Papers

Vision-and-Language Navigation (VLN) requires an agent to navigate through complex unseen environments based on natural language instructions. However, existing methods often struggle to effectively capture key semantic cues and accurately align them with visual observations. To address this limitation, we propose Beyond Textual Knowledge (BTK), a VLN framework that synergistically integrates environment-specific textual knowledge with generative image knowledge bases. BTK employs Qwen3-4B to extract goal-related phrases and utilizes Flux-Schnell to construct two large-scale image knowledge bases: R2R-GP and REVERIE-GP. Additionally, we leverage BLIP-2 to construct a large-scale textual knowledge base derived from panoramic views, providing environment-specific semantic cues. These multimodal knowledge bases are effectively integrated via the Goal-Aware Augmentor and Knowledge Augmentor, significantly enhancing semantic grounding and cross-modal alignment. Extensive experiments on the R2R dataset with 7,189 trajectories and the REVERIE dataset with 21,702 instructions demonstrate that BTK significantly outperforms existing baselines. On the test unseen splits of R2R and REVERIE, SR increased by 5% and 2.07% respectively, and SPL increased by 4% and 3.69% respectively. The source code is available at https://github.com/yds3/IPM-BTK/.

The rapid advancement of generative models has enabled highly realistic audio deepfakes, yet current detectors suffer from a critical bias problem, leading to poor generalization across unseen datasets. This paper proposes Artifact-Focused Self-Synthesis (AFSS), a method designed to mitigate this bias by generating pseudo-fake samples from real audio via two mechanisms: self-conversion and self-reconstruction. The core insight of AFSS lies in enforcing same-speaker constraints, ensuring that real and pseudo-fake samples share identical speaker identity and semantic content. This forces the detector to focus exclusively on generation artifacts rather than irrelevant confounding factors. Furthermore, we introduce a learnable reweighting loss to dynamically emphasize synthetic samples during training. Extensive experiments across 7 datasets demonstrate that AFSS achieves state-of-the-art performance with an average EER of 5.45%, including a significant reduction to 1.23% on WaveFake and 2.70% on In-the-Wild, all while eliminating the dependency on pre-collected fake datasets. Our code is publicly available at https://github.com/NguyenLeHaiSonGit/AFSS.

Markerless biomechanics increasingly relies on 3D skeletal keypoints extracted from video, yet downstream biomechanical mappings typically treat these estimates as deterministic, providing no principled mechanism for frame-wise quality control. In this work, we investigate predictive uncertainty as a quantitative measure of confidence for mapping 3D pose keypoints to 3D anatomical landmarks, a critical step preceding inverse kinematics and musculoskeletal analysis. Within a temporal learning framework, we model both uncertainty arising from observation noise and uncertainty related to model limitations. Using synchronized motion capture ground truth on AMASS, we evaluate uncertainty at frame and joint level through error–uncertainty rank correlation, risk–coverage analysis, and catastrophic outlier detection. Across experiments, uncertainty estimates, particularly those associated with model uncertainty, exhibit a strong monotonic association with landmark error (Spearman $\rho \approx 0.63$), enabling selective retention of reliable frames (error reduced to $\approx 16.8$ mm at 10% coverage) and accurate detection of severe failures (ROC-AUC $\approx 0.92$ for errors $>50$ mm). Reliability ranking remains stable under controlled input degradation, including Gaussian noise and simulated missing joints. In contrast, uncertainty attributable to observation noise provides limited additional benefit in this setting, suggesting that dominant failures in keypoint-to-landmark mapping are driven primarily by model uncertainty. Our results establish predictive uncertainty as a practical, frame-wise tool for automatic quality control in markerless biomechanical pipelines.

Voice anonymisation is used to conceal voice identity while preserving linguistic content. Even if anonymisation seems strong, non-timbral cues such as accent that remain post-anonymisation can help re-identification and reveal sensitive socio-demographic traits. We report a study of residual accent information involving multiple anonymisation systems. We highlight the role of accent using speaker verification, accent verification, and accent classification using a set of embeddings focusing on timbral, non-timbral and accent-related information and show the extent to which related cues facilitate reidentification post anonymisation. Results show that, while some systems are robust to reidentification attempts using accent cues, others leave residual, speaker-dependent, accentrelated cues which can be used to reveal the voice identity. We also highlight accent-dependent variation in anonymisation performance, raising fairness concerns, and show that a system with characterlevel conditioning can help obfuscate identity-revealing accent cues, reducing accent-identification accuracy by 68% on average and improving overall anonymisation performance by 11% relative.

Multimodal Emotion Recognition in Conversations (MERC) aims to predict speakers' emotional states in multi-turn dialogues through text, audio, and visual cues. In real-world settings, conversation scenarios differ significantly in speakers, topics, styles, and noise levels. Existing MERC methods generally neglect these cross-scenario variations, limiting their ability to transfer models trained on a source domain to unseen target domains. To address this issue, we propose a Dual-branch Graph Domain Adaptation framework (DGDA) for multimodal emotion recognition under cross-scenario conditions. We first construct an emotion interaction graph to characterize complex emotional dependencies among utterances. A dual-branch encoder, consisting of a hypergraph neural network (HGNN) and a path neural network (PathNN), is then designed to explicitly model multivariate relationships and implicitly capture global dependencies. To enable out-of-domain generalization, a domain adversarial discriminator is introduced to learn invariant representations across domains. Furthermore, a regularization loss is incorporated to suppress the negative influence of noisy labels. To the best of our knowledge, DGDA is the first MERC framework that jointly addresses domain shift and label noise. Theoretical analysis provides tighter generalization bounds, and extensive experiments on IEMOCAP and MELD demonstrate that DGDA consistently outperforms strong baselines and better adapts to cross-scenario conversations. Our code is available at https://github.com/Xudmm1239439/DGDA-Net.

Remote photoplethysmography (rPPG) allows for the contactless estimation of physiological signals from facial videos by analyzing subtle skin color changes. However, rPPG signals are extremely susceptible to illumination changes, motion, shadows, and specular reflections, resulting in low-quality signals in unconstrained environments. To overcome these issues, we present a Reliability-Aware Weighted Multi-Scale Spatio-Temporal (WMST) map that models pixel reliability through the suppression of environmental noises. These noises are modeled using different weighting strategies to focus on more physiologically valid areas. Leveraging the WMST map, we develop an SSL contrastive learning approach based on Swin-Unet, where positive pairs are generated from conventional rPPG signals and temporally expanded WMST maps. Moreover, we introduce a new High-High-High (HHH) wavelet map as a negative example that maintains motion and structural details while filtering out physiological information. Here, our aim is to estimate heart rate (HR), and the experiments on public rPPG benchmarks show that our approach enhances motion and illumination robustness with lower HR estimation error and higher Pearson correlation than existing Self-Supervised Learning (SSL) based rPPG methods.

Real-time 30-to-60 fps video frame interpolation on mobile neural processing units (NPUs) requires each synthesized frame within 33.3 ms. We show that mainstream flow-based video frame interpolation faces three structural deployment barriers on mobile NPUs: spatial sampling operators exceed the frame budget or lack hardware support, iterative flow refinement collapses under 8-bit integer post-training quantization, and memory-bound operators dominate the inference graph. ANVIL addresses these barriers by reusing motion vectors from the H.264/AVC decoder to prealign input frames, removing learned optical flow, spatial sampling, and iterative accumulation from the accelerator graph. The remaining residual is refined by a convolution-dominated network composed almost entirely of compute-bound operators. On a Snapdragon 8 Gen 3 device, ANVIL achieves 12.8 ms 1080p inference at 8-bit integer precision; an open-source Android player sustains 28.4 ms median end-to-end latency over 30-minute continuous playback. Per-operator causal analysis identifies quantized accumulation on recurrent flow states as a key mechanism behind integer quantization failure in iterative methods. The current design targets H.264/AVC playback with decoder-exposed motion vectors.

We propose a hybrid diffusion-based augmentation framework to overcome the critical challenge of ultrasound data augmentation in breast ultrasound (BUS) datasets. Unlike conventional diffusion-based augmentations, our approach improves visual fidelity and preserves ultrasound texture by combining text-to-image generation with image-to-image (img2img) refinement, as well as fine-tuning with low-rank adaptation (LoRA) and textual inversion (TI). Our method generated realistic, class-consistent images on an open-source Kaggle breast ultrasound image dataset (BUSI). Compared to the Stable Diffusion v1.5 baseline, incorporating TI and img2img refinement reduced the Frechet Inception Distance (FID) from 45.97 to 33.29, demonstrating a substantial gain in fidelity while maintaining comparable downstream classification performance. Overall, the proposed framework effectively mitigates the low-fidelity limitations of synthetic ultrasound images and enhances the quality of augmentation for robust diagnostic modeling.

Background and Aims: Reproducible external benchmarks for pneumothorax-related lung ultrasound (LUS) AI are scarce, and binary lung-sliding classification may obscure clinically important signs. We therefore developed a manifest-based external benchmark and used it to test both cross-domain generalization and task validity. Methods: We curated 280 clips from 190 publicly accessible LUS source videos and released a reconstruction manifest containing URLs, timestamps, crop coordinates, labels, and probe shape. Labels were normal lung sliding, absent lung sliding, lung point, and lung pulse. A previously published single-site binary classifier was evaluated on this benchmark; challenge-state analysis examined lung point and lung pulse using the predicted probability of absent sliding, P(absent). Results: The single-site comparator achieved ROC-AUC 0.9625 in-domain but 0.7050 on the heterogeneous external benchmark; restricting external evaluation to linear clips still yielded ROC-AUC 0.7212. In challenge-state analysis, mean P(absent) ranked absent (0.504) > lung point (0.313) > normal (0.186) > lung pulse (0.143). Lung pulse differed from absent clips (p=0.000470) but not from normal clips (p=0.813), indicating that the binary model treated pulse as normal-like despite absent sliding. Lung point differed from both absent (p=0.000468) and normal (p=0.000026), supporting its interpretation as an intermediate ambiguity state rather than a clean binary class. Conclusion: A manifest-based, multi-source benchmark can support reproducible external evaluation without redistributing source videos. Binary lung-sliding classification is an incomplete proxy for pneumothorax reasoning because it obscures blind-spot and ambiguity states such as lung pulse and lung point.

We study the stochastic multi-armed bandit (MAB) problem where an underlying network structure enables side-observations across related actions. We use a bipartite graph to link actions to a set of unknowns, such that selecting an action reveals observations for all the unknowns it is connected to. While previous works rely on the assumption that all actions are permanently accessible, we investigate the more practical setting of stochastic availability, where the set of feasible actions (the "activation set") varies dynamically in each round. This framework models real-world systems with both structural dependencies and volatility, such as social networks where users provide side-information about their peers' preferences, yet are not always online to be queried. To address this challenge, we propose UCB-LP-A, a novel policy that leverages a Linear Programming (LP) approach to optimize exploration-exploitation trade-offs under stochastic availability. Unlike standard network bandit algorithms that assume constant access, UCB-LP-A computes an optimal sampling distribution over the realizable activation sets, ensuring that the necessary observations are gathered using only the currently active arms. We derive a theoretical upper bound on the regret of our policy, characterizing the impact of both the network structure and the activation probabilities. Finally, we demonstrate through numerical simulations that UCB-LP-A significantly outperforms existing heuristics that ignore either the side-information or the availability constraints.

Traditional wireless power transfer (WPT) systems are largely limited to 1-D charging pads or 2-D charging surfaces and therefore do not support a truly ubiquitous device-powering experience. Although room-scale WPT based on multimode quasistatic cavity resonance (QSCR) has demonstrated full-volume coverage by leveraging multiple resonant modes, existing high-coverage implementations require obstructive internal conductive structures, such as a central pole. Here, we present a new structure, termed the patched-wall QSCR, that eliminates such internal obstructions while preserving full-volume coverage. By using conductive wall segments interconnected by capacitors, the proposed structure supports two complementary resonant modes that cover both the peripheral and central regions without obstructions within the charging volume. Electromagnetic simulations show that, by selectively exciting these two resonant modes, the proposed structure achieves a minimum power-transfer efficiency of 48.1% throughout the evaluated interior volume of a 54m^3 enclosure while preserving an unobstructed charging space.

Set operations are well understood for convex sets but become considerably more challenging in the non-convex case due to the loss of structural properties in their representation. Constrained polynomial zonotopes (CPZs) offer an effective compromise, as they can capture complex, typically non-convex geometries while maintaining an algebraic structure suitable for further manipulation. Building on this, we propose novel nonlinear encodings that provide sufficient conditions for testing inclusion between two CPZs and adapt them for seamless integration within optimization frameworks.

Synthesizing control barrier functions (CBFs) for complex nonlinear systems remains difficult, particularly when safety constraints have high relative degree with respect to the input. In this work, we study the class of composite port-Hamiltonian systems as a network of interconnected subsystems, and this network admits a natural graph structure. We propose a novel graph-theoretic framework for the CBF synthesis problem. The framework addresses the problem of CBF construction for complex nonlinear systems by offering a relative-degree constraint induced by graph distance and a barrier reshaping rule informed by the graph topology. Furthermore, we perform a feasibility analysis of the reshaped barrier under a bounded input constraint and identify the sufficient conditions for feasibility. Lastly, the proposed framework is validated on an LC ladder network, where the CBF feasibility on a capacitor charge constraint depends on where the input enters the network.

This paper considers a wideband millimeter-wave MIMO system with fully digital transceivers at both the base station and the user equipment (UE), focusing on mobile scenarios. To reduce the baseband processing burden at the UE, we propose a two-stage digital combining architecture, where the received signals are compressed from $K$ antennas to dimension $N_{\mathrm c}$ before baseband processing. The first-stage combining matrix exploits channel geometry and is updated on the beam-coherence timescale, which is longer than the channel coherence time, while the second stage is updated per channel coherence time. We develop a pilot-based channel estimation framework tailored to the proposed two-stage digital combining architecture, leveraging maximum likelihood estimation. Furthermore, we propose a time-domain method that exploits the finite delay spread to reconstruct the full channel from a reduced number of pilot subcarriers. Precoding and combining schemes are designed accordingly, and spectral efficiency expressions with imperfect channel state information are derived. Numerical results show that the proposed time-domain approach outperforms hybrid beamforming while reducing pilot overhead. We further demonstrate that the framework extends to multi-user MIMO and retains its performance advantages. These results highlight the potential of two-stage fully digital transceivers for future wideband systems.

Electrocardiograms (ECGs) are among the most widely available clinical signals and play a central role in cardiovascular diagnosis. While recent foundation models (FMs) have shown promise for learning transferable ECG representations, most existing pretraining approaches treat leads as independent channels and fail to explicitly leverage their strong structural redundancy. We introduce the latent attention masked autoencoder (LAMAE) FM that directly exploits this structure by learning cross-lead connection mechanisms during self-supervised pretraining. Our approach models higher-order interactions across leads through latent attention, enabling permutation-invariant aggregation and adaptive weighting of lead-specific representations. We provide empirical evidence on the Mimic-IV-ECG database that leveraging the cross-lead connection constitutes an effective form of structural supervision, improving representation quality and transferability. Our method shows strong performance in predicting ICD-10 codes, outperforming independent-lead masked modeling and alignment-based baselines.

This letter investigates the channel estimation problem in THz wireless communications where a RIS is employed to assist wireless transmission between different devices. Unlike existing studies, we consider a novel scenario where specific devices are all located in the radiative NF region of the RIS. Meanwhile, we also account for the impact on channel estimation of the random electromagnetic interference occurring at the RIS location. A linear minimum mean-square error estimator is employed, where the estimation error is fully determined by the RIS configuration. Optimizing the RIS involves solving a non-convex problem, which is addressed using an alternating optimization approach based on the diagonally scaled gradient descent algorithm. Numerical results in the THz band highlight the importance of leveraging NF channel statistics over far-field approximations and demonstrate that the proposed estimator achieves substantial improvements in normalized mean-square error compared to existing methods.

Deformable image registration remains a central challenge in medical image analysis, particularly under multi-modal scenarios where intensity distributions vary significantly across scans. While deep learning methods provide efficient feed-forward predictions, they often fail to generalize robustly under distribution shifts at test time. A straightforward remedy is full network fine-tuning, yet for modern architectures such as Transformers or deep U-Nets, this adaptation is prohibitively expensive in both memory and runtime when operating in 3D. Meanwhile, the naive fine-tuning struggles more with potential degradation in performance in the existence of drastic domain shifts. In this work, we propose a registration framework that integrates a frozen pretrained mono-modal registration model with a lightweight adaptation pipeline for multi-modal image registration. Specifically, we employ style transfer based on contrast-agnostic representation generation and refinement modules to bridge modality and domain gaps with instance optimization at test time. This design is orthogonal to the choice of backbone mono-modal model, thus avoids the computational burden of full fine-tuning while retaining the flexibility to adapt to unseen domains. We evaluate our approach on the Learn2Reg 2025 LUMIR validation set and observe consistent improvements over the pretrained state-of-the-art mono-modal backbone. In particular, the method ranks second on the multi-modal subset, third on the out-of-domain subset, and achieves fourth place overall in Dice score. These results demonstrate that combining frozen mono-modal models with modality adaptation and lightweight instance optimization offers an effective and practical pathway toward robust multi-modal registration.

Generalization under manipulation and dataset shift remains a core challenge in forged media detection for AI-driven edge sensing systems. Frozen vision foundation models with linear probes are strong baselines, but most pipelines use default backbone outputs without testing conditioning at the frozen feature interface. We present the first controlled probing study on DINOv3 ConvNeXt and show that, without task-specific fine-tuning, linear probing alone yields competitive forged-media detection performance, indicating that ViT-7B self-supervised distillation transfers to security-critical vision workloads at edge-compatible inference cost. Backbone, head, data, and optimization are fixed while conditioning is varied; LN-Affine, the default ConvNeXt head output, is the natural baseline. On FaceForensics++ c23, five conditioning variants are evaluated under in-distribution testing, leave-one-manipulation-out (LOMO), and cross-dataset transfer to Celeb-DF v2 and DeepFakeDetection. In ConvNeXt-Tiny, conditioning alone changes LOMO mean AUC by 6.1 points and reverses ID-vs-OOD ranking: LN-Affine is strongest on external datasets, while LayerNorm is strongest in-distribution. In ConvNeXt-Base replication, the OOD winner becomes protocol-dependent, and ID-optimal selection still fails as a robust deployment rule. Results show that feature conditioning is a first-order design variable and should be selected with robustness-oriented validation, not ID accuracy alone.

Channel estimation is a central bottleneck in BD-RIS-assisted MIMO systems. The richer inter-element coupling that enables large performance gains also makes training and hardware control substantially harder than in diagonal RIS architectures. Existing estimators either target only cascaded channels or require block-by-block reconfiguration of the BD-RIS interconnections, which is costly and difficult to implement in practice. To overcome this limitation, we propose a pilot-assisted tensor framework for group-connected BD-RIS under a two-timescale protocol, where the scattering structure is designed as a low-rank PARAFAC model with fixed factor matrices. This design keeps the interconnection topology constant across blocks and updates only phase shifts, enabling practical operation without sacrificing estimation quality. Building on this structure, we develop a PARAFAC-based alternating least-squares (PALS) receiver that recovers the individual channels. Numerical results confirm that PALS delivers markedly lower composite-channel NMSE than conventional LS, matches the accuracy of state-of-the-art tensor receivers, and sharply reduces BD-RIS design complexity

Foundation models learn transferable representations, motivating growing interest in their application to wireless systems. Existing wireless foundation models are predominantly based on transformer architectures, whose quadratic computational and memory complexity can hinder practical deployment for large-scale channels. In this work, we introduce WiMamba, a wireless foundation model built upon the recently proposed Mamba architecture, which replaces attention mechanisms with selective state-space models and enables linear-time sequence modeling. Leveraging this architectural advantage combined with adaptive preprocessing, WiMamba achieves scalable and low-latency inference while maintaining strong representational expressivity. We further develop a dedicated task-agnostic, self-supervised pre-training framework tailored to wireless channels, resulting in a genuine foundation model that learns transferable channel representations. Evaluations across four downstream tasks demonstrate that WiMamba matches or outperforms transformer-based wireless foundation models, while offering dramatic latency and memory reductions.

In haptics, guaranteeing stability is essential to ensure safe interaction with remote or virtual environments. One of the most relevant methods at the state-of-the-art is the Time Domain Passivity Approach (TDPA). However, its high conservatism leads to a significant degradation of transparency. Moreover, the stabilizing action may conflict with the device's physical limitations. State-of-the-art solutions have attempted to address these actuator limits, but they still fail to account simultaneously for the power limits of each actuator while maximizing transparency. This work proposes a new damping limitation method based on prioritized dissipation actions. It prioritizes an optimal dissipation direction that minimizes actuator load, while any excess dissipation is allocated to the orthogonal hyperplane. The solution provides a closed-form formulation and is robust in multi-DoF scenarios, even in the presence of actuator and motion anisotropies. The method is experimentally validated using a parallel haptic interface interacting with a virtual environment and tested under different operating conditions.

The complex Gaussian distribution has been widely used as a fundamental spectral and noise model in signal processing and communication. However, its Gaussian structure often limits its ability to represent the diverse amplitude characteristics observed in individual source signals. On the other hand, many existing non-Gaussian amplitude distributions derived from hyperspherical models achieve good empirical fit due to their power-law structures, while they do not explicitly account for the complex-plane geometry inherent in complex-valued observations. In this paper, we propose a new probabilistic model for complex-valued random variables, which can be interpreted as a power-weighted noncentral complex Gaussian distribution. Unlike conventional hyperspherical amplitude models, the proposed model is formulated directly on the complex plane and preserves the geometric structure of complex-valued observations while retaining a higher-dimensional interpretation. The model introduces a nonlinear phase diffusion through a single shape parameter, enabling continuous control of the distributional geometry from arc-shaped diffusion along the phase direction to concentration of probability mass toward the origin. We formulate the proposed distribution and analyze the statistical properties of the induced amplitude distribution. The derived amplitude and power distributions provide a unified framework encompassing several widely used distributions in signal modeling, including the Rice, Nakagami, and gamma distributions. Experimental results on speech power spectra demonstrate that the proposed model consistently outperforms conventional distributions in terms of log-likelihood.

We propose an Expected Free Energy-based acquisition function for Bayesian optimization to solve the joint learning and optimization problem, i.e., optimize and learn the underlying function simultaneously. We show that, under specific assumptions, Expected Free Energy reduces to Upper Confidence Bound, Lower Confidence Bound, and Expected Information Gain. We prove that Expected Free Energy has unbiased convergence guarantees for concave functions. Using the results from these derivations, we introduce a curvature-aware update law for Expected Free Energy and show its proof of concept using a system identification problem on a Van der Pol oscillator. On a two-dimensional benchmark with an oscillatory landscape, our adaptive Expected Free Energy acquisition achieves competitive performance in both regret and mean squared error, unlike the typical acquisition functions that perform well in only one metric.

Integrating grid-forming converters (GFMCs) into grid-following converter (GFLC)-dominated power systems enhances the grid strength, but GFMCs' current-limiting characteristic triggers dynamic switching between constant voltage control (CVC) and current limit control (CLC). This switching feature poses critical transient stability risks to GFLCs, requiring urgent investigation. This paper first develops a mathematical model for this switched system. Then, it derives switching conditions for droop-controlled GFMCs, which are separately GFMC angle-dependent and GFLC angle-dependent. On this basis, the stability boundaries of GFLC within each subsystem are analyzed, and the impact of GFMC switching arising from GFLC angle oscillation is investigated. The findings reveal that the switched system's stability boundary coincides with that of the CLC subsystem. To enhance GFLC's transient stability and ensure GFMC converges to the CVC mode, this paper introduces a virtual fixed d-axis control (VFDC) strategy. Compared with existing methods, this method achieves decoupling and self-stabilization using only local state variables from individual converters. The conclusions are validated through simulations and Controller Hardware-in-the-Loop tests.

Optimal dispatch of energy storage systems (ESSs) in distribution networks involves jointly improving operating economy and voltage security under time-varying conditions and possible topology changes. To support fast online decision making, we develop a topology-aware Reinforcement Learning architecture based on Twin Delayed Deep Deterministic Policy Gradient (TD3), which integrates graph neural networks (GNNs) as graph feature encoders for ESS dispatch. We conduct a systematic investigation of three GNN variants: graph convolutional networks (GCNs), topology adaptive graph convolutional networks (TAGConv), and graph attention networks (GATs) on the 34-bus and 69-bus systems, and evaluate robustness under multiple topology reconfiguration cases as well as cross-system transfer between networks with different system sizes. Results show that GNN-based controllers consistently reduce the number and magnitude of voltage violations, with clearer benefits on the 69-bus system and under reconfiguration; on the 69-bus system, TD3-GCN and TD3-TAGConv also achieve lower saved cost relative to the NLP benchmark than the NN baseline. We also highlight that transfer gains are case-dependent, and zero-shot transfer between fundamentally different systems results in notable performance degradation and increased voltage magnitude violations. This work is available at: https://github.com/ShuyiGao/GNNs_RL_ESSs and https://github.com/distributionnetworksTUDelft/GNNs_RL_ESSs.

Standard LLM-based speech recognition systems process utterances in isolation, limiting their ability to leverage conversational context. Here we study whether multimodal context from prior turns improves LLM-based ASR and how to represent that context efficiently. We find that, after supervised multi-turn training, conversational context mainly helps with the recognition of contextual entities. However, conditioning on raw context is expensive because the prior-turn audio embedding sequence grows rapidly with conversation length. To address this, we propose Abstract Compression, a learned latent abstraction of prior-turn audio that replaces each prior-turn audio sequence with a fixed number of latent tokens while retaining corresponding transcripts explicitly. On both in-domain and out-of-domain test sets, the compression-enabled model recovers part of the gains of raw-context conditioning with a smaller prior-turn audio footprint. We also provide targeted analyses of the compression setup and its tradeoffs.

Fluid antenna (FA) systems offer novel spatial degrees of freedom (DoFs) with the potential for significant performance gains. Compared to existing works focusing solely on optimizing FA positions at discrete time instants, we introduce the concept of continuous-trajectory fluid antenna (CTFA), which explicitly considers the antenna element's movement trajectory across continuous time intervals and incorporates the inherent kinematic constraints present in practical FA implementations. Accordingly, we formulate the total throughput maximization problem in CTFA-aided wireless communication systems, addressing the joint optimization of continuous antenna trajectories in conjunction with the transmit covariance matrices under kinematic constraints. To effectively solve this non-convex problem with highly coupled optimization variables, we develop an iterative algorithm based on block coordinate descent (BCD) and majorization-minimization (MM) principles with the aid of the weighted minimum mean square error (WMMSE) method. Finally, numerical results are presented to validate the efficacy of the proposed algorithms and to quantify the substantial total throughput advantages afforded by the conceived CTFA-aided system compared to conventional fixed-position antenna (FPA) benchmarks and alternative approaches employing simplified trajectories.

Existing approaches for battery health forecasting often rely on extensive cycling histories and continuously monitored cells. In contrast, many real-world scenarios provide only sparse information, e.g. a single diagnostic cycle. In our study, we investigate state of health (SoH)- and remaining useful life (RUL) estimation of previously unseen lithium-ion cells, relying on cycling data from begin of life (BOL) to end of life (EOL) of multiple similar cells by using the publicly available Oxford battery aging dataset. The estimator applies incremental capacity analysis (ICA)-based feature extraction in combination with data-efficient regression methods. Particular emphasis is placed on a multi-model Gaussian process regression ensemble approach (GPRn), which also provides uncertainty quantification. Due to a rather cell invariant behaviour, the mapping of ICA features to SoH estimation is highly precise and points out a normalized mean absolute error (NMAE) of 1.3%. The more cell variant mapping to RUL estimation is challenging, reflecting in a NMAE of 5.3%. Using the estimation results, a RUL monitoring strategy is derived. The objective is to safely operate a battery cell from BOL to EOL by only taking sparse diagnostic measurements. On average, only four diagnostic measurements are required during a cell's lifetime of 3300 to 5000 cycles.

This paper proposes a movable-antenna-based index modulation (MA-IM) framework that exploits the spatial mobility of a single reconfigurable antenna to create additional information-bearing dimensions for next-generation wireless systems. By discretizing the continuous movable region into a dense set of candidate sampling points and selecting representative anchors for indexing, the proposed framework converts spatial degrees of freedom into a practical modulation resource. Building on this framework, we develop a family of anchor-selection strategies with different levels of channel awareness, including geometry-based, SNR-based, max–min channel-domain, and joint constellation-aware designs. For the resulting MA-IM schemes, joint maximum-likelihood (ML) detectors are derived, along with a low-complexity two-stage detector, and unified analytical upper bounds on the average bit error probability (ABEP) are established based on the joint index–modulation constellation. The results reveal that directly indexing all sampling points is generally unreliable, highlighting the necessity of anchor optimization. The performance of MA-IM is shown to depend on key system parameters, including channel richness, spatial correlation, the number of index states, and the modulation order. In particular, increasing the number of index states and increasing the QAM order affect MA-IM in fundamentally different ways, even under the same transmission rate. Among the proposed schemes, the joint constellation-aware anchor design achieves the best error performance, demonstrating that optimizing channel-domain separation alone is insufficient and that effective MA-IM design must account for the geometry of the joint signal constellation. Simulation results further show that, with properly designed anchors, MA-IM can approach or even outperform same-spectral-efficiency QAM baselines.

This paper proposes a novel modulation and coding scheme (MCS) selection framework that integrates mutual information (MI) prediction based on vector similarity search (VSS) for massive multi-user multiple-input multiple-output orthogonal frequency-division multiplexing (MU-MIMO-OFDM) systems with advanced uplink multi-user detection (MUD). The framework performs MCS selection at the transport block (TB)-level MI and establishes the mapping from post-MUD MI to post-decoding block error rate (BLER) using a prediction function generated from extrinsic information transfer (EXIT) curves. A key innovation is the VSS-based MI prediction scheme, which addresses the challenge of analytically predicting MI in iterative detectors such as expectation propagation (EP). In this scheme, an offline vector database (VDB) stores feature vectors derived from channel state information (CSI) and average received signal-to-noise ratio (SNR), together with corresponding MI values achieved with advanced MUD. During online operation, an approximate nearest neighbor (ANN) search on graphics processing units (GPUs) enables ultra-fast and accurate MI prediction, effectively capturing iterative detection gains. Simulation results under fifth-generation new radio (5G NR)-compliant settings demonstrate that the proposed framework significantly improves both system and user throughput, ensuring that the detection gains of advanced MUD are faithfully translated into tangible system-level performance improvements.