Papers

Industrial chemical plants often operate under strict data confidentiality constraints, making centralized data-driven process modeling difficult. Federated learning (FL) provides a promising solution by enabling collaborative model training across distributed facilities without sharing raw operational data. This paper proposes a privacy-preserving federated learning framework for distributed chemical process optimization using data collected from multiple geographically separated plants. Each plant locally trains a neural-network-based process model using its own time-series sensor data, while only model parameters are transmitted to a central aggregation server through secure aggregation mechanisms. This design allows cross-plant knowledge sharing while maintaining strict data locality and industrial confidentiality. Experimental evaluation was conducted using process datasets from three independent chemical plants operating under heterogeneous conditions. The results demonstrate rapid convergence of the federated model, with the global mean squared error decreasing from approximately 2369 to below 50 within the first five communication rounds and stabilizing around 35 after 40 rounds. In comparison with local-only training, the proposed federated framework significantly improves prediction accuracy across all plants, while achieving performance comparable to centralized training. The findings indicate that federated learning provides an effective and scalable solution for collaborative industrial analytics, enabling privacy-preserving predictive modeling and process optimization across distributed chemical production facilities.

Supervised contrastive learning (SupCon) is widely used to shape representations, but has seen limited targeted study for audio deepfake detection. Existing work typically combines contrastive terms with broader pipelines; however, the focus on SupCon itself is missing. In this work, we run a controlled study on wav2vec2 XLS-R (300M) that varies (i) similarity in SupCon (cosine vs angular similarity derived from the hyperspherical angle) and (ii) negative scaling using a warm-started global cross-batch queue. Stage 1 fine-tunes the encoder and projection head with SupCon; Stage 2 freezes them and trains a linear classifier with BCE. Trained on ASVspoof 2019 LA and evaluated on ASV19 eval plus ITW and ASVspoof 2021 DF/LA, Cosine SupCon with a delayed queue achieves the best ITW EER (8.29%) and pooled EER (4.44), while angular similarity performs strongly without queued negatives (ITW 8.70), indicating reduced reliance on large negative sets.

This paper presents a systematic verification and validation (V&V) framework for the Evasive Minimum Risk Maneuver (EMRM) feature in autonomous vehicles, addressing a critical gap in existing safety assessment methods. We introduce the first formally integrated pipeline that unifies Hazard Analysis and Risk Assessment (HARA), System-Theoretic Process Analysis (STPA), and Finite State Machine (FSM) modeling into a single traceable workflow specifically designed for EMRM V&V. HARA and STPA are combined through a structured hazard-loss mapping to identify hazards and unsafe control actions; an FSM layer captures hazard-to-loss state transitions that neither method models individually; and the unified framework drives automated scenario generation with measurable parameter-space coverage. Applied to a T-junction EMRM case study, the framework guides 1{,}880 RRT-based simulations spanning ego speed, time-to-collision (TTC), and road friction, uncovering a key physical result: the T-junction geometry gives nearly equal difficulty to stopping and to navigating, so the intermediate mitigation mode occupies only 1.9% of the feasible parameter space. EMRM steering strategies achieve 81% collision-avoidance rate and reduce mean residual impact speed from 18.9~km/h to 9.0~km/h compared with emergency braking alone, while the framework attains 100% hazard, UCA, and parameter-space coverage versus $\leq$1% for traditional methods. These results demonstrate that the integrated HARA-STPA-FSM framework enables high-resolution, traceable EMRM V&V that is not achievable with any single method in isolation.

This paper presents the optimal-control suggestion for congestion on freeways using data assimilation (DA) of distributed fiber-optic sensing (DFOS). To simultaneously maximize throughput and avoid/mitigate congestion, it is necessary to execute optimal control for the current traffic state as active transportation and demand management (ATDM) according to multi-objective optimization with real-time monitoring data. However, optimal control cannot be estimated due to intermittent observed data obtained from conventional sensors. To solve the issue, this paper proposes the ATDM optimal control estimation with DA of DFOS, which can monitor traffic flow in real time without dead zones. Our real-time DA method enables us to estimate the effectiveness of control scenarios by simulation. This paper also provides a method to uniquely determine the optimal-control solution among the Pareto solutions for multi-objective optimization. Throughput and mean speed across the entire road are considered as the objective functions. Variable speed limit (VSL) and inflow control are taken as ATDM examples. Validation results on a Japanese freeway show that (i) the optimal control scenario varies depending on the traffic state, especially congestion level; (ii) optimal control considering VSL alone improves throughput by 5-14% while the improvement rate for mean speed is 0-8%; (iii) throughput and mean speed are improved by 10-15% and 20-30%, respectively when VSL and inflow control are considered. This paper also implies the importance of balance management for the lane occupancy and proactive optimal control before congestion occurs.

As LLM applications grow more complex, developers are increasingly adopting multi-agent architectures to decompose workflows into specialized, collaborative components, introducing structure that constrains agent behavior and exposes useful semantic predictability. Unlike traditional LLM serving, which operates under highly dynamic and uncertain conditions, this structured topology enables opportunities to reduce runtime uncertainty$\unicode{x2015}$yet existing systems fail to exploit it, treating agentic workloads as generic traffic and incurring significant inefficiencies. Our analysis of production traces from an agent-serving platform and an internal coding assistant reveals key bottlenecks, including low prefix cache hit rates, severe resource contention from long-context requests, and substantial queuing delays due to suboptimal scaling. To address these challenges, we propose Pythia, a multi-agent serving system that captures workflow semantics through a simple interface at the serving layer, unlocking new optimization opportunities and substantially improving throughput and job completion time over state-of-the-art baselines.

Contact variability, sensing uncertainty, and external disturbances make grasp execution stochastic. Expected-quality objectives ignore tail outcomes and often select grasps that fail under adverse contact realizations. Risk-sensitive POMDPs address this failure mode, but many use particle-filter beliefs that scale poorly, obstruct gradient-based optimization, and estimate Conditional Value-at-Risk (CVaR) with high-variance approximations. We instead formulate grasp acquisition as variational inference over latent contact parameters and object pose, representing the belief with a differentiable Gaussian mixture. We use Gumbel-Softmax component selection and location-scale reparameterization to express samples as smooth functions of the belief parameters, enabling pathwise gradients through a differentiable CVaR surrogate for direct optimization of tail robustness. In simulation, our variational neural belief improves robust grasp success under contact-parameter uncertainty and exogenous force perturbations while reducing planning time by roughly an order of magnitude relative to particle-filter model-predictive control. On a serial-chain robot arm with a multifingered hand, we validate grasp-and-lift success under object-pose uncertainty against a Gaussian baseline. Both methods succeed on the tested perturbations, but our controller terminates in fewer steps and less wall-clock time while achieving a higher tactile grasp-quality proxy. Our learned belief also calibrates risk more accurately, keeping mean absolute calibration error below 0.14 across tested simulation regimes, compared with 0.58 for a Cross-Entropy Method planner. We provide code, simulation assets, and a dataset of 243 force-closed grasps at the following link: www.github.com/coenwerem/vnb-grasp.

Current pedestrian crossing signals operate on fixed timing without adjustment to pedestrian behavior, which can leave vulnerable road users (VRUs) such as the elderly, disabled, or distracted pedestrians stranded when the light changes. We introduce No Pedestrian Left Behind (NPLB), a real-time adaptive traffic signal system that monitors VRUs in crosswalks and automatically extends signal timing when needed. We evaluated five state-of-the-art object detection models on the BGVP dataset, with YOLOv12 achieving the highest mean Average Precision at 50% (mAP@0.5) of 0.756. NPLB integrates our fine-tuned YOLOv12 with ByteTrack multi-object tracking and an adaptive controller that extends pedestrian phases when remaining time falls below a critical threshold. Through 10,000 Monte Carlo simulations, we demonstrate that NPLB improves VRU safety by 71.4%, reducing stranding rates from 9.10% to 2.60%, while requiring signal extensions in only 12.1% of crossing cycles.

It has long been of interest to understand when control and observer-design methodologies developed for linear dynamical systems can provide local guarantees when applied to nonlinear systems. However, beyond establishing local stability, it is also important to characterize the region of attraction and performance in terms of system parameters and design gains–something that linearization-based approaches cannot achieve. For instance, although increasing an observer gain improves the convergence rate in the linear setting (as in backstepping design), the relationship between this gain and the convergence rate in the nonlinear setting need not be monotonic; understanding these dependencies is therefore crucial for reliable operation. In light of this discussion, we consider the one-dimensional quasilinear heat equation with state-dependent diffusivity, and design a boundary-output observer based on the backstepping design for a linear heat equation with constant diffusivity. Viewing the quasilinear system as a perturbation of the linear one, we establish exponential stability of the origin for the estimation error dynamics in $H^1$, with an estimate of the region of attraction depending on the system parameters, observer gains, and the mismatch between the nonlinear diffusivity and the constant design diffusivity. Importantly, the estimation error converges to zero rather than merely to a neighborhood scaling with this mismatch, even though, in contrast to backstepping-based stabilization of nonlinear PDEs, the mismatch need not decay along trajectories and may remain bounded away from zero, acting as a persistent state-dependent multiplicative perturbation. A technical challenge was to perform a sufficiently-fine Lyapunov analysis that does not yield overly conservative results such as mere boundedness of the estimation error.

Federated inference enhances LLM performance in edge computing through weighted averaging of distributed model predictions. However, autoregressive LLM inference requires frequent full-model forward passes across workers, severely limiting decoding throughput. Distributed deployment further aggravates this due to a communication bottleneck: each worker must transmit full token probability distributions per draft token, dominating end-to-end latency. To address these challenges, we introduce speculative decoding to enable parallel LLM processing and propose a top-K compressed transmission scheme with two server-side reconstruction strategies. We theoretically analyze the robustness of our method in terms of local reconstruction error, aggregation bias, and acceptance-rate bias, and derive corresponding bounds. Experiments demonstrate that our scheme achieves high generation fidelity while significantly reducing communication overhead.

Open, unclassified research on secure autonomy is constrained by limited access to operational platforms, contested communications infrastructure, and representative adversarial test conditions. This paper presents a threat-oriented digital twinning methodology for cybersecurity evaluation of learning-enabled autonomous platforms. The approach is instantiated as an open-source, modular twin of a representative autonomy stack with separated sensing, autonomy, and supervisory-control functions; confidence-gated multi-modal perception; explicit command and telemetry trust boundaries; and runtime hold-safe behavior. The contribution is methodological: a reproducible design pattern that translates threat analysis into observable, controllable tests for spoofing, replay, malformed-input injection, degraded sensing, and adversarial ML stress. Although the implemented proxy is ground based, the architecture is intentionally framed around stack elements shared with UAV and space systems, including constrained onboard compute, intermittent or high-latency links, probabilistic perception, and mission-critical recovery behavior. The result is an implementable research scaffold for dependable and secure autonomy studies across UAV and space domains.

This paper considers a shifted dissipativity property of synchronous generators (SGs) with the aim of establishing stability of the synchronous orbit of a power network. Existing shifted-passivity analyses are formulated in a reference frame attached to the rotor and therefore suppress the problem of rotor-angle alignment in a multimachine network. The analysis takes two steps to reconcile angle-modulated power sources with dissipation. First, we define a weaker shifted dissipativity property whose supply rate retains a cross term between rotor-angle and terminal-voltage errors. We show that an SG connected to an infinite bus is locally asymptotically stable if it satisfies this property. Second, for networks of multiple SGs interconnected through dynamic transmission lines and static loads, we introduce a consensus-like rotor-angle feedback that consolidates the contributions of the distributed angle-modulated power sources into a single effective channel. This channel is then dominated by the dissipative dynamics of an auxiliary state variable representing the phase of the synchronous orbit, thereby establishing local orbital stability. Numerical examples demonstrate the feasibility of the shifted dissipativity condition.

Designing Doppler-resilient unimodular discrete phase-coded waveforms (DPWs) with low delay-Doppler sidelobes is critical for multiple-input multiple-output (MIMO) radar. Existing block coordinate descent (BCD) methods suffer from high computational cost for designing long sequences or large waveform sets. Meanwhile, learning-based alternatives such as the soft-quantization network (SQN) only address correlation optimization in the delay domain, without considering ambiguity function (AF) optimization in the joint delay-Doppler domain. To address these issues, this paper proposes a novel Doppler-resilient DPW design framework, termed SQNGD, for joint transmit-receive optimization that simultaneously optimizes the auto-AF, cross-AF (CAF), and signal-to-noise ratio loss (SNRL) under unimodular constraints. To solve the multi-objective optimization problem (MOOP), a joint transmit-receive design and an alternating optimization strategy are developed. The transmit waveforms are optimized via soft-quantization-based differentiable parameterization, while the receive filters are updated by gradient descent (GD) with an energy constraint and SNRL penalty. An FFT-accelerated evaluation of the AF and CAF is further incorporated, reducing the optimization time by 1.9x - 11x compared with the state-of-the-art (SOTA) majorization-minimization-coordinate descent (MMCD) method. Numerical results show that SQNGD achieves a peak sidelobe level (PSL) of approximately -43 dB over the Doppler range [-0.5,0.5] and -31 dB over [-600,600], respectively, outperforming MMCD by 5.85 dB and 3.45 dB, while maintaining the same SNRL of 0.5 dB.

Recent advancements in large audio language models have extended Chain-of-Thought (CoT) reasoning into the auditory domain, enabling models to tackle increasingly complex acoustic and spoken tasks. To elicit and sustain these extended reasoning chains, the prevailing paradigm – driven by the success of text-based reasoning models – overwhelmingly relies on Reinforcement Learning with Verified Rewards (RLVR). However, as models are strictly optimized to distill rich, continuous auditory contexts into isolated, verifiable text labels, a fundamental question arises: are we fostering true audio intelligence, or merely reducing a continuous sensory medium into a discrete puzzle? We identify this as the "verifiable reward trap." While RLVR yields remarkable scores on standardized objective benchmarks, it systematically degrades the real-world conversational feel of audio models. By prioritizing isolated correctness over acoustic nuance, RLVR reduces dynamic interactions to mechanical "answering machines," severely compromising prosodic naturalness, emotional continuity, and user immersion, particularly in long-turn dialogues. To bridge the gap between mechanical objective verification and genuine sensory empathy, we introduce Step-Audio-R1.5, marking a paradigm shift toward Reinforcement Learning from Human Feedback (RLHF) in audio reasoning. Comprehensive evaluations demonstrate that Step-Audio-R1.5 not only maintains robust analytical reasoning but profoundly transforms the interactive experience, redefining the boundaries of deeply immersive long-turn spoken dialogue.

Foundation segmentation models such as the Segment Anything Model (SAM) have demonstrated strong generalization across natural images; however, their robustness under clinically realistic medical imaging domain shifts remains insufficiently quantified. We present a systematic slice-level robustness audit of SAM (ViT-B) for spleen segmentation in abdominal CT using 1,051 nonempty slices from 41 volumes in the Medical Segmentation Decathlon. A standardized ground-truth-derived bounding-box protocol was used to isolate encoder robustness from prompt uncertainty. Controlled perturbations simulating inter-scanner variability, including Gaussian noise, blur, contrast scaling, gamma correction, and resolution mismatch, were applied across ten conditions. The clean baseline achieved a mean Dice score of 0.9145 (95% CI: [0.909, 0.919]) with a failure rate of 0.67%. Across all perturbations, the absolute mean {\Delta}Dice remained below 0.01. Paired Wilcoxon signed-rank tests with Benjamini-Hochberg false discovery rate correction identified statistically significant but small-magnitude changes under selected conditions, while McNemar analysis showed no significant increase in failure probability. These findings indicate that SAM exhibits stable segmentation behavior under moderate CT domain shifts, supporting its role as a robust foundation baseline for medical image segmentation research. As health digital twins increasingly incorporate foundation segmentation models for anatomical modeling and organ-level monitoring, formal characterization of robustness under real-world imaging variability is a necessary step toward trustworthy deployment.

Unaddressed pain in neonates can lead to adverse effects, including delayed development and slower weight gain, emphasising the need for more objective and reliable pain assessment methods. Hence, automated methods using behavioural and physiological pain indicators have been developed to aid healthcare professionals in the Neonatal ICU. Traditional contact-based methods for physiological parameter estimation are unsuitable for long-term monitoring and increase the risk of spreading diseases like COVID-19. We introduce a novel approach using remote photoplethysmography (rPPG) to estimate pulse signals in a non-contact manner and employ them for neonatal pain detection. The temporal signals acquired from regions-of-interest (ROIs) affected by skin deformations may exhibit lower quality and provide erroneous rPPG signals. Therefore, we incorporated a quality parameter to select the temporal signals obtained from ROIs that are least affected by skin deformations. Further, we employed signal-to-noise ratio as a fitness parameter to extract the rPPG signal corresponding to the clip that is least affected by noise. Experimental findings demonstrate that the rPPG signals provide useful information for neonatal pain detection, and signals extracted from the blue colour channel outperform those extracted from other colour channels. We also show that combining rPPG and audio features provides better results than individual modalities.

We propose a human in the loop approach for black-box testing of Functional Mock-up Units (FMUs) using Large Language Models (LLMs). The goal is to reduce the manual effort in defining test scenarios for dynamic simulation models and to improve the interpretability of results. The approach takes the functional and interface specifications of an FMU as input, and prompts an LLM to generate structured scenario goals in Given-When-Then format that define the initial input conditions of the simulation, a possible change in those conditions, and the expected output behaviour of the system against those changes. The corresponding scenario plans specify input patterns and add assertion oracles that describe expected output patterns defined in scenario goals. The approach generates a complete input time series for the scenario plans, runs the FMU simulation, and evaluates assertions on the recorded outputs. It produces human-readable logs and plots that show statistics for each scenario with overlays, aggregate pass rates, and per-goal outcomes. The generated scenarios and results are stored for evaluation and later re-execution. We evaluate the approach on a Lube Oil Cooling system and discuss design choices that make the approach practical for everyday use. Results suggest that LLM-assisted scenario generation can facilitate automatic test design and verification of dynamic simulation models.

The joint training of speech enhancement and speaker embedding networks for speaker recognition is widely adopted under noisy acoustic environments. While effective, this paradigm often fails to leverage the generalization and robustness benefits inherent in large-scale speech enhancement pre-training. Moreover, maintaining the speaker information in the denoised speech is not an explicit objective of the speech enhancement process. To address these limitations, we proposed a scalable UNet-based Fusion framework (UF-EMA) that considers the noisy and enhanced speech as a multi-channel input, thereby enabling the speaker encoder to exploit speaker information effectively. In addition, an Exponential Moving Average strategy is applied to a speaker encoder pre-trained on clean speech to mitigate overfitting and facilitate a smooth transition from clean to noisy conditions. Experimental results on multiple noise-contaminated test sets showcase the superiority of the proposed approach.

Non-invasive electrophysiology lacks methods that accurately reconstruct whole-brain spatiotemporal dynamics while incorporating individual cortical geometry, leaving current electroencephalography and magnetoencephalography source imaging limited by simplistic or biologically implausible priors. Here, we show that embedding participant-specific Geometric Basis Functions (GBFs), eigenmodes derived from each individual's cortical surface, provides a powerful anatomic constraint that resolves the inverse problem and improves reconstruction fidelity. The method reconstructs neural sources as linear combinations of geometric basis functions, thereby aligning source estimates with the geometric organization of neural dynamics. We validate GBF across the Meta-Source Benchmark, task-evoked data, resting-state networks, intracranial stimulation, and epilepsy data. The results demonstrate that GBF yields high localization accuracy and captures fast spatiotemporal dynamics consistent with anatomical pathways. These findings suggest that both spontaneous and evoked whole-brain activity can be described by hundreds of geometric modes, providing a compact yet accurate representation of neural sources. By linking cortical geometry to electrophysiological dynamics, GBF offers a versatile source imaging tool for both scientific and clinical applications.

Recent audio-aware large language models (ALLMs) have demonstrated strong capabilities across diverse audio understanding and reasoning tasks, but they still frequently produce hallucinated or overly confident outputs. While uncertainty estimation has been extensively studied in text-only LLMs, it remains largely unexplored for ALLMs, where audio-conditioned generation introduces additional challenges such as perceptual ambiguity and cross-modal grounding. In this work, we present the first systematic empirical study of uncertainty estimation in ALLMs. We benchmark five representative methods, including predictive entropy, length-normalized entropy, semantic entropy, discrete semantic entropy, and P(True), across multiple models and diverse evaluation settings spanning general audio understanding, reasoning, hallucination detection, and unanswerable question answering. Our results reveal two key findings. First, semantic-level and verification-based methods consistently outperform token-level baselines on general audio reasoning benchmarks. Second, on trustworthiness-oriented benchmarks, the relative effectiveness of uncertainty methods becomes notably more model- and benchmark-dependent, indicating that conclusions drawn from general reasoning settings do not straightforwardly transfer to hallucination and unanswerable-question scenarios. We further explore uncertainty-based adaptive inference as a potential downstream application. We hope this study provides a foundation for future research on reliable, uncertainty-aware audio-language systems.

Mobile robots that move between outdoor and indoor environments still struggle with consistent positioning. Satellite-based and terrestrial ranging each work well in their home domains, but combining them at the raw measurement level has received little attention, and the building boundary is precisely where both classes degrade. This paper reports preliminary observations from the HYMN dataset, which time-synchronizes raw measurements from GNSS, Ultra-Wideband (UWB), WiFi Fine Time Measurement (FTM), and Bluetooth Low Energy (BLE) against millimeter-level ground truth in an industrial setting. Per-zone measurement availability and ranging-residual behavior are characterised. The two technology classes turn out to be complementary, and the indoor-outdoor transition is where their weaknesses overlap. The dataset is publicly available.

In digital sliding mode control implementations, discretization-induced chattering and inter-sample blindness can severely degrade the closed-loop performance, especially in case of fast perturbations. This paper addresses these challenges for a discrete-time implementation of the super-twisting sliding mode controller. Building upon recent results on barrier-function-modulated super-twisting algorithms, a nested architecture employing multiple barriers is discretized using an eigenvalue-based exact matching approach. The resulting discrete-time controller preserves the adaptive and robustness properties established in continuous time, while ensuring consistent stability behavior at the sampling level. The proposed framework is validated through numerical simulations. The results highlight the effectiveness of multi-layer barrier adaptation for discrete-time sliding mode control applications.

Unbalanced three-phase systems still lack a compact phasor-domain representation of power that makes the joint cross-phase voltage-current structure explicit while remaining consistent with established apparent-power definitions. This paper addresses that point through a complex-vector power (CVP) formulation for sinusoidal steady-state operation. The proposed CVP is an element of a direct-sum space whose constituents are the conventional complex-power scalar and a cross-phase unbalance (CPU) vector obtained from the cross product of the voltage and current phasor vectors. The scalar constituent retains the usual active and reactive powers, whereas the CPU vector characterizes their antisymmetric cross-phase relation. Apparent power is thereby separated into intraphase and cross-phase contributions, and its norm is preserved under the power-invariant Fortescue transformation. The formulation is extended to three-phase four-wire systems by introducing equivalent coordinates that preserve the effective apparent-power norm for the chosen voltage reference. Only standard complex numbers and matrices are required. Numerical examples identify operating conditions in which a non-negligible part of the apparent-power structure is associated with cross-phase unbalance and is not displayed by the active- and reactive-power pair alone. A proof-of-concept shunt-compensation study further illustrates how the CPU descriptor can be incorporated into reproducible current references. The proposed formulation therefore provides a compact phasor-based description of unbalance that complements established scalar apparent-power representations.

This paper studies the distributed adaptiveestimation problems for stochastic large regression modelswith an infinite number of parameters. By constructing a re-cursive local cost function, we propose a novel distributedrecursive least squares algorithm to estimate the unknownsystem parameters, where the growth rate of regressors'dimension is characterized by a non-decreasing positivefunction. The almost sure convergence of the proposedalgorithm is established under a cooperative excitationcondition, which incorporates the temporal information andthe spatial information to reflect the cooperative effectamong multiple agents. Moreover, we analyze the predic-tion error by establishing the asymptotic upper boundof the accumulated regret without any excitation condi-tions. The main difficulty of theoretical analysis lies in howto analyze properties of the product of non-independentand non-stationary random matrices, whose dimensionschange over time simultaneously. Some techniques, suchas stochastic Lyapunov function, double-array martingaletheory and algebraic graph theory, are employed to dealwith the above issue. Our theoretical results are derivedwithout imposing independence or stationarity assump-tions on the regression vectors, thereby not excluding thecorrelated feedback signals.

Metasurfaces' ability to control electromagnetic wave propagation has led to a rapid paradigm shift in wireless operation. These metasurfaces are often called reconfigurable intelligent surfaces (RISs) due to active tuning elements distributed across the meta-atoms comprising the metasurface array. However, each of these dynamic meta-atoms requires additional DC power lines and biasing circuitry for active tuning. Additionally, achieving polarization diverse operations using compact metasurface configurations is challenging due to the complexity involved in polarization detection. To address these limitations, we propose a passive metasurface array architecture that is both polarization sensitive and capable of altering radiation patterns with frequency diversity. In particular, we designed a polarization-sensitive meta-atom model with added randomness in the scattering behavior and extended it to a polarization-diverse-frequency-selective array. By capturing the electric fields scattered off from the metasurface, we can numerically acquire the polarization information of the incoming signal. The proposed polarization-diverse array can simplify the polarization measurement techniques and may find its application in polarization sensitive sensing and imaging operations.

Commercial TTS systems produce near-native Indic audio, but the best open-source bases (Chatterbox, Indic Parler-TTS, IndicF5) trail them on measured phonological dimensions, and the most widely adopted multilingual base (Chatterbox, 23 languages) does not even tokenise Telugu or Tamil. We ask: what is the minimum intervention that brings such a non-Indic-native base to commercial-class output on Telugu, Tamil, and Hindi, without training a new acoustic decoder and without any commercial TTS training data? We combine three pieces: (1) BUPS, a Brahmic Unified Phoneme Space that deterministically romanises seven Indic scripts to ISO-15919 so Chatterbox's Latin tokeniser can process them; (2) a LoRA adapter on only the text-token predictor (Chatterbox's t3), trained on ~1,220h of licensed Indic audio with a Hindi-proxy language_id; (3) a voice-prompt recovery recipe – an 8-11s same-language reference clip plus three sampling overrides (exaggeration 0.7, temperature 0.6, min_p 0.1; "Config B") – that recovers commercial-class acoustic output with no acoustic-decoder training. On Hindi, the LoRA regresses accuracy and we instead use vanilla Chatterbox + Config B, giving a two-branch deployment. Evaluated on 10-utterance pilot sets with the companion PSP benchmark, Praxy Voice matches or slightly leads commercial baselines: 26.7% retroflex collapse on Telugu (vs Sarvam Bulbul 33.3%), 71% Tamil-zha collapse (vs commercial trio's 86%), 0.025 LLM-WER on Hindi (tied with Cartesia Sonic-3). For intra-sentential code-mix we add a third branch (IndicF5 + native-script transliteration) that drops code-mix LLM-WER from 0.80-0.85 to 0.14-0.27 across Hi/Te/Ta. We release R6 LoRA weights (Apache-2.0), inference code and router (MIT), and a Gradio demo.

In this paper, a broadband 1-bit coding metasurface-based reconfigurable intelligent surface (RIS) is presented. The unit cell of the metasurface consists of a wide dipole modified with interdigital capacitors and loaded with an SMP 1340-040LF PIN diode. The proposed element offers cell miniaturization and a stable angular response. A phase difference of 180$\degree \pm$ 30$\degree$ is achieved for a frequency range of 4.85-6.05 GHz between the ON and OFF states for the normal incidence of the TE polarized wave, whereas it provides a fairly stable response with reflection loss of less than 3 dB and phase difference of 180$\degree$ $\pm$ 50$\degree$ for oblique incidence up to 45$\degree$. The RF is isolated from the DC on the bias lines using properly designed butterfly-shaped radial stubs. Using this unit cell, a prototype with an array of 16 $\times$ 10 elements is constructed. A low-cost microcontroller-based control circuit is designed, which can be plugged-in for biasing the PIN diodes of such array. The theoretically calculated and full-wave simulated radiation patterns of the array are validated using experiments inside anechoic chamber. Furthermore, the capability of the RIS for non-line of sight (NLOS) user equipment (UE) localization and robust uplink communication is demonstrated using LTE communication framework. This shows great potential of our RIS for applications, such as in unmanned aerial vehicle (UAV) localization and its uplink communication at NLOS or extended range.

Efficient GPU execution of convolution operators is governed by memory-access efficiency, on-chip data reuse, and execution mapping rather than arithmetic throughput alone. This paper presents a controlled operator-level study of CUDA kernel optimization for the depthwise convolution used in Structured State Space Model Convolutional Diagonal (S4ConvD), together with a cloud-compatible, counter-free performance analysis methodology. The operator, model, dataset, and training configuration are fixed, and only the CUDA kernel implementation is varied. The evaluated CUDA kernels comprise naive, global-memory-coalesced, shared-memory cache-blocked, and warp-tiled variants, covering forward, input-gradient, and weight-gradient execution paths under steady-state training conditions. Performance is characterized using a counter-free methodology that combines CUDA-event timing, execution-path decomposition, analytically derived memory-traffic modeling, effective-bandwidth estimation, and roofline analysis. This enables profiling-like architectural insights without requiring hardware performance counters or privileged profiling access. The warp-tiled kernel reduces convolution runtime by $3.26\times$ relative to the naive CUDA baseline, while end-to-end training speedup reaches $1.29\times$. A PyTorch implementation is used separately for numerical validation and runtime context, but is not treated as a controlled architectural baseline. Forward and input-gradient paths benefit substantially from improved locality and on-chip data reuse, whereas the reduction-dominated weight-gradient path remains the primary bottleneck. The results demonstrate that meaningful architecture-level GPU kernel analysis can be performed reproducibly in restricted cloud environments, even without access to hardware performance counters.

Direction-of-arrival (DOA) estimation is an important task in microphone array processing and many downstream applications. The steered response power with phase transform (SRP-PHAT) method has been widely adopted for DOA estimation in recent years. However, accurate SRP-PHAT estimation in 3D scenarios requires evaluating steering responses over thousands of candidate directions, severely limiting real-time performance on resource-constrained platforms. This challenge becomes even more critical for planar arrays, which are widely used in robotics due to their structural simplicity. Motivated by the fact that azimuth estimation is usually more reliable than elevation estimation for most arrays, we propose ASAP, an azimuth-priority strip-based search approach to planar microphone array DOA estimation in 3D. In the first stage, ASAP performs coarse-to-fine region contraction within azimuthal strips to lock azimuth angles while retaining multiple maxima through spherical caps. In the second stage, it refines elevation along the great-circle arc between two close candidates. Extensive simulations and real-world experiments validate the efficiency and merits of the proposed method over existing approaches.

To establish empathy with machines, it is essential to fully understand human emotional changes. However, research in multimodal emotion recognition often overlooks one problem: individual expressive traits vary significantly, which means that different people may express emotions differently. In our daily lives, we can see this. When communicating with different people, some express "happiness" through their facial expressions and words, while others may hide their happiness or express it through their actions. Both are expressions of 'happiness,' but such differences in emotional expression are still too difficult for machines to distinguish. Current emotion recognition remains at a 'static' level, using a single recognition model to identify all emotional styles. This "simplification" often affects the recognition results, especially in multi-turn dialogues. To address this problem, this paper introduces a novel Multi-Level Speaker Adaptive Network (ML-SAN), which, specifically, effectively addresses the challenge of speaker identity information confusion. ML-SAN does not simply assign a speaker's ID after recognition; instead, it employs a three-stage adaptive process: First, Input-level Calibration uses Feature-Level Linear Modulation (FiLM) to adjust the raw audio and visual features into a neutral space unrelated to the speaker. Then, Interaction-level Gating re-adjusts the trust level for each modality (e.g., voice or facial features) based on the speaker's identity information. Finally, Output-level Regularization maintains the consistency of speaker features in the latent space. Tests on the MELD and IEMOCAP datasets show that our model (ML-SAN) achieves better results, performs exceptionally well in handling challenging tail sentiment categories, and better addresses the diversity of speakers in real-world scenarios.

While it is generally understood that zeroth-order (ZO) algorithms have an extra dependency on their number of iterations for any choice of parameters, compared to their first-order (FO) counterparts, in this work, we show that under several conditions, in expectation, ZO methods do not suffer from extra dimension dependencies in their convergence rates with respect to their FO counterparts. We look at optimisation algorithms from the dynamical systems perspective and analyse the conditions under which one can formulate the average of a ZO algorithm as the average of its FO counterpart with bounded perturbations with values dependent on design parameters. Then, using input-to-state stability properties, we show ZO methods follow the same decay rate as their FO counterparts and converge to a neighbourhood of the fixed point of FO methods, where its radius depends on the bound of the norm of the perturbations, which can be made arbitrarily small. The theoretical findings are illustrated via numerical examples.