Search Results (190)

With the rapid development of social media, the speed and breadth of information dissemination have increased substantially, leading to more complex patterns in the emergence and evolution of online public opinion. Compared to unsigned networks, signed networks more accurately capture supportive and adversarial relationships among users. Although the traditional Polarity-Related Independent Cascade model (IC-P) can describe opinion propagation in signed networks, its capability remains limited when applied to complex social environments. To address this issue, this paper improves the IC-P model by incorporating a Prisoner’s Dilemma game to establish a user propagation-choice mechanism. Furthermore, activation probability and activation thresholds are redesigned from the perspectives of authority effect, homophily, and temporal decay, resulting in an Independent Cascade model incorporating Communication Choice and Polarity (ICC-P). Using three real-world negative public opinion datasets collected from the Sina Weibo platform spanning from March to April 2024, Monte Carlo simulations were conducted and compared with the main baseline models. Experimental results indicate that, relative to the best existing baselines, ICC-P reduces the mean absolute error of the prediction of the propagation scale by approximately 43% and reduces the mean absolute error of the prediction of the sentiment distribution of the nodes by approximately 57%, demonstrating significant improvements in both propagation fitting accuracy and sentiment prediction performance. Full article

Although existing research has found that parent–child conflict significantly predicts children’s psychological adjustment, it remains unclear whether father–child and mother–child conflict exert distinct effects on psychological adjustment, the mediating processes through which they operate, and whether these processes vary across primary and secondary school stages. Using a three-wave longitudinal design, this study examined 1210 primary school students (M_age = 10.17, SD_age = 0.85) and 973 secondary school students (M_age = 12.62, SD_age = 1.36). A multiple mediation model integrating parallel and serial paths was constructed to investigate how father–child and mother–child conflict frequency respectively predicted four indicators of psychological adjustment (internalizing problems, externalizing problems, life satisfaction, and prosocial behavior) and to test the mediating roles of parental psychological control and basic psychological needs. Results showed the following: (1) parental psychological control and basic psychological needs served as significant independent mediators of the relationship between conflict frequency and psychological adjustment. In primary school, maternal psychological control emerged as the core mediator; in secondary school, the mediating role of paternal psychological control was significantly strengthened, and the basic psychological need mediated all associations between mother–child conflict and every adjustment indicator. (2) The serial mediating pathway “parental psychological control → basic psychological needs” was robust across both school stages. As a distal family stressor, parent–child conflict is indirectly transformed into maladjustment through a sequential process that first elevates psychological control and then thwarts basic psychological need. These findings illuminate a cascading mechanism underlying the impact of parent–child conflict on multifaceted adjustment and offer stage-specific guidance for targeted family interventions in primary and secondary school settings. Full article

Background/Objectives: Accurate post-mortem interval (PMI) estimation remains challenging in forensic pathology. Although potassium (K⁺) is the most well-validated single biomarker in vitreous humor (VH), multivariate approaches may enhance precision by capturing the complex cascade of post-mortem biochemical changes. This study aimed to develop and validate a multivariate PMI estimation model incorporating three biochemical markers—potassium, ammonium (NH₄⁺), and albumin (ALB)—in vitreous humor using automated clinical chemistry platforms for practical forensic application. Methods: Vitreous humor samples from 38 autopsy cases with documented PMIs (39.5–285 h; mean, 105.5 h) were analyzed for K⁺ (Cobas C8000), NH₄⁺ (Cobas C8000), and ALB (Immage 800 nephelometry). Univariate and multivariate regression analyses were performed, with the residual standard error (RSE) as the primary measure of accuracy. Model validation was conducted by back-calculating PMI in four samples completely distinct from the training cohort. Results: All three analytes demonstrated strong individual correlations with PMI (R²: K⁺ = 0.88, ALB = 0.78, NH₄⁺ = 0.69; all p < 0.001). The multivariate regression model [PMI = 40.25[Alb] + 0.01573[NH₄⁺] + 5.339[K⁺] − 53.032] yielded an RMSE of ±15.5 h (MSE = 240.25 h²), outperforming potassium-only models (RMSE = ±22.6 h). Although NH₄⁺ showed limited statistical significance in the multivariate model (p = 0.128), its inclusion improved overall predictive accuracy. External validation in an independent cohort of four subjects (distinct from the 38 subjects in the training set) demonstrated a mean absolute error (MAE) of 20.4 h. Conclusions: The multivariate approach combining K⁺, NH₄⁺, and ALB in VH improves PMI estimation accuracy compared with single-marker methods. The use of automated clinical chemistry platforms enhances reproducibility and facilitates practical implementation in forensic laboratories. Full article

Early non-invasive correction of neonatal auricular deformities is highly dependent on timely and precise diagnosis. However, clinical practice is often compromised by the subjectivity of visual assessments and the lack of objective tracking metrics, which frequently leads to missed optimal treatment windows. To address these challenges, we developed an interpretable deep learning-based diagnostic system for the automated screening and fine-grained classification of these deformities. Methodologically, a large-scale, multi-source dataset (n = 4644) was curated to support model training. The system pairs an automated object detector (YOLOv11) for background-reduced region-of-interest isolation with a cascaded classification pipeline optimized via ConvNeXt-Tiny. Crucially, we introduced a supervised contrastive learning module to project high-dimensional morphological features into a continuous severity score, enabling quantitative longitudinal tracking of therapeutic efficacy. To evaluate generalization and robustness, the framework underwent rigorous evaluation across three independent real-world cohorts and one controlled synthetic stress test. The system achieved 88.2% accuracy (Area Under the Curve (AUC): 0.949) in binary screening and 87.4% accuracy (macro-AUC: 0.976) in multi-class subtyping on the internal baseline. To enhance interpretability and build clinical trust, Gradient-weighted Class Activation Mapping (Grad-CAM) was utilized to explore the spatial distribution of the model’s attention, which frequently aligned with key anatomical landmarks. Furthermore, the learned severity scores robustly quantified post-intervention improvements (p = 0.0004), effectively capturing subtle anatomical normalization. While validation for rare subtypes remains underpowered, and the severity score currently functions mainly as a learned morphological similarity index requiring future clinical calibration, this study ultimately provides an objective and standardized web-based tool to facilitate the early intervention and precision management of neonatal auricular anomalies. Full article

Background/Objectives: Spontaneous intracranial hypotension (SIH) is caused by spinal cerebrospinal fluid (CSF) leakage and is typically diagnosed by clinical presentation and characteristic MRI signs; however, objective tools for monitoring physiological changes and treatment response remain limited. Cine phase-contrast MRI (PC-MRI) enables noninvasive quantification of aqueductal CSF dynamics, yet reliable analysis is challenging since the cerebral aqueduct is extremely small and susceptible to low contrast, partial volume effects, and ROI-dependent measurement variability—particularly in SIH where CSF pulsatility is often reduced. Methods: We propose an end-to-end automated framework that integrates (1) a cascade localization–segmentation strategy, consisting of Tiny YOLOv4 detection followed by MultiResUNet segmentation on a YOLOv4-derived cropped ROI; (2) physiology-informed pulsatility-based segmentation (PUBS) to refine anatomical masks into functional flow ROIs; and (3) one-dimensional convolutional neural networks (1D-CNNs) to extract exploratory waveform morphology features from 32-phase cardiac-cycle velocity waveforms. The study includes 39 participants, yielding 59 cine PC-MRI examinations: 11 controls, 28 Pre-treatment SIH scans and 20 Post-treatment Recovery scans. Results: The cascade model significantly improves segmentation robustness compared with a full-image baseline, achieving higher Dice scores and markedly lower boundary errors across cohorts (e.g., Pre-treatment SIH HD95: 1.66 ± 0.74 px vs. 15.37 ± 44.98 px). PUBS refinement reduces quantification deviation from expert manual references in SIH (mean relative error: 7.4% to 5.6%) and improves diagnostic performance for multiple hemodynamic parameters (e.g., downward mean flow AUC: 0.747 to 0.792). For waveform morphology analysis, the end-to-end 1D-CNN classifier was evaluated using repeated-seed participant-level grouped LOOCV. The repeated-seed ensemble prediction showed modest out-of-sample discrimination between Normal controls and Pre-treatment SIH scans, with an AUC of 0.646, a bootstrap 95% confidence interval of 0.455–0.826, and a permutation-test p-value of 0.072. Separately, exploratory analysis of the final baseline-trained 1D-CNN latent space showed marked, apparent Normal-versus-SIH separability and an intermediate recovery distribution in PCA space, suggesting that aqueductal waveform morphology may encode SIH-related physiological information. Conclusions: These findings suggest that SIH-related information may be reflected not only in flow magnitude but also in aqueductal CSF waveform morphology. However, the modest and statistically non-significant out-of-sample performance of the end-to-end 1D-CNN classifier indicates that morphology-based AI features should currently be regarded as exploratory biomarker candidates rather than validated stand-alone diagnostic tools. Larger independent cohorts are required to confirm their reproducibility, physiological meaning, and clinical utility. Full article

Selective laser melting (SLM) equipment operates as a complex cyber–physical system, wherein strong implicit coupling among internal subsystems presents significant challenges for condition monitoring and fault diagnosis. Existing deep learning methods often suffer from feature submersion when processing multi-source heterogeneous data and lack the capability for system-level topological causal inference. To address these issues, we propose a multisystem coupling modeling and anomaly traceability method based on a decoupled graph attention network (ST-DBGAE). Independent local spatiotemporal feature alignment modules are constructed to map heterogeneous sensory data into a unified latent space. This eliminates dimensional discrepancies while strictly maintaining the feature independence of underlying hardware subsystems, such as optical and gas circuits. A dynamic graph attention mechanism with sparse priors is subsequently introduced to adaptively capture time-varying coupling weights triggered by implicit interactions (e.g., thermal fluids), bypassing the need for predefined rigid physical connections. Furthermore, a dual-branch two-stage decoupled optimization architecture is designed. By blocking the cross-interference of global backpropagation, this architecture outputs a continuous equipment health index (HI) based on reconstruction errors and employs a topological difference matrix inference mechanism to reversely anchor the root-cause nodes responsible for cross-system cascading degradation. Experimental results based on over 310,000 real operational monitoring records from industrial SLM equipment demonstrate that the proposed model achieves a comprehensive diagnostic Macro-F1 score of 96.5% across eight operating states. The single-class detection rates (ACCs) of specific underlying anomalies are significantly improved. This method not only enables high-precision equipment health warnings but also provides a physically interpretable microscopic fault propagation mapping for predictive maintenance. Full article

Wearable ECG monitors enable continuous cardiac surveillance, but most still rely on cloud-based analysis with limited on-device support for multi-class arrhythmia detection. Spiking neural networks (SNNs) are promising for low-power edge inference, yet it remains unclear how class-imbalance loss design interacts with RR-interval features in directly trained quantized SNNs, and FPGA validation in this setting is largely unexplored. We propose a quantized convolutional spiking neural network (QCSNN) for real-time arrhythmia detection on resource-constrained hardware. The model uses a dual-head architecture that jointly trains binary and four-class classifiers, subsequently reorganized into a cascaded pipeline that routes only abnormal beats to the second stage. At inference, beats classified as Normal exit at Stage 1; only beats classified as Abnormal are routed to the four-class head, so the bulk of the inference cost is absorbed by Stage 1. We evaluate two loss functions, Cross-Entropy and Focal Loss, under four RR-feature routing strategies. Without RR features, Focal Loss improves macro F1 by 2.3–2.5% over Cross-Entropy (mean Δ = +0.013 in Stage-2 macro F1; Wilcoxon two-sided p = 0.031). With RR features, this advantage largely disappears (Wilcoxon two-sided p ≥ 0.219 at all RR routings); meanwhile, RR features at the strongest routing improve Stage-2 macro F1 by +0.028 to +0.034 depending on loss function—a gain that exceeds the entire Focal-Loss-over-Cross-Entropy advantage, suggesting that RR features provide discriminative information that compensates for class imbalance at the input level. Based on clinically prioritized sensitivity, the CE:RR→Both configuration was deployed on a PYNQ-Z2 FPGA, achieving 99.02% cascaded accuracy, 11.54 ms per-beat latency, and 0.33 W accelerator power—a 31.66× power reduction and 4.01× energy reduction versus GPU inference, within 1% macro F1. These results demonstrate quantized SNNs as a practical solution for real-time edge arrhythmia monitoring that operates independently of cloud connectivity—removing the network-dependent latency, connectivity-dropout failure modes, and continuous-transmission energy burden that constrain current wearable monitors and, to our knowledge, represent one of the first systematic studies of loss-function/RR-feature interactions in directly trained SNN arrhythmia classification and one of the first FPGA deployments of a fully quantized, directly trained SNN for multi-class ECG arrhythmia detection. All code generated and used in this study has been made publicly available. Full article

Background/Objectives: Trichomonas vaginalis is one of the most prevalent non-viral sexually transmitted infections worldwide. Although Gram staining is routinely performed in clinical microbiology laboratories for the evaluation of vaginal samples, it is not considered a diagnostic method for T. vaginalis, which represents a missed diagnostic opportunity in routine practice. This study aimed to evaluate an artificial intelligence (AI)-assisted diagnostic approach for the identification of T. vaginalis directly from routine Gram-stained vaginal smears. Methods: A retrospective dataset of Gram-stained vaginal smear images was analysed using a cascaded AI-based framework combining image processing and classification. The image selection and quality control were performed under the supervision of a specialised clinical microbiologist. All cases were independently confirmed by polymerase chain reaction (PCR), which served as the reference diagnostic standard. Model performance was assessed using standard diagnostic metrics, including accuracy, sensitivity, specificity, area under the receiver operating characteristic curve (AUC), Cohen’s kappa, and Matthews correlation coefficient (MCC). Held-out independent testing was used to assess generalisability beyond the internal validation subset. Results: The proposed AI-assisted approach demonstrated high diagnostic performance for the identification of T. vaginalis, achieving an AUC of 0.973, Cohen’s kappa of 0.87, and an MCC of 0.87. The system showed high diagnostic concordance with PCR results across both internal and external validation datasets, supporting the feasibility and reproducibility of the approach under routine laboratory conditions. Conclusions: This study shows that artificial intelligence may enhance the diagnostic utility of routinely performed Gram-stained vaginal smears by enabling reliable identification of T. vaginalis. The proposed approach could be integrated into standard microbiology workflows as an objective decision-support or triage adjunct, facilitating early identification and supporting clinical decision-making without altering existing laboratory procedures. Full article

Establishing long-term human settlements in deep space presents significant challenges. Environmental conditions, such as extreme temperature fluctuations, micrometeorite impacts, seismic activity, and exposure to solar and cosmic radiation, pose obstacles to the design and operation of habitat systems. Prolonged mission duration and vast distances from Earth introduce further complications in the form of delayed communication and limited resources, making Earth independence through appropriate autonomous management systems especially desirable. Enabling the modeling and simulation of the consequences of disruptions and faults, and their propagation through the various habitat subsystems, is critically needed for the development of resilience-based design frameworks and methods for autonomous operation. While existing simulation tools can assist in modeling isolated aspects of damage, the integration of damage propagation and the capacity to enable detection and repair are rarely considered in a computational model. This paper introduces and demonstrates an architecture designed specifically to enable the modeling and integration of faults and damage, as well as their cascading effects. By combining physics-based and phenomenological models, our approach balances computational efficiency with model fidelity. After describing the modeling approach and corresponding architecture, we demonstrate its application within HabSim, a system-level space habitat model developed by the NASA-funded Resilient Extraterrestrial Habitat Institute (RETHi), as a simulation-based design aid suited to early-phase trade studies. Fire hazard propagation within a lunar habitat is used as an illustrative example of how the architecture supports modeling of disruption consequences, propagation, detection, and repair, and of how HabSim can be leveraged for stochastic simulations to support resilience assessment. Resilience-focused studies that apply this architecture can quantify and compare design alternatives. Full article

Background/Objectives: Empowered mothers are more likely to adopt recommended childcare practices, thereby contributing to reduced childhood undernutrition. However, the magnitude of the association between maternal empowerment and childhood undernutrition in Bangladesh has not been comprehensively assessed. This study aims to address this research gap. Methods: The Bangladesh Demographic and Health Survey (BDHS) 2017-18 and BDHS 2022 served as data sources. Maternal empowerment was assessed across three domains, e.g., attitude to violence, social independence, and decision making, using the Survey-based Women’s Empowerment (SWPER) index. The undernutrition status of children was assessed through z-score based indicators, including stunting (height-for-age z-score < −2 SD), wasting (weight-for-height z-score < −2 SD), and underweight (weight-for-age < −2 SD). Children with at least two and any of these undernutrition conditions were categorized as multiple undernutrition and any undernutrition, respectively. Multivariable logistic regression models were utilized to observe the survey-specific and pooled association between maternal empowerment and childhood undernutrition. Results: The analysis includes 11,647 mother–child pairs. The association between maternal empowerment and childhood undernutrition was consistent across individual surveys and the pooled sample, although the significance level varied by empowerment domains and undernutrition categories. Maternal social independence was found to be a significant protective factor against both multiple and any childhood undernutrition status in individual surveys and the pooled sample. For example, in the pooled sample, high maternal empowerment in the social independence domain was significantly associated with 18% (AOR: 0.82; 95% CI: 0.69, 0.98; p = 0.026) lower odds of multiple undernutrition statuses and 18% (AOR: 0.82; 95% CI: 0.71, 0.95; p = 0.009) lower odds of any undernutrition statuses than those of low maternal empowerment. Conclusions: Improving the status of maternal social independence can potentially result in reduced childhood undernutrition. The scope remains to cascade the benefits of the other two maternal empowerment domains, e.g., attitude to violence and decision making, to child nutrition in Bangladesh. Full article

A coupled multi-stage learning framework is presented for joint bit error rate (BER) prediction and beam angle optimization in massive multiple-input multiple-output (MIMO) systems under a controlled simulation protocol. Unlike purely sequential benchmarking pipelines, the proposed method jointly coordinates BER prediction and beam-angle selection through a shared latent representation, an uncertainty-guided refinement mechanism, a cross-stage consistency loss and alternating optimization. Ten diverse approaches are systematically evaluated across two task-specific stages: Stage 1 examines six classical and adapted methods for BER prediction, including polynomial regression and deep unfolding networks; Stage 2 investigates four machine-learning and generative adversarial network (GAN)-based approaches for angle optimization, including conditional GANs and the proposed Direct-Angle neural network. Stage 3 couples the best-performing methods into a unified hybrid architecture through a shared encoder, explicit consistency regularization and alternating cross-stage updates, thereby producing an integrated beamforming decision strategy rather than an independent cascade. It is shown through the evaluation that the coupled hybrid framework achieves 96.0% overall angle-selection accuracy, a mean BER of $8.0\times {10}^{-5}$ and 100% BER tolerance compliance within $\pm 3$ dB. In this framework, a differentiable BER surrogate initialized from a second-degree polynomial-regression teacher is coupled with the proposed Direct-Angle-NN for angle optimization. Relative to the strongest reimplemented literature baseline under the same controlled simulation assumptions, a 33.3% reduction in mean BER is achieved. Ablation experiments show that the coupling mechanism provides a modest but consistent improvement over the decoupled sequential baseline, increasing angle-selection accuracy from 93.5% to 96.0% and reducing mean BER from $1.05\times {10}^{-4}$ to $8.0\times {10}^{-5}$; the shared encoder accounts for the largest part of this gain while the consistency loss adds 0.6 percentage points. These results indicate that the shared encoder, consistency regularization and uncertainty-guided refinement improve the final beamforming decision, although the gain should be interpreted as incremental rather than as a large architectural breakthrough. A spectral efficiency of 38.0 bps/Hz and an energy efficiency of 0.466 Gbps/W are achieved with a power consumption of only 32.6 W. The theoretical discussion is presented as an analytical characterization of BER sensitivity, complemented by a computational-complexity assessment and empirical convergence diagnostics for the alternating optimization, rather than as a formal optimality proof. The effectiveness of the framework across multiple performance metrics is supported by Monte Carlo simulations, while the limitations of the current setup, including perfect CSI, uncoded QPSK, ideal hardware assumptions and a fixed beam codebook, are explicitly discussed. The complete simulation framework, including code and trained models, can be made available by the corresponding author upon reasonable request to facilitate reproducible research in massive MIMO optimization. Full article

This paper proposes a hybrid time–frequency domain identification method for second-order plus dead time models with an additional process zero (SOPDT+Z). A dual-relay experiment combined with step response data provides six independent equations for five model parameters, whose mathematical well-posedness is established through Jacobian rank analysis. A cascaded initialization strategy (Sundaresan → SIMC → Jin → proposed) guarantees monotonically improving accuracy. An Internal Model Control (IMC) framework yields equivalent PID parameters with a single tuning parameter λ, supported by a formal robust stability theorem. Simulation studies on five benchmark systems demonstrate 60–100% reduction in open-loop IAE compared to existing SOPDT methods, 36% faster settling, and 100% closed-loop stability under ±20% Monte Carlo perturbation (N = 200). Noise robustness analysis under SNR = 20–40 dB and additional performance metrics (ITAE, ISE) further validate the method. Full article

Coronavirus disease 2019 (COVID-19) has been associated with an increased long-term cardiovascular risk, potentially mediated by magnitude of the acute inflammatory response inflammation. Interleukin-6 (IL-6) and serum amyloid A (SAA) are key components of the inflammatory cascade and may serve as biomarkers of post-COVID cardiovascular vulnerability. This longitudinal observational study investigated the association between post- COVID-19 infection IL-6 and SAA levels and major cardiovascular events over a six-year follow-up period. A total of 97 individuals with documented prior SARS-CoV-2 infection were included. Circulating IL-6 and SAA concentrations were measured in the acute phase. The composite endpoint included incident arrhythmia, myocardial infarction, and all-cause mortality. Biomarker distributions were right-skewed and were therefore analyzed using non-parametric methods and penalized logistic regression models. During follow-up, 14.4% of participants experienced the composite endpoint. Individuals with adverse outcomes had significantly higher IL-6 and SAA levels compared with event-free participants. IL-6 demonstrated the strongest association with mortality, whereas SAA showed particularly robust associations with the composite endpoint, and with myocardial infarction. Both biomarkers independently predicted long-term adverse events. Circulating IL-6 and SAA concentrations measured during the acute phase of SARS-CoV-2 infection were analyzed in relation to long-term cardiovascular outcomes. These findings support the hypothesis that the magnitude of the acute inflammatory response during SARS-CoV-2 infection may be associated with long-term cardiovascular outcomes and suggest that combined assessment of IL-6 and SAA may have potential utility for hypothesis-generating prognostic signal requiring validation, pending validation in larger studies. Full article

Cascade hydropower scheduling is a high-dimensional, tightly constrained multi-objective optimization problem in which classical genetic and evolutionary algorithms struggle to find feasible solutions. Under random initialization, algorithms such as the Non-dominated Sorting Genetic Algorithm II (NSGA-II), the Non-dominated Sorting Genetic Algorithm III (NSGA-III), and the Constrained Two-Archive Evolutionary Algorithm (C-TAEA) rarely produce any feasible solution when the feasible region occupies a vanishingly small fraction of the search space. This paper presents a three-phase framework that combines physics-guided deep learning with evolutionary computation to accelerate both NSGA-II and NSGA-III. The method trains a Physics-Embedded Temporal Convolutional Network (PeTCN) as a differentiable surrogate model that explicitly incorporates physical constraints, applies gradient-based inverse optimization to obtain a feasible or near-feasible solution of high quality, and warm-starts NSGA-II or NSGA-III with that solution for efficient Pareto front exploration. Experiments on a real-world six-station cascade system show that, under a 1500 s fixed-time budget across 20 independent runs, Boosted NSGA-II and Boosted NSGA-III both find feasible solutions in all runs. Boosted NSGA-II and Boosted NSGA-III both reach the first feasible solution within roughly 50–60 generations of Phase 3 search on average, whereas standard NSGA-II produces no feasible run within the same budget and standard NSGA-III requires thousands of generations among its successful runs. The mean final hypervolume reaches $43.84\times {10}^6$ for Boosted NSGA-II and $46.52\times {10}^6$ for Boosted NSGA-III, and both boosted algorithms reach a target hypervolume of $35.00\times {10}^6$ in all 10 target-hypervolume runs. These results demonstrate that coupling physics-embedded surrogates with gradient-based initialization is an effective strategy for constrained multi-objective problems in which feasible solutions are extremely sparse. Full article

The Influence Maximization Problem addresses the challenge of selecting a set of spreaders with the aim of maximizing the spread influence on the network. In this work we propose a novel influence spreading model formulated for Social Signed Multi-Networks. We combine the assumption that only infected nodes express their opinion with the bounded confidence assumption. We solve the Influence Maximization Problem under this model by proposing a novel proxy-based Greedy method. We validate the effectiveness of our method through simulations on multi-networks constructed from two real online social signed networks, Epinions and Slashdot, and compare it against several baseline methods. The simulation results demonstrate that our proposed method consistently outperforms all other baselines in each case tested. Full article