Papers

Reliable quantification of uncertainty in Mobile Laser Scanning (MLS) point clouds is essential for ensuring the accuracy and credibility of downstream applications such as 3D mapping, modeling, and change analysis. Traditional backward uncertainty modeling heavily rely on high-precision reference data, which are often costly or infeasible to obtain at large scales. To address this issue, this study proposes a machine learning-based framework for point-level uncertainty evaluation that learns the relationship between local geometric features and point-level errors. The framework is implemented using two ensemble learning models, Random Forest (RF) and XGBoost, which are trained and validated on a spatially partitioned real-world dataset to avoid data leakage. Experimental results demonstrate that both models can effectively capture the nonlinear relationships between geometric characteristics and uncertainty, achieving mean ROC-AUC values above 0.87. The analysis further reveals that geometric features describing elevation variation, point density, and local structural complexity play a dominant role in predicting uncertainty. The proposed framework offers a data-driven perspective of uncertainty evaluation, providing a scalable and adaptable foundation for future quality control and error analysis of large-scale point clouds.

Tractography fiber clustering using diffusion MRI (dMRI) is a crucial method for white matter (WM) parcellation to enable analysis of brains structural connectivity in health and disease. Current fiber clustering strategies primarily use the fiber geometric characteristics (i.e., the spatial trajectories) to group similar fibers into clusters, while neglecting the functional and microstructural information of the fiber tracts. There is increasing evidence that neural activity in the WM can be measured using functional MRI (fMRI), providing potentially valuable multimodal information for fiber clustering to enhance its functional coherence. Furthermore, microstructural features such as fractional anisotropy (FA) can be computed from dMRI as additional information to ensure the anatomical coherence of the clusters. In this paper, we develop a novel deep learning fiber clustering framework, namely Deep Multi-view Fiber Clustering (DMVFC), which uses joint multi-modal dMRI and fMRI data to enable functionally consistent WM parcellation. DMVFC can effectively integrate the geometric and microstructural characteristics of the WM fibers with the fMRI BOLD signals along the fiber tracts. DMVFC includes two major components: (1) a multi-view pretraining module to compute embedding features from each source of information separately, including fiber geometry, microstructure measures, and functional signals, and (2) a collaborative fine-tuning module to simultaneously refine the differences of embeddings. In the experiments, we compare DMVFC with two state-of-the-art fiber clustering methods and demonstrate superior performance in achieving functionally meaningful and consistent WM parcellation results.

This work aims to train Deep Learning models to perform Automatic Target Recognition (ATR) on Synthetic Aperture Radar (SAR) images. To circumvent the lack of real labelled measurements, we resort to synthetic data produced by SAR simulators. Simulation offers full control over the virtual environment, which enables us to generate large and diversified datasets at will. However, simulations are intrinsically grounded on simplifying assumptions of the real world (i.e. physical models). Thus, synthetic datasets are not as representative as real measurements. Consequently, ATR models trained on synthetic images cannot generalize well on real measurements. Our contributions to this problem are twofold: on one hand, we demonstrate and quantify the impact of the simulation paradigm on the ATR. On the other hand, we propose a new approach to tackle the ATR problem: combine two SAR simulators that are grounded on different (but complementary) paradigms to produce synthetic datasets. To this end, we use two simulators: MOCEM, which is based on a scattering centers model approach, and Salsa, which resorts on a ray tracing strategy. We train ATR models using synthetic dataset generated both by MOCEM and Salsa and our Deep Learning approach called ADASCA. We reach an accuracy of almost 88 % on the MSTAR measurements.

Pretrained diffusion models have demonstrated strong capabilities in zero-shot inverse problem solving by incorporating observation information into the generation process of the diffusion models. However, this presents an inherent dilemma: excessive integration can disrupt the generative process, while insufficient integration fails to emphasize the constraints imposed by the inverse problem. To address this, we propose Noise Combination Sampling, a novel method that synthesizes an optimal noise vector from a noise subspace to approximate the measurement score, replacing the noise term in the standard Denoising Diffusion Probabilistic Models process. This enables conditional information to be naturally embedded into the generation process without reliance on step-wise hyperparameter tuning. Our method can be applied to a wide range of inverse problem solvers, including image compression, and, particularly when the number of generation steps $T$ is small, achieves superior performance with negligible computational overhead, significantly improving robustness and stability.

Generative modeling has emerged as a powerful paradigm for representation learning, but its direct applicability to challenging fields like medical imaging remains limited: mere generation, without task alignment, fails to provide a robust foundation for clinical use. We propose MAGIC-Flow, a conditional multiscale normalizing flow architecture that performs generation and classification within a single modular framework. The model is built as a hierarchy of invertible and differentiable bijections, where the Jacobian determinant factorizes across sub-transformations. We show how this ensures exact likelihood computation and stable optimization, while invertibility enables explicit visualization of sample likelihoods, providing an interpretable lens into the model's reasoning. By conditioning on class labels, MAGIC-Flow supports controllable sample synthesis and principled class-probability estimation, effectively aiding both generative and discriminative objectives. We evaluate MAGIC-Flow against top baselines using metrics for similarity, fidelity, and diversity. Across multiple datasets, it addresses generation and classification under scanner noise, and modality-specific synthesis and identification. Results show MAGIC-Flow creates realistic, diverse samples and improves classification. MAGIC-Flow is an effective strategy for generation and classification in data-limited domains, with direct benefits for privacy-preserving augmentation, robust generalization, and trustworthy medical AI.

Robustness has become one of the most critical problems in machine learning (ML). The science of interpreting ML models to understand their behavior and improve their robustness is referred to as explainable artificial intelligence (XAI). One of the state-of-the-art XAI methods for computer vision problems is to generate saliency maps. A saliency map highlights the pixel space of an image that excites the ML model the most. However, this property could be misleading if spurious and salient features are present in overlapping pixel spaces. In this paper, we propose a caption-based XAI method, which integrates a standalone model to be explained into the contrastive language-image pre-training (CLIP) model using a novel network surgery approach. The resulting caption-based XAI model identifies the dominant concept that contributes the most to the models prediction. This explanation minimizes the risk of the standalone model falling for a covariate shift and contributes significantly towards developing robust ML models. Our code is available at https://github.com/patch0816/caption-driven-xai

This paper suggests a novel rotor-cooling shaft concept for high-performance electric motors that increases the effectiveness of cooling and is yet simple and cost-effective to manufacture. We investigate the thermal performance of four shaft geometries for rotor cooling in automotive applications. The proposed tooth-guided liquid-cooling shaft design aims to solve the high churning loss of conventional cooled rotor shafts due to internal vortex formation and their still limited heat transfer. Therefore, we optimize heat transfer efficiency and pressure management by incorporating cold-formed internal channels that restrict vortex formation beyond a degree that improves heat transfer. We evaluated key performance metrics, including heat transfer rate, outlet temperature, pressure drop, and velocity profiles, under varying rotational speeds, inlet flow rates, and coolant temperatures. Computational fluid analysis demonstrates that the tooth-guided design outperforms conventional hollow shafts and achieves up to 110% higher cooling efficiency at low rotational speeds, while it maintains comparable pressure levels. These findings provide practical insight into geometry-driven thermal optimization and offer a path toward improving the performance and durability of electric motors.

Wilson-Cowan and Amari-type models capture nonlinear neural population dynamics, providing a fundamental framework for modeling how sensory and other exogenous inputs shape activity in neural tissue. We study the controllability properties of Amari-type neural fields subject to piecewise/constant-in-time inputs. The model describes the time evolution of the polarization of neural tissue within a spatial continuum, with synaptic interactions represented by a convolution kernel. We study the synthesis of piecewise/constant-in-time inputs to achieve two-point boundary-type control objectives, namely, steering neural activity from an initial state to a prescribed target state. This approach is particularly relevant for predicting the emergence of paradoxical neural representations, such as discordant visual illusions that occur in response to overt sensory stimuli. We first present a control synthesis based on the Banach fixed-point theorem, which yields an iterative construction of a constant-in-time input under minimal regularity assumptions on the kernel and transfer function; however, it exhibits practical limitations, even in the linear case. To overcome these challenges, we then develop a generic synthesis framework based on the flow of neural dynamics drift, enabling explicit piecewise constant and constant-in-time inputs. Extensive numerical results in one and two spatial dimensions confirm the effectiveness of the proposed syntheses and demonstrate their superior performance compared to inputs derived from naive linearization at the initial or target states when these states are not equilibria of the drift dynamics. By providing a mathematically rigorous framework for controlling Amari-type neural fields, this work advances our understanding of nonlinear neural population control with potential applications in computational neuroscience, psychophysics, and neurostimulation.

Safety-critical applications of machine learning require uncertainty estimates that support reliable worst-case analysis. Neural networks (NNs) provide expressive function approximation, while Gaussian processes (GPs) offer principled probabilistic uncertainty, but both face limitations in this setting: most modern neural architectures lack tractable worst-case error bounds, and Gaussian processes become computationally expensive at scale. For uncertainty to be interpretable, a central requirement is distance-awareness: that is, uncertainty increases with the distance between a test input and the nearest relevant training data. We present a general framework for worst-case distance-aware error bounds for NNs that combine dense layers with spline-based components. Our approach establishes error bounds that are both distance-aware, reflecting proximity of a test point to its nearest training data, and worst-case, providing deterministic guarantees under known Lipschitz constraints rather than probabilistic assumptions. Our algorithm, K-DAREK (Distance-Aware Error for Kurkova-Kolmogorov Networks), provides efficient and interpretable uncertainty quantification for NNs. K-DAREK is about four times faster and ten times more computationally efficient than an ensemble of KANs, 8.6 times more scalable than GP, eliminates up to 8.2% error-bound violation rate observed in DAREK, and reduces the average collision rate from 1.8% to 1.1% in the multi-agent safe control experiment. On high-dimensional real-world regression tasks (e.g., Real Estate Valuation), K-DAREK preserves distance-aware error bounds and achieves zero coverage violations, addressing the overgeneralization and inducing-point-coverage limitations exhibited by SNGP and DUE, respectively.

Conventional AC Power Flow (ACPF) solvers like Newton-Raphson (NR) face significant computational and convergence challenges in modern, large-scale power systems. This paper proposes a novel, two-stage hybrid method that integrates a Physics-Informed Graph Neural Network (GNN) with a robust, iterative Linear State Estimation (LSE) refinement step to produce fast and physically-consistent solutions. The GNN, trained with a physics-informed loss function featuring an efficient dynamic weighting scheme, rapidly predicts a high-quality initial system state. This prediction is then refined using an iterative, direct linear solver inspired by state estimation techniques. This LSE refinement step solves a series of linear equations to enforce physical laws, effectively bypassing the non-linearities and convergence issues of traditional solvers. The proposed GNN-LSE framework is comprehensively validated on systems ranging from small radial distribution networks (IEEE 33-bus, 69-bus) to a large, meshed transmission system (IEEE 118-bus). Results show that our GNN variants are up to $8.4 \times 10^3$ times faster than NR. The LSE refinement provides a fast route to a physically-consistent solution, while heavy-loading stress tests (120%-150% of nominal) and N-1 contingencies demonstrate the method's reliability and generalization. This work presents a powerful and flexible framework for bridging fast, data-driven models with the rigorous constraints of power system physics, offering a practical tool for real-time operations and analysis.

We present an algorithm for planning trajectories that avoid obstacles and satisfy key-door precedence specifications expressed with a fragment of signal temporal logic. Our method includes a novel exact convex partitioning of the obstacle free space that encodes connectivity among convex free space sets, key sets, and door sets. We then construct an augmented graph of convex sets that exactly encodes the key-door precedence specifications. By solving a shortest path problem in this augmented graph of convex sets, our pipeline provides an exact solution up to a finite parameterization of the trajectory. To illustrate the effectiveness of our approach, we present a method to generate key-door mazes that provide challenging problem instances, and we perform numerical experiments to evaluate the proposed pipeline. Our pipeline is faster by several orders of magnitude than recent state-of-the art methods that use general purpose temporal logic tools.

The remarkable success of diffusion and flow-matching models has ignited a surge of works on adapting them at test time for controlled generation tasks. Examples range from image editing to restoration, compression and personalization. However, due to the iterative nature of the sampling process in those models, it is computationally impractical to use gradient-based optimization to directly control the image generated at the end of the process. As a result, existing methods typically resort to manipulating each timestep separately. Here we introduce FlowOpt - a zero-order (gradient-free) optimization framework that treats the entire flow process as a black box, enabling optimization through the whole sampling path without backpropagation through the model. Our method is both highly efficient and allows users to monitor the intermediate optimization results and perform early stopping if desired. We prove a sufficient condition on FlowOpt's step-size, under which convergence to the global optimum is guaranteed. We further show how to empirically estimate this upper bound so as to choose an appropriate step-size. We demonstrate how FlowOpt can be used for image editing, showcasing two options: (i) inversion (determining the initial noise that generates a given image), and (ii) directly steering the edited image to be similar to the source image while conforming to a target text prompt. In both cases, FlowOpt achieves state-of-the-art results while using roughly the same number of neural function evaluations (NFEs) as existing methods. Code and examples are available on the project's webpage.

Detecting single-trial P300 from EEG is difficult when only a few labeled trials are available. When attempting to boost a small target set with a large source dataset through transfer learning, cross-dataset shift arises. To address this challenge, we study transfer between two public visual-oddball ERP datasets using five shared electrodes (Fz, Pz, P3, P4, Oz) under a strict small-sample regime (target: 10 trials/subject; source: 80 trials/subject). We introduce Adaptive Split Maximum Mean Discrepancy Training (AS-MMD), which combines (i) a target-weighted loss with warm-up tied to the square root of the source/target size ratio, (ii) Split Batch Normalization (Split-BN) with shared affine parameters and per-domain running statistics, and (iii) a parameter-free logit-level Radial Basis Function kernel Maximum Mean Discrepancy (RBF-MMD) term using the median-bandwidth heuristic. Implemented on an EEG Conformer, AS-MMD is backbone-agnostic and leaves the inference-time model unchanged. Across both transfer directions, it outperforms target-only and pooled training (Active Visual Oddball: accuracy/AUC 0.66/0.74; ERP CORE P3: 0.61/0.65), with gains over pooling significant under corrected paired t-tests. Ablations attribute improvements to all three components.

We propose a generalization of the Bass diffusion model in discrete-time that explicitly models the effect of price in adoption. Our model is different from earlier price-incorporated models and fits well to adoption data for various products. We then utilize this model to study two decision-making problems. First, we provide a series of structural results on optimal pricing strategies to maximize profits from product sales by a monopolist over a finite horizon. We fully characterize the optimal pricing strategy in the single-period problem, and establish several structural properties of the same for the multi-period counterpart. Second, we study a Stackelberg game between a policy-maker and a monopolist, where the former seeks to maximize adoption through rebates, while the latter focuses on profits. For this problem, we analytically characterize crucial properties of the equilibrium path of the single-period game, and demonstrate how they carry over to the multi-period variant.

We formulate and solve the fixed horizon linear quadratic covariance steering problem in continuous time with a terminal cost measured in Hilbert-Schmidt (i.e., Frobenius) norm error between the desired and the controlled terminal covariances. For this problem, the necessary conditions of optimality become a coupled matrix ODE two-point boundary value problem. To solve this system of equations, we design a matricial recursive algorithm and prove its convergence. The proposed algorithm and its analysis make use of the linear fractional transforms parameterized by the state transition matrix of the associated Hamiltonian matrix. To illustrate the results, we provide two numerical examples: one with a two dimensional and another with a six dimensional state space.

Inverse design of metasurfaces for the joint optimization of optical modulation and algorithmic decoding in computational optics presents significant challenges, especially in applications such as hyperspectral imaging. We introduce a physics-data co-driven framework for designing reconfigurable metasurfaces fabricated from the phase-change material Ge2Sb2Se4Te1 to achieve compact, compressive spectral imaging in the shortwave infrared region. Central to our approach is a differentiable neural simulator, trained on over 320,000 simulated geometries, that accurately predicts spectral responses across 11 crystallization states. This differentiability enables end-to-end joint optimization of the metasurface geometry, its spectral encoding function, and a deep reconstruction network. We also propose a soft shape regularization technique that preserves manufacturability during gradient-based updates. Experiments show that our optimized system improves reconstruction fidelity by up to 7.6 dB in the peak-signal-to-noise ratio, with enhanced noise resilience and improved measurement matrix conditioning, underscoring the potential of our approach for high-performance hyperspectral imaging.

This paper presents a unified algebraic, topological, and logical framework for electrical one-port networks based on Šare's $m$-theory. Within this formalism, networks are represented by $m$-words (jorbs) over an ordered alphabet, where series and parallel composition induce an $m$-topology on $m$-graphs with a theta mapping $\vartheta$ that preserves one-port equivalence. The study formalizes quasi-orders, shells, and cores, showing their structural correspondence to network boundary conditions and impedance behavior. The $\lambda--\Delta$ metric, together with the valuation morphism $\Phi$, provides a concise descriptor of the impedance-degree structure. In the computational domain, the framework is extended with algorithmic procedures for generating and classifying non-isomorphic series-parallel topologies, accompanied by programmatic Cauer/Foster synthesis workflows and validation against canonical examples from Ladenheim's catalogue. The resulting approach enables symbolic-to-topological translation of impedance functions, offering a constructive bridge between algebraic representation and electrical realization. Overall, the paper outlines a self-consistent theoretical and computational foundation for automated network synthesis, classification, and formal verification within the emerging field of Jorbology.

This paper focuses on rate-limited control of the generalized Ornstein-Uhlenbeck process where the control action can be either multiplicative or additive, and the noise variance can depend on the control action. We derive a lower bound on the data rate necessary to achieve the desired control cost. The lower bound is attained with equality if the control is performed via an additive white Gaussian channel. The system model approximates the dynamics of a discrete-state molecular birth-death process, and the result has direct implications on the control of a biomolecular system via chemical reactions, where the multiplicative control corresponds to the degradation rate, the additive control corresponds to the production rate, and the control objective is to decrease the fluctuations of the controlled molecular species around their desired concentration levels.

Generalized Nash equilibria are used in multi-agent control applications to model strategic interactions between agents that are coupled in the cost, dynamics, and constraints, and provide the foundations for game-theoretic MPC (Receding Horizon Games). We study properties of finite-horizon dynamic GNE trajectories from a system-theoretic perspective. We show how strict dissipativity generates the turnpike phenomenon in GNE solutions. Moreover, we establish a converse turnpike result, i.e., the implication from turnpike to strict dissipativity. We derive conditions under which the steady-state GNE is the optimal operating point and, using a game value function, we give a local characterization of the geometry of storage functions. Finally, we design linear terminal penalties that ensure dynamic GNE trajectories applied in open-loop converge to and remain at the steady-state GNE. These connections provide the foundation for future system-theoretic analysis of GNEs similar to those existing in optimal control as well as for recursive feasibility and closed-loop stability results of game-theoretic MPC.

This paper proposes a scalable decentralized safety filter for multi-agent systems based on high-order control barrier functions (HOCBFs) and auction-based responsibility allocation. While decentralized HOCBF formulations ensure pairwise safety under input bounds, they face feasibility and scalability challenges as the number of agents grows. Each agent must evaluate an increasing number of pairwise constraints, raising the risk of infeasibility and making it difficult to meet real-time requirements. To address this, we introduce an auction-based allocation scheme that distributes constraint enforcement asymmetrically among neighbors based on local control effort estimates. The resulting directed responsibility graph guarantees full safety coverage while reducing redundant constraints and per-agent computational load. Simulation results confirm safe and efficient coordination across a range of network sizes and interaction densities.

Weak turbulence is commonly modeled using the log-normal distribution. Our experimental results show that this distribution fails to capture irradiance fluctuations in this regime. The Gamma-Gamma model is shown to be more accurate.

Deployment of optimization algorithms over communication networks face challenges associated with time delays and corruptions. Fixed time delays can destabilize popular gradient-based algorithms, and this degradation is exacerbated by time-varying delays that may arise from packet drops. This work concentrates on the analysis and synthesis of discrete-time optimization algorithms with certified exponential convergence rates that are robust against switched network dynamics between the optimizer and the gradient oracle. Analysis is accomplished by solving linear matrix inequalities under bisection in the exponential convergence rate, searching over Zames-Falb filter coefficients that can certify convergence. Synthesis is performed by alternating between a search over filter coefficient for a fixed controller, and a search over controllers for a fixed filter. Effectiveness is demonstrated by the synthesis of convergent optimization algorithms over networks with time-varying delays, and networks with unstable channel dynamics.

Speech separation and enhancement (SSE) has advanced remarkably and achieved promising results in controlled settings, such as a fixed number of speakers and a fixed array configuration. Towards a universal SSE system, single-channel systems have been extended to deal with a variable number of speakers (i.e., outputs). Meanwhile, multi-channel systems accommodating various array configurations (i.e., inputs) have been developed. However, these attempts have been pursued separately. In this paper, we propose a flexible input and output SSE system, named FlexIO. It performs conditional separation using prompt vectors, one per speaker as a condition, allowing separation of an arbitrary number of speakers. Multi-channel mixtures are processed together with the prompt vectors via an array-agnostic channel communication mechanism. Our experiments demonstrate that FlexIO successfully covers diverse conditions with one to five microphones and one to three speakers. We also confirm the robustness of FlexIO on CHiME-4 real data.

Table look-up realization of image restoration CNNs has the potential of achieving competitive image quality while being much faster and resource frugal than the straightforward CNN implementation. The main technical challenge facing the LUT-based CNN algorithm designers is to manage the table size without overly restricting the receptive field. The prevailing strategy is to reuse the table for small pixel patches of different orientations (apparently assuming a degree of isotropy) and then fuse the look-up results. The fusion is currently done by average pooling, which we find being ill suited to anisotropic signal structures. To alleviate the problem, we investigate and discuss anisotropic pooling methods to replace naive averaging for improving the performance of the current LUT-realizable CNN restoration methods. First, we introduce the method of generalized median pooling which leads to measurable gains over average pooling. We then extend this idea by learning data-dependent pooling coefficients for each orientation, so that they can adaptively weigh the contributions of differently oriented pixel patches. Experimental results on various restoration benchmarks show that our anisotropic pooling strategy yields both perceptually and numerically superior results compared to existing LUT-realizable CNN methods.

This paper proposes a robust regret control framework in which the performance baseline adapts to the realization of system uncertainty. The plant is modeled as a discrete-time, uncertain linear time-invariant system with real-parametric uncertainty. The performance baseline is the optimal non-causal controller constructed with full knowledge of the disturbance and the specific realization of the uncertain plant. We show that a controller achieves robust additive regret relative to this baseline if and only if it satisfies a related, robust $H_\infty$ performance condition on a modified plant. One technical issue is that the modified plant can, in general, have a complicated nonlinear dependence on the uncertainty. We use a linear approximation step so that the robust additive regret condition can be recast as a standard $\mu$-synthesis problem. A numerical example is used to demonstrate the proposed approach.

Conventional Convolutional Neural Networks (CNNs) in the real domain have been widely used for audio classification. However, their convolution operations process multi-channel inputs independently, limiting the ability to capture correlations among channels. This can lead to suboptimal feature learning, particularly for complex audio patterns such as multi-channel spectrogram representations. Quaternion Convolutional Neural Networks (QCNNs) address this limitation by employing quaternion algebra to jointly capture inter-channel dependencies, enabling more compact models with fewer learnable parameters while better exploiting the multi-dimensional nature of audio signals. However, QCNNs exhibit higher computational complexity due to the overhead of quaternion operations, resulting in increased inference latency and reduced efficiency compared to conventional CNNs, posing challenges for deployment on resource-constrained platforms. To address this challenge, this study explores knowledge distillation (KD) and pruning, to reduce the computational complexity of QCNNs while maintaining performance. Our experiments on audio classification reveal that pruning QCNNs achieves similar or superior performance compared to KD while requiring less computational effort. Compared to conventional CNNs and Transformer-based architectures, pruned QCNNs achieve competitive performance with a reduced learnable parameter count and computational complexity. On the AudioSet dataset, pruned QCNNs reduce computational cost by 50% and parameter count by 80%, while maintaining performance comparable to the conventional CNNs. Furthermore, pruned QCNNs generalize well across multiple audio classification benchmarks, including GTZAN for music genre recognition, ESC-50 for environmental sound classification and RAVDESS for speech emotion recognition.

This work investigates an integrated sensing and edge artificial intelligence (ISEA) system, where multiple devices first transmit probing signals for target sensing and then offload locally extracted features to the access point (AP) via analog over-the-air computation (AirComp) for collaborative inference. To characterize the relationship between AirComp error and inference performance, two proxies are established: the computation-optimal proxy that minimizes the aggregation distortion, and the decision-optimal proxy that maximizes the inter-class separability, respectively. Optimal transceiver designs in terms of closed-form power allocation are derived for both time-division multiplexing (TDM) and frequency-division multiplexing (FDM) settings, revealing threshold-based and dual-decomposition structures, respectively. Experimental results validate the theoretical findings.

Obtaining control barrier functions (CBFs) with large safe sets for complex nonlinear systems and constraints is a challenging task. Predictive CBFs address this issue by using an online finite-horizon optimal control problem that implicitly defines a large safe set. The optimal control problem, also known as the predictive safety filter (PSF), involves predicting the system's flow under a given backup control policy. However, for non-smooth systems and constraints, some key elements, such as CBF gradients and the sensitivity of the flow, are not well-defined, making the current methods inadequate for ensuring safety. Additionally, for control-non-affine systems, the PSF is generally nonlinear and non-convex, posing challenges for real-time computation. This paper considers piecewise affine systems, which are usually control-non-affine, under nonlinear state and polyhedral input constraints. We solve the safety issue by incorporating set-valued generalized Clarke derivatives in the PSF design. We show that enforcing CBF constraints across all elements of the generalized Clarke derivatives suffices to guarantee safety. Moreover, to lighten the computational overhead, we propose an explicit approximation of the PSF. The resulting control methods are demonstrated through numerical examples.

Significant research efforts are currently being dedicated to non-intrusive quality and intelligibility assessment, especially given how it enables curation of large scale datasets of in-the-wild speech data. However, with the increasing capabilities of generative models to synthesize high quality speech, new types of artifacts become relevant, such as generative hallucinations. While intrusive metrics are able to spot such sort of discrepancies from a reference signal, it is not clear how current non-intrusive methods react to high-quality phoneme confusions or, more extremely, gibberish speech. In this paper we explore how to factor in this aspect under a fully unsupervised setting by leveraging language models. Additionally, we publish a dataset of high-quality synthesized gibberish speech for further development of measures to assess implausible sentences in spoken language, alongside code for calculating scores from a variety of speech language models.

This paper presents a novel data-driven stochastic MPC design for discrete-time nonlinear systems with additive disturbances by leveraging the Koopman operator and a distributionally robust optimization (DRO) framework. By lifting the dynamical system into a linear space, we achieve a finite-dimensional approximation of the Koopman operator. We explicitly account for the modeling approximation and additive disturbance error by a mixed stochastic-deterministic tube for the lifted linear model. This ensures the regulation of the original nonlinear system while complying with the prespecified constraints. Stochastic and deterministic tubes are constructed using a DRO and a hyper-cube hull, respectively. We provide finite sample error bounds for both types of tubes. The effectiveness of the proposed approach is demonstrated through numerical simulations.