint. Experiments across multiple analytic surface families and real-robot validation on a sinusoidal surface demonstrate improved geometric continuity, reduced large orientation jumps, and robust contact maintenance compared with direct tiling. These results show that decoupling offline geometric remapping from lightweight online projection enables stable and repeatable surface-embedded trajectory execution under sensor-lite feedbacks.

Autonomous platforms require accurate positioning to complete their tasks. To this end, a Kalman filter-based algorithms, such as the extended Kalman filter or invariant Kalman filter, utilizing inertial and external sensor fusion are applied. To cope with real-world scenarios, adaptive noise estimation methods have been developed primarily for classical Euclidean formulations. However, these methods remain largely unexplored in the tangent Lie space, despite it provides a principled geometric framework with favorable error dynamics on Lie groups. To fill this gap, we combine invariant filtering theory with neural-aided adaptive noise estimation in real-world settings. To this end, we derive a novel theoretical extension of classical innovation-based process noise adaptation formulated directly within the Lie-group framework. We further propose a lightweight neural network that estimates the process noise covariance parameters directly from raw inertial data. Trained entirely in a sim2real framework via domain adaptation, the network captures motion-dependent and sensor-dependent noise characteristics without requiring labeled real-world data. To examine our proposed neural-aided adaptive invariant Kalman filter, we focus on the challenging real-world scenario of autonomous underwater navigation. Experimental results demonstrate superior performance compared to existing methods in terms of position root mean square error. These results validate our sim2real pipeline and further confirm that geometric invariance significantly enhances learning-based adaptation and that adaptive noise estimation in the tangent Lie space offers a powerful mechanism for improving navigation accuracy in nonlinear autonomous platforms.

The recently deployed 6 GHz spectrum in the U.S. utilizes distinct power categories, with the latest proposed "Geofenced Variable Power" (GVP) category permitting indoor and outdoor operations without continuous Automated Frequency Coordination (AFC) by relying instead on local databases of exclusion zones. Consequently, the safe operation of GVP devices depends entirely on reliable GNSS localization to respect these geofences. However, GNSS accuracy is highly variable and significantly degrades in environments like urban canyons or indoors. This paper presents the first comprehensive empirical study evaluating GNSS reliability specifically for GVP compliance. Utilizing the SigCap Android application, we document and compare GNSS accuracy across an extensive array of real-world conditions, encompassing urban versus suburban landscapes, varying mobility states (stationary, walking, driving), and indoor versus outdoor settings. The results demonstrate that while device hardware causes variations in GNSS accuracy, the operational environment is the primary driver of error. Indoor settings and dense urban areas consistently degrade localization. Moreover, outdoor positions adjacent to buildings often surprisingly produce significant inaccuracies, even near low-elevation structures. We further analyze the contribution of different GNSS constellations to device positioning and show that satellites from non-U.S.-licensed constellations-although currently used in a substantial portion of location fixes-are not permitted for regulatory geolocation under FCC requirements.

We investigate three distinct methods of incorporating all-sky imager (ASI) images into deep learning (DL) irradiance nowcasting. The first method relies on a convolutional neural network (CNN) to extract features directly from raw RGB images. The second method uses state-of-the-art algorithms to engineer 2D feature maps informed by domain knowledge, e.g., cloud segmentation, the cloud motion vector, solar position, and cloud base height. These feature maps are then passed to a CNN to extract compound features. The final method relies on aggregating the engineered 2D feature maps into time-series input. Each of the three methods were then used as part of a DL model trained on a high-frequency, 29-day dataset to generate multi-horizon forecasts of global horizontal irradiance up to 15 minutes ahead. The models were then evaluated using root mean squared error and skill score on 7 selected days of data. Aggregated engineered ASI features as model input yielded superior forecasting performance, demonstrating that integration of ASI images into DL nowcasting models is possible without complex spatially-ordered DL-architectures and inputs, underscoring opportunities for alternative image processing methods as well as the potential for improved spatial DL feature processing methods.

Remote photoplethysmography (rPPG) based on traditional frame-based cameras often struggles with motion artifacts and limited temporal resolution. To address these limitations, we introduce EMPD (Event-based Multimodal Physiological Dataset), the first benchmark dataset specifically designed for non-contact physiological sensing via event cameras. The dataset leverages a laser-assisted acquisition system where a high-coherence laser modulates subtle skin vibrations from the radial artery into significant signals detectable by a neuromorphic sensor. The hardware platform integrates a high-resolution event camera to capture micro-motions and intensity transients, an industrial RGB camera to provide traditional rPPG benchmarks, and a clinical-grade pulse oximeter to record ground truth PPG waveforms. EMPD contains 193 valid records collected from 83 subjects, covering a wide heart rate range (40-110 BPM) under both resting and post-exercise conditions. By providing precisely synchronized multimodal data with microsecond-level temporal precision, EMPD serves as a crucial resource for developing robust algorithms in the field of neuromorphic physiological monitoring. The dataset is publicly available at: https://doi.org/10.5281/zenodo.18765701

In electricity markets, customers are increasingly constrained by their budgets. A budget constraint for a user is an upper bound on the price multiplied by the quantity. However, since prices are determined by the market equilibrium, the budget constrained welfare maximization problem is difficult to define rigorously and to work with. In this letter, we show that a natural dual-ascent algorithm converges to a unique competitive equilibrium under budget constraints. Further, this budget-constrained equilibrium is exactly the solution of a convex welfare maximization problem in which each user's utility is replaced by a modified utility that splices the original utility with a logarithmic function where the budget binds. We also provide an explicit piecewise construction of this modified utility and demonstrate the results on examples with quadratic and square root utility functions.

Quality assessment of brain structural MR images is critical for large-scale neuroimaging studies, where motion artifacts can significantly bias clinical estimates. While visual rating remains the gold standard, it is time-consuming and subjective. This study evaluates the relative performance and generalization capabilities of two prominent Automated Quality Assessment (AQA) methods: MRIQC, which uses hand-crafted image-quality metrics with traditional machine learning, and CNNQC, which utilizes a deep learning (DL) architecture. Using a heterogeneous dataset of 1,098 T1-weighted volumes from 17 different sites, we assessed performance on both seen sites and entirely new sites using a leave-one-site-out (LOSO) approach. Our results indicate that both DL and traditional ML methods struggle to generalize to new scanners or sites. While MRIQC generally achieved higher accuracy across most unseen sites, CNNQC demonstrated higher sensitivity for detecting poor-quality scans. Given that DL-based methods like CNNQC offer higher computational efficiency and do not require expensive pre-processing, they may be preferred for widespread deployment, provided that future work focuses on improving cross-site generalizability.

Recent Audio Multimodal Large Language Models (Audio MLLMs) demonstrate impressive performance on speech benchmarks, yet it remains unclear whether these models genuinely process acoustic signals or rely on text-based semantic inference. To systematically study this question, we introduce DEAF (Diagnostic Evaluation of Acoustic Faithfulness), a benchmark of over 2,700 conflict stimuli spanning three acoustic dimensions: emotional prosody, background sounds, and speaker identity. Then, we design a controlled multi-level evaluation framework that progressively increases textual influence, ranging from semantic conflicts in the content to misleading prompts and their combination, allowing us to disentangle content-driven bias from prompt-induced sycophancy. We further introduce diagnostic metrics to quantify model reliance on textual cues over acoustic signals. Our evaluation of seven Audio MLLMs reveals a consistent pattern of text dominance: models are sensitive to acoustic variations, yet predictions are predominantly driven by textual inputs, revealing a gap between high performance on standard speech benchmarks and genuine acoustic understanding.

Multi-site neuroimaging analysis is fundamentally confounded by scanner-induced covariate shifts, where the marginal distribution of voxel intensities $P(\mathbf{x})$ varies non-linearly across acquisition protocols while the conditional anatomy $P(\mathbf{y}|\mathbf{x})$ remains constant. This is particularly detrimental to radiomic reproducibility, where acquisition variance often exceeds biological pathology variance. Existing statistical harmonization methods (e.g., ComBat) operate in feature space, precluding spatial downstream tasks, while standard deep learning approaches are theoretically bounded by local effective receptive fields (ERF), failing to model the global intensity correlations characteristic of field-strength bias. We propose SA-CycleGAN-2.5D, a domain adaptation framework motivated by the $H\Delta H$-divergence bound of Ben-David et al., integrating three architectural innovations: (1) A 2.5D tri-planar manifold injection preserving through-plane gradients $\nabla_z$ at $O(HW)$ complexity; (2) A U-ResNet generator with dense voxel-to-voxel self-attention, surpassing the $O(\sqrt{L})$ receptive field limit of CNNs to model global scanner field biases; and (3) A spectrally-normalized discriminator constraining the Lipschitz constant ($K_D \le 1$) for stable adversarial optimization. Evaluated on 654 glioma patients across two institutional domains (BraTS and UPenn-GBM), our method reduces Maximum Mean Discrepancy (MMD) by 99.1% ($1.729 \to 0.015$) and degrades domain classifier accuracy to near-chance (59.7%). Ablation confirms that global attention is statistically essential (Cohen's $d = 1.32$, $p < 0.001$) for the harder heterogeneous-to-homogeneous translation direction. By bridging 2D efficiency and 3D consistency, our framework yields voxel-level harmonized images that preserve tumor pathophysiology, enabling reproducible multi-center radiomic analysis.

Linear-quadratic Gaussian games provide a framework for modeling strategic interactions in multi-agent systems, where agents must estimate system states from noisy observations while also making decisions to optimize a quadratic cost. However, these formulations usually require agents to utilize the full set of available observations when forming their state estimates, which can be unrealistic in large-scale or resource-constrained settings. In this paper, we consider linear-quadratic Gaussian games with sparse interagent observations. To enforce sparsity in the estimation stage, we design a distributed estimator that balances estimation effectiveness with interagent measurement sparsity via a group lasso problem, while agents implement feedback Nash strategies based on their state estimates. We provide sufficient conditions under which the sparse estimator is guaranteed to trigger a corrective reset to the optimal estimation gain, ensuring that estimation quality does not degrade beyond a level determined by the regularization parameters. Simulations on a formation game show that the proposed approach yields a significant reduction in communication resources consumed while only minimally affecting the nominal equilibrium trajectories.

This paper develops a multi-period optimization framework to design a voluntary renewable program (VRP) for an electric utility company, aiming to maximize total renewable energy deployments. In the business model of VRP, the utility must ensure it generates renewable energy up to the total amount of contract during each market episode (i.e., a year), while all the revenue collected from the VRP must either be used to invest in procuring renewable capacities or to maintain the current renewable fleet and infrastructure. We thus formulate the problem as an optimal pricing problem coupled with revenue allocation and renewable deployment decisions. We model the demand function of voluntary renewable contracts as an exponential decay function based on survey data. We analytically derive the optimal pricing policy of the VRP as a function of the current grid carbon intensity. We prove that a myopic policy is conditionally optimal, which maximizes renewable capacity in each period, attains the long-run optimum due to the utility's revenue-neutral constraint. We show different binding conditions and marginal values of decision variables correspond to different phases of the energy transition, and that the utility should strategically design its revenue-sharing decisions, balancing investments in renewable expansion and subsidizing existing renewable fleets. Finally, we show that voluntary renewable programs can only extend renewable penetration but cannot achieve net-zero emissions or a fully renewable grid. This pricing-allocation-expansion framework highlights both the potential and limitations of voluntary renewable demand, providing analytical insight into optimal policy design and the qualitative shifts occurring during the energy transition process.

This paper presents a particle swarm optimization algorithm that leverages surrogate modeling to replace the conventional global best solution with the minimum of an n-dimensional quadratic form, providing a better-conditioned dynamic attractor for the swarm. This refined convergence target, informed by the local landscape, enhances global convergence behavior and increases robustness against premature convergence and noise, while incurring only minimal computational overhead. The surrogate-augmented approach is evaluated against the standard algorithm through a numerical study on a set of benchmark optimization functions that exhibit diverse landscapes. To ensure statistical significance, 400 independent runs are conducted for each function and algorithm, and the results are analyzed based on their statistical characteristics and corresponding distributions. The quadratic surrogate attractor consistently outperforms the conventional algorithm across all tested functions. The improvement is particularly pronounced for quasi-convex functions, where the surrogate model can exploit the underlying convex-like structure of the landscape.

Polarization imaging is a technique that creates a pixel map of the polarization state in a scene. Although invisible to the human eye, polarization can assist various sensing and computer vision tasks. Existing polarization cameras use spatial or temporal multiplexing, which increases the camera volume, weight, cost, or all of the above. Recent lensless imaging approaches, such as DiffuserCam, have demonstrated that compact imaging systems can be realized by replacing the lens with a coding element and performing computational reconstruction. In this work, we propose a compact lensless polarization camera composed of a diffuser and a simple striped polarization mask. By combining this optical design with a reconstruction algorithm that explicitly models the polarization-encoded lensless measurements, four linear polarization images are recovered from a single snapshot. Our results demonstrate the potential of lensless approaches for polarization imaging and reveal the physical factors that govern reconstruction quality, guiding the development of high-quality practical systems.

Networked multi-agent dynamical systems have been used to model how individual opinions evolve over time due to the opinions of other agents in the network. Particularly, such a model has been used to study how a planning agent can be used to steer opinions in a desired direction through repeated, budgeted interventions. In this paper, we consider the problem where individuals' susceptibilities to external influences are unknown. We propose an online algorithm that alternates between estimating this susceptibility parameter, and using the current estimate to drive the opinion to a desired target. We provide conditions that guarantee stability and convergence to the desired target opinion when the planning agent faces budgetary or temporal constraints. Our analysis shows that the key advantage of estimating the susceptibility parameter is that it helps achieve near-optimal convergence to the target opinion given a finite amount of intervention rounds, and, for a given intervention budget, quantifies how close the opinion can get to the desired target.

This paper presents a new dynamic integral quadratic constraint (IQC) for the repeated Rectified Linear Unit (ReLU). These dynamic IQCs can be used to analyze stability and induced $\ell_2$-gain performance of discrete-time, recurrent neural networks (RNNs) with ReLU activation functions. These analysis conditions can be incorporated into learning-based controller synthesis methods, which currently rely on static IQCs. We show that our proposed dynamic IQCs for repeated ReLU form a superset of the dynamic IQCs for repeated, slope-restricted nonlinearities. We also prove that the $\ell_2$-gain bounds are nonincreasing with respect to the horizon used in the dynamic IQC filter. A numerical example using a simple (academic) RNN shows that our proposed IQCs lead to less conservative bounds than existing IQCs.

Converter-based generators and loads are growing in prevalence on power grids across the globe. The rise of these resources necessitates controllers that handle the power electronic devices' strict current limits without jeopardizing stability or overly constraining behavior. Existing controllers often employ complex, cascaded control loop architecture to saturate currents, but these controllers are challenging to tune properly and can destabilize following large disturbances. In this paper, we extend previous analysis to prove the feasible output region of a grid-connected converter is convex regardless of filter topology. We then formulate a convex optimal control problem from which we derive a projected gradient descent-based controller with convergence guarantees. This approach drives the converter toward optimality in real-time and differs from conventional control strategies that regulate converter outputs around predefined references regardless of surrounding grid conditions. Simulation results demonstrate safe and stabilizing behavior of the proposed controller, in both the single-converter-infinite-bus systems and multi-converter networks.

This paper presents Neural-NPV Control, a learning-based framework for joint synthesis of a parameter-dependent (PD) controller and a PD Lyapunov function using neural networks for an NPV system under input constraints. At the first stage, the proposed framework utilizes a gradient-based counterexample-guided procedure to synthesize a PD controller and a PD Lyapunov function candidate. The second stage relies on a level-set guided procedure to refine the controller and Lyapunov function candidate while maximizing the robust region of attraction (R-ROA). The learned controller, Lyapunov function, and R-ROA are empirically evaluated. We demonstrate the advantages of Neural-NPV over SOS-based methods in terms of applicability, performance, and scalability through numerical experiments involving a simple inverted pendulum with one scheduling parameter and a quadrotor system with three scheduling parameters.

Many wireless vision applications, such as autonomous driving, require preservation of global structural information rather than only per-pixel fidelity. However, existing Deep joint source-channel coding (DeepJSCC) schemes mainly optimize pixel-wise losses and provide no explicit protection of connectivity or topology. This letter proposes TopoJSCC, a topology-aware DeepJSCC framework that integrates persistent-homology regularizers to end-to-end training. Specifically, we enforce topological consistency by penalizing Wasserstein distances between cubical persistence diagrams of original and reconstructed images, and between Vietoris–Rips persistence of latent features before and after the channel to promote a robust latent manifold. TopoJSCC is based on end-to-end learning and requires no side information. Experiments show improved topology preservation and peak signal-to-noise ratio (PSNR) in low signal-to-noise ratio (SNR) and bandwidth-ratio regimes.

Ensuring the safety of control systems often requires the satisfaction of constraints on states (such as position or velocity), control inputs (such as force), and a mixture of states and inputs (such as power that depends on both velocity and force). This paper presents a safety-critical control framework for enforcing mixed state-input constraints through a generalization of backup control barrier functions (backup CBFs). First, we extend the backup CBF approach to maintain multiple decoupled state and input constraints using a single backup set-backup controller pair. Second, we address mixed state-input constraints by converting them into state constraints using a projection from the state-input space to the state space along the backup controller. In the special case of decoupled state and input constraints, the proposed method simplifies the synthesis of backup CBFs by eliminating the need for saturating backup control laws. Finally, we demonstrate the efficacy of the proposed method on an inverted pendulum example, where constraints on the angle (state), torque (input), and power (mixture of state and input) are satisfied simultaneously.

Data-driven model predictive control based on Willems' fundamental lemma has proven effective for linear systems, but extending stability guarantees to nonlinear systems remains an open challenge. In this paper, we establish conditions under which data-driven MPC, applied directly to input-output data from a nonlinear system, yields practical exponential stability. The key insight is that the existence of an approximate Koopman linear embedding certifies that the nonlinear data can be interpreted as noisy data from a linear time-invariant system, enabling the application of existing robust stability theories. Crucially, the Koopman embedding serves only as a theoretical certificate; the controller itself operates on raw nonlinear data without knowledge of the lifting functions. We further show that the proportional structure of the embedding residual can be exploited to obtain an ultimate bound that depends only on the irreducible offset, rather than the worst-case embedding error. The framework is demonstrated on a synchronous generator connected to an infinite bus, for which we construct an explicit physics-informed embedding with error bounds.

In this work, we study the design of receivers for uplink multi-user systems, aiming to estimate both the channel and the transmitted symbols. We consider two estimation strategies: (i) a joint estimation approach, where the channel and symbols are estimated simultaneously, and (ii) a sequential estimation approach, where the channel is first estimated and then used for symbol detection. For both strategies, we derive the Cramér-Rao Bound (CRB) for symbol estimation to characterize fundamental performance limits. When efficient receivers achieving the CRB exist, these bounds provide accurate lower bounds on the mutual information. In general, however, such receivers may not be available, and we instead use these same CRB-based metrics as practical proxies for achievable throughput. Leveraging tools from random matrix theory (RMT), we analyze the asymptotic behavior of these lower bounds under various asymptotic regimes for both estimation strategies. This analysis enables the derivation of generic power allocation guidelines that asymptotically maximize the proxy metrics. Simulation results confirm the accuracy of the asymptotic expressions and their effectiveness in guiding resource allocation decisions.

Collecting everyday speech data for prosodic analysis is challenging due to the confounding of prosody and semantics, privacy constraints, and participant compliance. We introduce and empirically evaluate a content-controlled, privacy-first smartphone protocol that uses scripted read-aloud sentences to standardize lexical content (including prompt valence) while capturing naturalistic variation in prosodic delivery. The protocol performs on-device prosodic feature extraction, deletes raw audio immediately, and transmits only derived features for analysis. We deployed the protocol in a large study (N = 560; 9,877 recordings), evaluated compliance and data quality, and conducted diagnostic prediction tasks on the extracted features, predicting self-reported speaker sex and momentary affective states (valence, arousal). We discuss implications and directions for advancing and deploying the protocol.

Nash equilibria provide a principled framework for modeling interactions in multi-agent decision-making and control. However, many equilibrium-seeking methods implicitly assume that each agent has access to the other agents' objectives and constraints, an assumption that is often unrealistic in practice. This letter studies a class of asymmetric-information two-player constrained games with decoupled feasible sets, in which Player 1 knows its own objective and constraints while Player 2 is available only through a best-response map. For this class of games, we propose an asymmetric projected gradient descent-best response iteration that does not require full mutual knowledge of both players' optimization problems. Under suitable regularity conditions, we establish the existence and uniqueness of the Nash equilibrium and prove global linear convergence of the proposed iteration when the best-response map is exact. Recognizing that best-response maps are often learned or estimated, we further analyze the inexact case and show that, when the approximation error is uniformly bounded by $\varepsilon$, the iterates enter an explicit $O(\varepsilon)$ neighborhood of the true Nash equilibrium. Numerical results on a benchmark game corroborate the predicted convergence behavior and error scaling.

Human listeners exhibit the remarkable ability to segregate a desired sound from complex acoustic scenes through selective auditory attention, motivating the study of Targeted Sound Detection (TSD). The task requires detecting and localizing a target sound in a mixture when a reference audio of that sound is provided. Prior approaches, rely on generating a sound-discriminative conditional embedding vector for the reference and pairing it with a mixture encoder, jointly optimized with a multi-task learning approach. In this work, we propose a unified encoder architecture that processes both the reference and mixture audio within a shared representation space, promoting stronger alignment while reducing architectural complexity. This design choice not only simplifies the overall framework but also enhances generalization to unseen classes. Following the multi-task training paradigm, our method achieves substantial improvements over prior approaches, surpassing existing methods and establishing a new state-of-the-art benchmark for targeted sound detection, with a segment-level F1 score of 83.15% and an overall accuracy of 95.17% on the URBAN-SED dataset.

Hamilton-Jacobi (HJ) reachability provides formal safety guarantees for dynamical systems, but solving high-dimensional HJ partial differential equations limits its use in real-time planning. This paper presents a contingency-aware multi-goal navigation framework that integrates learning-based reachability with sampling-based planning in unknown environments. We use Fourier Neural Operator (FNO) to approximate the solution operator of the Hamilton-Jacobi-Isaacs variational inequality under varying obstacle configurations. We first provide a theoretical under-approximation guarantee on the safe backward reach-avoid set, which enables formal safety certification of the learned reachable sets. Then, we integrate the certified reachable sets with an incremental multi-goal planner, which enforces reachable-set constraints and a recovery policy that guarantees finite-time return to a safe region. Overall, we demonstrate that the proposed framework achieves asymptotically optimal navigation with provable contingency behavior, and validate its performance through real-time deployment on KUKA's youBot in Webots simulation.

Generalized Nash Equilibrium Problems (GNEPs) arise in many applications, including non-cooperative multi-agent control problems. Although many methods exist for finding generalized Nash equilibria, most of them rely on assuming knowledge of the objective functions or being able to query the best responses of the agents. We present a method for learning solutions of GNEPs only based on querying agents for their preference between two alternative decisions. We use the collected preference data to learn a GNEP whose equilibrium approximates a GNE of the underlying (unknown) problem. Preference queries are selected using an active-learning strategy that balances exploration of the decision space and exploitation of the learned GNEP. We present numerical results on game-theoretic linear quadratic regulation problems, as well as on other literature GNEP examples, showing the effectiveness of the proposed method.

We propose a constricting Control Barrier Function (CBF) framework for prescribed-time control of control-affine systems with input constraints. Given a system starting outside a target safe set, we construct a time-varying safety tube that shrinks from a relaxed set containing the initial condition to the target set at a user-specified deadline. Any controller rendering this tube forward invariant guarantees prescribed-time recovery by construction. The constriction schedule is bounded and tunable by design, in contrast to prescribed-time methods where control effort diverges near the deadline. Feasibilityå under input constraints reduces to a single verifiable condition on the constriction rate, yielding a closed-form minimum recovery time as a function of control authority and initial violation. The framework imposes a single affine constraint per timestep regardless of state dimension, scaling to settings where grid-based reachability methods are intractable. We validate on an 18-dimensional multi-agent system, demonstrating scalability and prescribed-time recovery with bounded control effort.

Automatic speech recognition systems based on neural networks are vulnerable to adversarial attacks that alter transcriptions in a malicious way. Recent works in this field have focused on making attacks work in over-the-air scenarios, however such attacks are typically detectable by human hearing, limiting their potential applications. In the present work we explore different approaches of making over-the-air attacks less detectable, as well as the impact these approaches have on the attacks' effectiveness.

Traditional speaker diarization systems have primarily focused on constrained scenarios such as meetings and interviews, where the number of speakers is limited and acoustic conditions are relatively clean. To explore open-world speaker diarization, we extend this task to the visual media domain, encompassing complex audiovisual programs such as films and TV series. This new setting introduces several challenges, including long-form video understanding, a large number of speakers, cross-modal asynchrony between audio and visual cues, and uncontrolled in-the-wild variability. To address these challenges, we propose Cinematic Speaker Registration & Diarization (CineSRD), a unified multimodal framework that leverages visual, acoustic, and linguistic cues from video, speech, and subtitles for speaker annotation. CineSRD first performs visual anchor clustering to register initial speakers and then integrates an audio language model for speaker turn detection, refining annotations and supplementing unregistered off-screen speakers. Furthermore, we construct and release a dedicated speaker diarization benchmark for visual media that includes Chinese and English programs. Experimental results demonstrate that CineSRD achieves superior performance on the proposed benchmark and competitive results on conventional datasets, validating its robustness and generalizability in open-world visual media settings.