Search Results (61)

Empathetic response generation stands as a pivotal endeavor in the development of human-like dialogue systems. An effective approach in previous research is integrating external knowledge to generate empathetic responses. However, existing approaches only focus on identifying a user’s current emotional state, and they overlook the user’s emotional transition during context, and fail to propel the sustainability of the dialogue. To tackle the aforementioned issues, we propose an empathetic response generation model based on an emotional transition prompt and dual-semantic contrastive learning (EPDC). Specifically, we first compute the transition in users’ sentiment polarity during the conversation and incorporate it into the conversation embedding as sentiment prompts. Then, we generate two distinct fine-grained contextual representations and treat them as positive examples for contrastive learning, respectively, aiming at extracting high-order semantic information to guide the subsequent turn of dialogue. Finally, we also leverage commonsense knowledge to enhance the contextual representations, and the empathetic responses are generated by decoding the combination of semantic and emotional states. Notably, our work represents the pioneering application of emotional prompts and contrastive learning to augment the sustainability of empathetic dialogue. Extensive experiments conducted on the benchmark dataset EMPATHETICDIALOGUES demonstrate that EPDC outperforms the baselines in both automatic evaluations and human evaluations. Full article

Precise extraction of rural settlements in alpine regions is critical for geographic data production, rural development, and spatial optimization. However, existing deep learning models are hindered by insufficient datasets and suboptimal algorithm structures, resulting in blurred boundaries and inadequate extraction accuracy. Therefore, this study uses high-resolution unmanned aerial vehicle (UAV) remote sensing images to construct a specialized dataset for the extraction of rural settlements in alpine mountainous areas, while introducing an innovative shadow mitigation technique that integrates multiple spectral characteristics. This methodology effectively addresses the challenges posed by intense shadows in settlements and environmental occlusions common in mountainous terrain analysis. Based on the comparative experiments with existing deep learning models, the Swin Transformer was selected as the baseline model. Building upon this, the Feature Fusion Swin Transformer (FF-SwinT) model was constructed by optimizing the data processing, loss function, and multi-view feature fusion. Finally, we rigorously evaluated it through ablation studies, generalization tests and large-scale image application experiments. The results show that the FF-SwinT has improved in many indicators compared with the traditional Swin Transformer, and the recognition results have clear edges and strong integrity. These results suggest that the FF-SwinT establishes a novel framework for rural settlement extraction in alpine mountain regions, which is of great significance for regional spatial optimization and development policy formulation. Full article

Large-language-model (LLM) APIs demonstrate impressive reasoning capabilities, but their size, cost, and closed weights limit the deployment of knowledge-aware AI within biomedical research groups. At the other extreme, standard attention-based neural language models (SANLMs)—including encoder–decoder architectures such as Transformers, Gated Recurrent Units (GRUs), and Long Short-Term Memory (LSTM) networks—are computationally inexpensive. However, their capacity for semantic reasoning in noisy, open-vocabulary knowledge bases (KBs) remains unquantified. Therefore, we investigate whether compact SANLMs can (i) reason over hybrid OpenIE-derived KBs that integrate commonsense, general-purpose, and non-communicable-disease (NCD) literature; (ii) operate effectively on commodity GPUs; and (iii) exhibit semantic coherence as assessed through manual linguistic inspection. To this end, we constructed four training KBs by integrating ConceptNet (600k triples), a 39k-triple general-purpose OpenIE set, and an 18.6k-triple OpenNCDKB extracted from 1200 PubMed abstracts. Encoder–decoder GRU, LSTM, and Transformer models (1–2 blocks) were trained to predict the object phrase given the subject + predicate. Beyond token-level cross-entropy, we introduced the Meaning-based Selectional-Preference Test (MSPT): for each withheld triple, we masked the object, generated a candidate, and measured its surplus cosine similarity over a random baseline using word embeddings, with significance assessed via a one-sided t-test. Hyperparameter sensitivity (311 GRU/168 LSTM runs) was analyzed, and qualitative frame–role diagnostics completed the evaluation. Our results showed that all SANLMs learned effectively from the point of view of the cross entropy loss. In addition, our MSPT provided meaningful semantic insights: for the GRUs (256-dim, 2048-unit, 1-layer): mean similarity $\left({\mu }_{sts}\right)$ of 0.641 to the ground truth vs. 0.542 to the random baseline (gap 12.1%; $p<{10}^{-180}$). For the 1-block Transformer: ${\mu }_{sts}=0.551$ vs. $0.511$ (gap 4%; $p<{10}^{-25}$). While Transformers minimized loss and accuracy variance, GRUs captured finer selectional preferences. Both architectures trained within <24 GB GPU VRAM and produced linguistically acceptable, albeit over-generalized, biomedical assertions. Due to their observed performance, LSTM results were designated as baseline models for comparison. Therefore, properly tuned SANLMs can achieve statistically robust semantic reasoning over noisy, domain-specific KBs without reliance on massive LLMs. Their interpretability, minimal hardware footprint, and open weights promote equitable AI research, opening new avenues for automated NCD knowledge synthesis, surveillance, and decision support. Full article

In this work, we employ two modeling approaches—a mean-field model and a network model—for the purpose of modeling respiratory infection outbreaks in Russia. The presented approaches and their software implementation combine heterogeneity and structural simplicity and, in this sense, they close the gap between the compartmental SEIR models and complex detailed solutions based on agent-based approaches—the two most common modeling techniques for influenza and COVID-19 dynamics. The mathematical description of the approaches is presented, with SEIR compartmental model serving as a baseline for comparison. The experiments demonstrate the similarity of the modeling output of the presented approaches, which allows their interchangeable usage in replicating real outbreak dynamics in Russian cities. The ability of the discussed approaches to mimic data from Russian epidemic surveillance is shown by fitting a mean-field model to data from an influenza outbreak in Saint Petersburg in 2014–2015. The comparison of model complexity and their performance is made using synthetic scenarios. Following the results of numerical experiments, the comparative advantages and drawbacks of the approaches in the application to respiratory infection outbreaks are discussed. The presented modeling techniques, in addition to classical SEIR models and agent-based models as a part of epidemic surveillance, allow one to select the best modeling option for any particular task in outbreak surveillance and control, based on the computational resources at hand, data availability, and data quality. Full article

Due to the special properties of synthetic aperture radar (SAR) images, they are widely used in maritime applications, such as detecting ships at sea. To perform ship detection in SAR images, existing algorithms commonly utilize convolutional neural network (CNN). However, the challenges in acquiring SAR images and the imaging noise hinder CNN in performing SAR ship-detection tasks. In this paper, we revisit the relationship between SAR expert features and network abstract features, and propose an expert-feature-based convolutional neural network (EFCNet). Specifically, we exploit the inherent physical properties of SAR images by manually extracting a range of expert features, including electromagnetic scattering, geometric structure, and grayscale statistics. These expert features are then adaptively integrated with abstract CNN features through a newly designed multi-source features association module, which improves the common CNN’s capability to recognize ship targets. Experiment results on the SSDD demonstrate that EFCNet outperforms general CNN approaches. Furthermore, EFCNet achieves comparable detection performance to baseline methods while utilizing only 70% of the data capacity, highlighting its efficiency. This work aims to reignite interest in leveraging expert features in remote sensing tasks and offers promising avenues for improved SAR image interpretation. Full article

Despite the introduction of preventive safety measures, such as near-miss reporting, to mitigate accidents and minimize fatalities, construction workers are constantly exposed to stressful situations that negatively affect their safety behavior and reporting efficiency. Occupational stress is induced by various factors, with mental stress and auditory stress being common workplace stressors that impact workers on the job site. While previous studies have demonstrated the effect of stressor conditions on workers’ hazard recognition and safety performance, research gaps persist regarding the direct impact of workplace stressors on workers’ stress levels and near-miss recognition performance. This study investigates workers’ near-miss recognition ability through an eye-tracking experiment conducted in a controlled environment under mental and auditory stress conditions. The findings from this study reveal that workplace stressors triggered by mental and auditory stress can adversely affect worker stress levels, safety behavior, and cognitive processing toward near-miss recognition. Visual attention towards near-miss scenarios was reduced by 26% for mental stress conditions and by 46% for auditory stress conditions compared to baseline. The results may potentially open avenues for developing wearable stress prediction and safety intervention models using bio-sensing technology and personalized safety training programs tailored to individuals with low identification abilities. Full article

In the field of question answering (QA), the methods of large language models (LLMs) cannot learn vertical domain knowledge during the pre-training stage, leading to low accuracy in domain QA. Conversely, knowledge base question answering (KBQA) can combine the knowledge base (KB) that contains domain knowledge with small language models to achieve high accuracy with a low cost. In KBQA, the inference subgraph is composed of entity nodes and their relationships pertinent to the question, with the final answers being derived from the subgraph. However, there are still two critical problems in this field: (i) fixed or decreased scopes of the inference subgraphs over the reasoning process may lead to limited knowledge, restricted in KBQA, and (ii) a lack of alignment between the inference subgraph and the question leads to low accuracy. In this work, we propose a dynamic graph reasoning model with an auxiliary task, the DGRMWAT, which addresses the above challenges through two key innovations, as follows: (i) dynamic graph reasoning, whereby we update the scope of the inference subgraph during each reasoning step to obtain more relevant knowledge and reduce irrelevant knowledge, and (ii) an auxiliary task to enhance the correlation between the inference subgraph and the question by computing the similarities between the inference subgraph and the QA context node. The experiments on two QA benchmark datasets, CommonsenseQA and OpenbookQA, indicate that the DGRMWAT allowed improvements compared to the baseline models and LLMs. Full article

Background: The wearable cardioverter defibrillator (WCD) has emerged as a valuable tool used for temporary protection from sudden cardiac death. However, since the WCD uses surface electrodes to detect arrhythmias, it is susceptible to inappropriate detection. Although shock conversion rates for the WCD are reported to be high for detected events, its efficacy in clinical practice tends to be degraded by patient noncompliance. Reasons for this include wearer discomfort and frequent false alarms, which may interrupt sleep and generate anxiety. Up to now, data on the incidence of false alarms emitted by the WCD and their predictors are rare. Objectives: The aim of our study was to assess the relationship between both artifact sensing and episode misclassification burden and wearing compliance in patients with a WCD (ZOLL LifeVest™ 4000 system, ZOLL CMS GmbH, Cologne, Germany). Methods and Results: We conducted a single-center retrospective observational study, analyzing patients with a WCD prescribed at our institution. A total of 134 patients (mean age 51.7 ± 13.8 years, 79.1% male) were included. Arrhythmia recordings were analyzed and categorized as non-sustained ventricular tachycardia, sustained ventricular tachycardia or fibrillation, artifact sensing or misclassified episodes. Indication for WCD prescription was both primary and secondary prophylaxis. A total of 3019 false WCD alarms were documented in 78 patients (average number of false alarms 38.7 ± 169.5 episodes per patient) over a mean WCD wearing time of 71.5 ± 70.9 days (daily WCD wearing time 20.2 ± 5.0 h). In a total of 78 patients (58.2% of the study population), either artifact sensing (76.9%), misclassified episodes (6.4%), or both (16.7%) occurred. Misclassified episodes included sinus tachycardias, atrial flutter, atrial fibrillation, premature ventricular contractions (PVCs), and intermittent bundle branch block. A multiple linear regression identified loop diuretics (regression coefficient [B] −0.11; 95% CI −0.21–(−0.0001); p = 0.049), angiotensin receptor–neprilysin inhibitors (ARNIs) (B −0.11; 95% CI 0.22–(−0.01); p = 0.033), and a higher R-amplitude of the WCD baseline electrocardiogram (ECG) (B −0.17; 95% CI −0.27–(−0.07); p = 0.001) as independent predictors for a lower number of artifact episodes per day. In addition, atrial fibrillation (B 0.05; 95% CI 0.01–0.08; p = 0.010), and calcium antagonists (B 0.07; 95% CI 0.02–0.12; p = 0.012) were independent predictors for increased numbers of misclassified episodes per day, while beta-blockers seemed to reduce them (B −0.06; 95% CI −0.10–(−0.01); p = 0.013). Patients terminated 61.0% of all false alarms manually by pressing the response button on average 1.9 times per false alarm with overall 3.6 manual terminations per affected patient per month. Conclusions: In conclusion, false alarms from the ZOLL LifeVest™ system were frequent, with artifact sensing being the most common cause. Hence, the occurrence of false alarms represents a significant side effect of WCD therapy, and efforts should be made to minimize false alarms. Full article

Background/Objectives: Olfactory dysfunction (OD) is associated with a variety of neurologic deficits and impacts socialization decisions, mood, and overall quality of life. As a common symptom comprising the long COVID condition, persistent COVID-19-associated olfactory dysfunction (C19OD) may further impact the presentations of neuropsychiatric sequelae. Our study aims to characterize the longitudinal burden of depression, anxiety, and neuropsychiatric symptoms in a population with C19OD. Methods: Individuals with perceived C19OD completed a psychophysical screening evaluation of their sense of smell using the comprehensive Sniffin’ Sticks olfactory assessment. Only those with validated psychophysical OD were included in this prospective longitudinal study for baseline and one-year follow-up. Participants also completed PHQ-9, Beck Anxiety Inventory (BAI), and neuropsychiatric symptom questionnaires at each time point. Anxiety, depression, and neuropsychiatric symptom prevalence was calculated and compared between time points with Pearson’s chi-squared, Fisher’s exact, and Wilcoxon rank sum tests. Results: Each neuropsychiatric symptom evaluated in this study was reported by 13–49% of longitudinal cohort participants at both baseline and follow-up, except for seizure (0% at baseline and follow-up) and word-finding difficulty (61–68% at baseline and follow-up). Word-finding and focus difficulties were the most commonly reported symptoms. In total, 41% of participants reported some level of depression at baseline and 38% of participants reported depression at one-year follow-up, while 29% and 27% of participants reported some level of anxiety at respective time points. Conclusions: Individuals with C19OD are at risk for developing persistent neuropsychiatric conditions. These neurologic and psychiatric sequelae are persistent with repeated longitudinal assessment, even at nearly 2.5 years following initial COVID-19 diagnosis. Full article

Commonsense question answering (CSQA) is a challenging task in the field of knowledge graph question answering. It combines the context of the question with the relevant knowledge in the knowledge graph to reason and give an answer to the question. Existing CSQA models combine pretrained language models and graph neural networks to process question context and knowledge graph information, respectively, and obtain each other’s information during the reasoning process to improve the accuracy of reasoning. However, the existing models do not fully utilize the textual representation and graph representation after reasoning to reason about the answer, and they do not give enough semantic representation to the edges during the reasoning process of the knowledge graph. Therefore, we propose a novel parallel fusion framework for text and knowledge graphs, using the fused global graph information to enhance the semantic information of reasoning answers. In addition, we enhance the relationship embedding by enriching the initial semantics and adjusting the initial weight distribution, thereby improving the reasoning ability of the graph neural network. We conducted experiments on two public datasets, CommonsenseQA and OpenBookQA, and found that our model is competitive when compared with other baseline models. Additionally, we validated the generalizability of our model on the MedQA-USMLE dataset. Full article

Artificial intelligence (AI), including machine learning (ML) and deep learning (DL), learns by training and is restricted by the amount and quality of training data. Training involves a tradeoff between prediction bias and variance controlled by model complexity. Increased model complexity decreases prediction bias, increases variance, and increases overfitting possibilities. Overfitting is a significantly smaller training prediction error relative to the trained model prediction error for an independent validation set. Uncertain data generate risks for AI–ML because they increase overfitting and limit generalization ability. Specious confidence in predictions from overfit models with limited generalization ability, leading to misguided water resource management, is the uncertainty-related negative consequence. Improved data is the way to improve AI–ML models. With uncertain water resource data sets, like stream discharge, there is no quick way to generate improved data. Data assimilation (DA) provides mitigation for uncertainty risks, describes data- and model-related uncertainty, and propagates uncertainty to results using observation error models. A DA-derived mitigation example is provided using a common-sense baseline, derived from an observation error model, for the confirmation of generalization ability and a threshold identifying overfitting. AI–ML models can also be incorporated into DA to provide additional observations for assimilation or as a forward model for prediction and inverse-style calibration or training. The mitigation of uncertain data risks using DA involves a modified bias–variance tradeoff that focuses on increasing solution variability at the expense of increased model bias. Increased variability portrays data and model uncertainty. Uncertainty propagation produces an ensemble of models and a range of predictions. Full article

Multicenter international clinical trials demonstrated the clinical safety and efficacy by using stem cell educator therapy to treat type 1 diabetes (T1D) and other autoimmune diseases. Previous studies characterized the peripheral blood insulin-producing cells (PB-IPC) from healthy donors with high potential to give rise to insulin-producing cells. PB-IPC displayed the molecular marker glucose transporter 2 (GLUT2), contributing to the glucose transport and sensing. To improve the clinical efficacy of stem cell educator therapy in the restoration of islet β-cell function, we explored the GLUT2 expression on PB-IPC in recent onset and longstanding T1D patients. In the Food and Drug Administration (FDA)-approved phase 2 clinical studies, patients received one treatment with the stem cell educator therapy. Peripheral blood mononuclear cells (PBMC) were isolated for flow cytometry analysis of PB-IPC and other immune markers before and after the treatment with stem cell educator therapy. Flow cytometry revealed that both recent onset and longstanding T1D patients displayed very low levels of GLUT2 on PB-IPC. After the treatment with stem cell educator therapy, the percentages of GLUT2⁺CD45RO⁺ PB-IPC were markedly increased in these T1D subjects. Notably, we found that T1D patients shared common clinical features with patients with other autoimmune and inflammation-associated diseases, such as displaying low or no expression of GLUT2 on PB-IPC at baseline and exhibiting a high profile of the inflammatory cytokine interleukin (IL)-1β. Flow cytometry demonstrated that their GLUT2 expressions on PB-IPC were also markedly upregulated, and the levels of IL-1β-positive cells were significantly downregulated after the treatment with stem cell educator therapy. Stem cell educator therapy could upregulate the GLUT2 expression on PB-IPC and restore their function in T1D patients, leading to the improvement of clinical outcomes. The clinical data advances current understanding about the molecular mechanisms underlying the stem cell educator therapy, which can be expanded to treat patients with other autoimmune and inflammation-associated diseases. Full article

In the near future, international space agencies have planned to achieve significant milestones in investigating the utilization of Global Navigation Satellite Systems (GNSS) within and beyond the current space service volume up to their application to lunar missions. These initiatives aim to demonstrate the feasibility of GNSS navigation at lunar altitudes. Based on the outcomes of such demonstrations, dozens of lunar missions will likely be equipped with a GNSS receiver to support autonomous navigation in the lunar proximity. Relying on non-invasive, consolidated differential techniques, GNSS will enable baseline estimation, thus supporting a number of potential applications to lunar orbiters such as collaborative navigation, formation flight, orbital manoeuvers, remote sensing, augmentation systems and beyond. Unfortunately, the large dynamics and the geometry of such differential GNSS scenarios set them apart from current terrestrial and low-earth orbit use cases. These characteristics result in an increased sensitivity to measurements time misalignment among orbiters. Hence, this paper offers a review of baseline estimation methods and characterizes the divergences and limitations w.r.t. to terrestrial applications. The study showcases the estimation of the baseline length between a lunar CubeSat mission, VMMO, and the communication relay Lunar Pathfinder mission. Notably, real GNSS measurements generated by an Engineering Model of the NaviMoon receiver in the European Space Agency (ESA/ESTEC) Radio Navigation Laboratory are utilized. A radio-frequency constellation simulator is used to generate the GNSS signals in these hardware-in-the-loop tests. The performed analyses showed the invalidity of common terrestrial differential GNSS ranging techniques for space scenarios due to the introduction of significant biases. Improved ranging algorithms were proposed and their potential to cancel ranging errors common to both receivers involved was confirmed. Full article

In the domain of automated story generation, the intricacies of the Arabic language pose distinct challenges. This study introduces a novel methodology that moves away from conventional event-driven narrative frameworks, emphasizing the restructuring of narrative constructs through sophisticated language models. Utilizing mBERT, our approach begins by extracting key story entities. Subsequently, XLM-RoBERTa and a BERT-based linguistic evaluation model are employed to direct beam search algorithms in the replacement of these entities. Further refinement is achieved through Low-Rank Adaptation (LoRA), which fine-tunes the extensive 3 billion-parameter BLOOMZ model specifically for generating Arabic narratives. Our methodology underwent thorough testing and validation, involving individual assessments of each submodel. The ROCStories dataset provided the training ground for our story entity extractor and new entity generator, and was also used in the fine-tuning of the BLOOMZ model. Additionally, the Arabic ComVE dataset was employed to train our commonsense evaluation model. Our extensive analyses yield crucial insights into the efficacy of our approach. The story entity extractor demonstrated robust performance with an F-score of 96.62%. Our commonsense evaluator reported an accuracy of 84.3%, surpassing the previous best by 3.1%. The innovative beam search strategy effectively produced entities that were linguistically and semantically superior to those generated using baseline models. Further subjective evaluations affirm our methodology’s capability to generate high-quality Arabic stories characterized by linguistic fluency and logical coherence. Full article

The emotional support dialogue system is an emerging and challenging task in natural language processing to alleviate people’s emotional distress. Each utterance in the dialogue has features such as emotion, intent, and commonsense knowledge. Previous research has indicated subpar performance in strategy prediction accuracy and response generation quality due to overlooking certain underlying factors. To address these issues, we propose Advanced Multi-Task Learning and Feature-Fusion for Emotional Support Conversation (AdMISC), which extracts various potential factors influencing dialogue through neural networks, thereby improving the accuracy of strategy prediction and the quality of generated responses. Specifically, we extract features affecting dialogue through dynamic emotion extraction and commonsense enhancement and then model strategy prediction. Additionally, the model learns these features through attention networks to generate higher quality responses. Furthermore, we introduce a method for automatically averaging loss function weights to improve the model’s performance. Experimental results using the emotional support conversation dataset ESConv demonstrate that our proposed model outperforms baseline methods in both strategy label prediction accuracy and a range of automatic and human evaluation metrics. Full article