Search Results (239)

The advancement of generative artificial intelligence (GAI) in healthcare is driving a transition from dialogue-based medical chatbots to workflow-embedded clinical AI agents. These agentic systems incorporate persistent state management, coordinated tool invocation, and bounded autonomy, enabling multi-step reasoning within institutional processes. As a result, traditional response-level evaluation frameworks are insufficient for understanding system behavior. This review provides a conceptual synthesis of the evolution from conversational systems to agentic architectures and proposes a system-level modeling framework for ethical clinical AI agents. We identify core architectural dimensions, including autonomy gradients, state persistence, tool orchestration, workflow coupling, and human–AI co-agency, and examine how these features reshape bias propagation pathways, error cascade dynamics, trust calibration, and accountability structures. Emphasizing that ethical risks emerge from longitudinal system interactions rather than isolated outputs, we argue for embedding fairness constraints, transparency mechanisms, and lifecycle governance directly within AI design. By outlining trajectory-level evaluation strategies, equity-aware development approaches, collaborative oversight models, and adaptive regulatory frameworks, this paper establishes a foundation for the responsible and trustworthy integration of agentic AI in healthcare. Full article

Referring Remote Sensing Image Segmentation (RRSIS) is a representative multimodal understanding task for remote sensing, which segments designated targets from remote images according to free-form natural language descriptions. However, complex remote sensing characteristics, such as cluttered backgrounds, large-scale variations, small scattered targets and repetitive textures, lead to unstable visual grounding and further spatial grounding drift, resulting in inaccurate segmentation results. Existing approaches typically perform implicit visual–linguistic fusion across encoding and decoding stages, entangling spatial grounding with mask refinement. This tightly coupled formulation lacks explicit structural constraints and is prone to cross-modal ambiguity, especially in complex remote sensing layouts. To address these limitations, we propose a Structurally consistent and Grounding-aware Stagewise Reasoning Framework (SGSRF) that follows a grounding-first, segmentation-second paradigm. The framework decomposes inference into three cascaded stages with progressively imposed structural constraints. First, Cross-modal Consistency Refinement (CCR) lays the foundation for stable spatial grounding by enhancing visual–textual structural alignment via CLIP-based features and Structural Consistency Regularization (SCR), producing well-aligned multimodal representations and reliable grounding cues. Second, Grounding-aware Prompt (GPG) Generation bridges grounding and segmentation by converting aligned representations into complementary sparse and dense prompts, which serve as explicit grounding guidance for the segmentation model. Third, Grounding Modulated Segmentation (GMS) leverages the Segment Anything Model (SAM) to generate fine-grained mask prediction under the joint guidance of prompts and grounding cues, improving spatial grounding stability and robustness to background interference and scale variation. Extensive experiments on three remote sensing benchmarks, namely RefSegRS, RRSIS-D, and RISBench, demonstrate that SGSRF achieves state-of-the-art performance. The proposed stagewise paradigm integrates structural alignment, explicit grounding, and prompt-driven segmentation into a unified framework, providing a practical and robust solution for RRSIS in real-world Earth observation applications. Full article

Anti-unmanned aerial vehicle (Anti-UAV) detection is critical for airspace security, yet existing single-modality approaches suffer from severe performance degradation under adverse illumination, thermal crossover, and extreme scale variation. In this paper, we propose CSFADet, a dual-modal detection framework that jointly exploits visible and infrared imagery through four tightly integrated modules. First, a Cross-Spectral Feature Alignment (CSFA) module performs early-stage spectral calibration by computing cross-modal query–value attention maps, generating modality-aware channel descriptors that re-weight and concatenate the two spectral streams. Second, a Dual-path Texture Enhancement Module (DTEM) enriches fine-grained spatial details via cascaded convolutions with residual connections. Third, a Dual-path Cross-Attention Module (DCAM) introduces a feature-shrinking token generation strategy followed by symmetric cross-attention branches with learnable scaling factors, Squeeze-and-Excitation recalibration, and a $1\times 1$ convolution fusion head, enabling deep bidirectional interaction between modalities. Fourth, a Dual-path Information Refinement Module (DIRM) embeds Adaptive Residual Groups (ARGs) that cascade Multi-modal Spatial Attention Blocks (MSABs) with channel and dynamic spatial attention, culminating in a Multi-scale Scale-aware Fusion Refinement (MSFR) unit that employs three parallel multi-head attention branches with a Scale Reasoning Gate and Channel Fusion Layer to produce scale-discriminative enhanced features. Experiments on the public Anti-UAV300 benchmark show that CSFADet achieves 91.4% mAP@0.5 and 58.7% mAP@0.5:0.95, surpassing fifteen representative detectors spanning single-stage, two-stage, YOLO-family, and Transformer-based categories. Ablation studies confirm the complementary contributions of each module, and heatmap visualizations verify the model’s capacity to focus on small, distant UAV targets under challenging conditions. Full article

The Democratic Republic of the Congo faces a high tuberculosis (TB) burden. In 2022, 61% of an estimated 402,000 TB cases were reported (World Health Organization Global tuberculosis report). To enhance case detection, the national TB program (NTP) introduced a program quality and efficiency approach (PQE), integrating systematic TB screening into outpatient departments (OPDs). Observational data of the PQE on the TB care cascade (from screening to treatment) across 70 sites in Kinshasa that initiated PQE during the first quarter of 2023 are presented. Data were collected monthly and validated during supervision visits, and disaggregated by sex, healthcare facility type (public, private, or faith-based), facility level (primary or secondary), and OPD within each facility. In 2024, 639,464 individuals were consulted in various OPDs in the participating facilities, 57% of which were female. The median number needed to screen (NNS) was 22.1, with an interquartile range of [9.5–104.3]. There was a significantly lower NNS observed in general practice and human immunodeficiency virus departments. Throughout the TB care cascade, women were less likely than men to be screened, tested, or treated. These findings, to be interpreted within the context of Kinshasa pilot facilities, provide insights to the NTP for developing PQE implementation research aimed at understanding the reasons for these discrepancies and informing NTP scale-up at the national level. Full article

Vision-based robot control remains a significant challenge due to the sample inefficiency and prolonged training times associated with traditional deep reinforcement learning methods. We propose a novel approach inspired by biological gene regulation, leveraging Gene Regulatory Networks (GRNs) for efficient and robust robot control. In our approach, robot states are encoded as gene expression levels, and evolutionary optimization is used to learn GRN parameters that map raw visual inputs to motor commands. We evaluate this method on the KukaDiverseObjectEnv benchmark, where robots must grasp diverse objects using only RGB images. Our GRN-based controller achieves a $57.5\%$ success rate while reducing training time by $13.7\times$ compared to Proximal Policy Optimization baselines. It also outperforms NEAT, standard reinforcement learning algorithms, and deep Q-learning in terms of both efficiency and performance. The controller maintains $91.8\%$ performance under noisy visual conditions. This bio-inspired design naturally enables hierarchical control via expression cascades, computational efficiency through bounded dynamics, and temporal reasoning without explicit memory modules. Full article

This review evaluates the emerging paradigm of multi-agent systems (MASs) for biomedical and clinical data analysis, focusing on their ability to overcome the reasoning and reliability limitations of standalone large language models (LLMs). We synthesize findings from recent architectural frameworks, specifically LangGraph, CrewAI, and the Model Context Protocol (MCP), to examine how specialized agent teams divide labor, utilize precision tools, and cross-verify outputs. We find that MAS architectures yield significant performance gains in various domains: recent implementations improved oncology decision-making accuracy from 30.3% to 87.2% and reached a peak of 93.2% accuracy on USMLE-style benchmarks through simulated clinical evolution. In clinical trial matching, multi-agent frameworks achieved 87.3% accuracy and enhanced clinician screening efficiency by 42.6% (p < 0.001). However, we also highlight critical operational challenges, including an unreliability tax of 15–50× higher token consumption compared to standalone models and the risk of cascading errors where initial hallucinations are amplified across the agent collective. We conclude that while MAS enables a shift toward collaborative intelligence in biomedicine, its clinical and research adoption requires the development of deterministic orchestration and rigorous cost-utility frameworks to ensure safety and expert-centered oversight. Full article

In the current context of environmental sustainability and reduced agricultural inputs, biostimulants represent one of the most efficient, eco-friendly and innovative strategies to preserve plants from biotic and abiotic stresses and to ensure sustainable agriculture. Ranging from benefic microorganisms, seaweed extracts, and humic acids to complex carbohydrates such as polysaccharides and oligosaccharides, these biostimulants are able to increase plant growth, photosynthetic efficiency, root development and nutrient uptake when they are applied during seed priming as foliar sprays or as liquid and solid soil amendments. The mechanisms underlying their effective action on plants are mainly related to the enhancement of antioxidant defenses and the regulation of hormonal pathways, particularly auxin homeostasis and transport. Several studies reported the relevance of biostimulant application in promoting root growth. In plants, roots play crucial roles, performing a variety of functions such as nutrients and water uptake, mediating stress perception and adaptation, influencing the rhizosphere microbiome, and providing structural support. The effectiveness and perception of polysaccharide-based biostimulants (PBs) are highly dependent on crucial factors, including the degree of depolymerization and the chemical modifications such as acetylation, methylation, sulfation, and oxidation. Furthermore, not all receptors and co-receptors involved in the recognition of PBs have yet been identified. However, there remain many gaps in our understanding regarding the interaction between biostimulants and roots, which is still far from fully elucidated. For these reasons, the present review provides a comprehensive overview of current research on biostimulants–root interactions, with a particular focus on polysaccharide-based biostimulants. It highlights the mechanisms involved in their recognition by plants roots, from perception to response, and the subsequent signaling cascades and the molecular pathways activated, with special emphasis on existing knowledge gaps and future research perspectives. Full article

Electrochemical energy storage is pivotal in constructing new-type power systems. However, the large-scale deployment of energy storage stations poses severe safety challenges, particularly the risk of deflagration. The coupling of combustible accumulation within battery systems and the confined structure of storage units can trigger cascading thermal runaway and deflagration accidents. Existing research still falls short in systematically analyzing the deflagration risks and process evolution mechanisms in energy storage stations. To address this gap, this study develops a probabilistic risk assessment model that enables analysis of risk propagation through the integration of fault tree analysis (FTA) with a static fuzzy Bayesian network (BN). The proposed approach delineates the complete risk evolution pathway from battery thermal runaway to deflagration in a confined space. Diagnostic reasoning identifies a dominant risk escalation path initiated by internal short circuits, leading to thermal runaway, flammable gas release, and pressure accumulation due to inadequate pressure relief. Sensitivity analysis highlights gases ejected during thermal runaway (C22) and lack of pressure relief devices or insufficient venting area (C31) as the most influential risk drivers. This study thus offers a practical, model-based framework for enhancing targeted risk prevention and safety resilience in electrochemical energy storage station infrastructure. Full article

(1) Background: Post-arrival screening for infectious diseases is routinely offered to newly arrived refugees in Denmark, including tests for hepatitis B virus (HBV), hepatitis C virus (HCV), human immunodeficiency virus (HIV), and syphilis. This study aimed to examine the cascade of care following positive screening results in a local cohort of refugees in Denmark, with a focus on subsequent clinical management, follow-up, and outcomes. (2) Methods: This retrospective cohort study included 1506 newly arrived refugees of all ages and countries of origin. All were offered a post-arrival infectious disease screening in Denmark. Clinical records were reviewed to assess progression through the cascade of care, including referral, evaluation, follow-up, and clinical outcomes among individuals with positive screening results. (3) Results: Of the 1506 screened refugees, 33 (2.2%) had at least one positive screening result. Among the 15 individuals with detectable hepatitis B surface antigen, six (43%) attended regular follow-up, while eight (57%) were lost during the cascade of care. Two participants screened positive for HCV antibodies; both underwent initial clinical evaluation, but their subsequent care trajectories differed due to repeated non-attendance or undocumented reasons. Only one participant with non-specific syphilis antibodies completed follow-up in accordance with national guidelines. One participant was diagnosed with HIV and successfully linked to care. (4) Conclusions: The prevalence of screened infectious diseases in this local Danish refugee cohort was low and consistent with findings from comparable settings. Although post-arrival screening facilitates the identification of infectious diseases, substantial loss to follow-up occurred after initial diagnosis, limiting the effectiveness of follow-up and treatment. These findings highlight the need for targeted, interdisciplinary strategies addressing organisational, social, and individual barriers to improve continuity of care following screening. Full article

Existing inspection systems for railway point machines often suffer from high latency and poor interpretability, which impedes the real-time detection of critical mechanical anomalies, thereby increasing the risks of derailment and leading to cascading schedule delays. Addressing these challenges, this study proposes a lightweight computer vision-based detection framework deployed on the RK3588S edge platform. First, to overcome the accuracy degradation of segmentation networks on constrained edge NPUs, a Sensitivity-Aware Mixed-Precision Quantization and Heterogeneous Scheduling (SMPQ-HS) strategy is proposed. Second, a Multimodal Semantic Diagnostic Framework is constructed. By integrating geometric engagement depths—calculated via perspective rectification—with visual features, a Hard-Constrained Knowledge Embedding Paradigm is designed for the Qwen2.5-VL model. This approach constrains the stochastic reasoning of the Qwen2.5-VL model into standardized diagnostic conclusions. Experimental results demonstrate that the optimized model achieves an inference speed of 38.5 FPS and an mIoU of 0.849 on the RK3588S, significantly outperforming standard segmentation models in inference speed while maintaining high precision. Furthermore, the average depth-estimation error remains approximately 3%, and the VLM-based fault identification accuracy reaches 88%. Overall, this work provides a low-cost, deployable, and interpretable solution for intelligent point machine maintenance under edge-computing constraints. Full article

Synthetic aperture radar (SAR), with its all-weather and all-day observation capabilities, plays a significant role in the field of remote sensing. However, due to the unique imaging mechanism of SAR, its interpretation is challenging. Translating SAR images into optical remote sensing images has become a research hotspot in recent years to enhance the interpretability of SAR images. This paper proposes a deep learning-based method for SAR-to-optical remote sensing image translation. The network comprises three parts: a global representor, a generator with cascaded multi-head attention, and a multi-scale discriminator. The global representor, built upon InternImage with deformable convolution v3 (DCNv3) as its core operator, leverages its global receptive field and adaptive spatial aggregation capabilities to extract global semantic features from SAR images. The generator follows the classic “encoder-bottleneck-decoder” structure, where the encoder focuses on extracting local detail features from SAR images. The cascaded multi-head attention module within the bottleneck layer optimizes local detail features and facilitates feature interaction between global semantics and local details. The discriminator adopts a multi-scale structure based on the local receptive field PatchGAN, enabling joint global and local discrimination. Furthermore, for the first time in SAR image translation tasks, structural similarity index metric (SSIM) loss is combined with adversarial loss, perceptual loss, and feature matching loss as the loss function. A series of experiments demonstrate the effectiveness and reliability of the proposed method. Compared to mainstream image translation methods, our method ultimately generates higher-quality optical remote sensing images that are semantically consistent, texturally authentic, clearly detailed, and visually reasonable appearances. Full article

In an increasingly interconnected world, the capacity of societies to withstand, adapt to, and recover from crises is a central challenge for security and sustainable development. Yet, despite extensive research on resilience, the mechanisms through which systemic vulnerabilities emerge and propagate across social domains remain poorly understood. This paper addresses this gap by proposing a network-based framework: the Complex Analysis for Socio-Environmental Adaptation (CASA), which models resilience as a graph-structured system. Each node in CASA represents a social or infrastructural component whose resistance is derived from indicators of installed capacities, while edges capture interdependencies among sectors. We formalize a damage propagation model in which the loss of capacity in one node dynamically affects connected components, revealing the topological patterns that drive systemic fragility. Comparative simulations demonstrate that CASA’s topology amplifies the impact of highly connected nodes, rendering them crucial for resilience planning. An application to a real-world case demonstrates how initial disruptions in access to drinking water cascade into governance, economic, and social instabilities. The results provide both theoretical and operational insights, highlighting that resilience depends not only on the strength of individual components but also on the network architecture that links them. CASA thus offers a replicable and data-informed approach for identifying drivers of social conflict and guiding anticipatory resilience strategies in complex territorial systems. Full article

Urban flood disasters have become one of the most significant natural hazards under the dual pressures of rapid urbanization and intensified climate change. With the increasing interconnection among urban subsystems, these disasters often evolve into urban flood disaster chains, characterized by cascading failures across infrastructure, environment, and society. Current research hotspots mainly focus on three key aspects: the formation mechanisms, identification methods, and risk assessment approaches of urban flood disaster chains. In terms of formation mechanisms, most studies qualitatively describe the triggering and transmission processes of cascading events, revealing how interactions among hazard-inducing factors, disaster-formative environments, and disaster receptor generate chain reactions. Identification methods are categorized into four paradigms: qualitative identification based on experiential reasoning, semantic identification driven by data, structural identification through model inference, and behavioral identification using simulation modeling. Risk assessment approaches include historical disaster analysis, indicator-based evaluation models, uncertainty models, numerical simulation models, and intelligent algorithm models that integrate machine learning with physical simulations. The review finds that, due to the scarcity and heterogeneity of disaster chain event data, existing studies lack a unified quantitative framework to represent the mechanisms of urban flood disaster chains, as well as dynamic identification and assessment methods that can adapt to their evolutionary processes. Future research should focus on developing integrated mathematical paradigms, enhancing multisource data fusion and causal reasoning, and constructing hybrid models to support real-time risk assessment for urban flooding disaster chains. Full article

The analysis of changing trends of river runoff and sediment discharge and the exploration of their causes are of great significance for formulating sustainable development measures for river basin systems. Based on methods such as trend test, mutation detection, and regression analysis, this study conducts a systematic comparative research on the water–sediment processes in the river reach where large-scale cascaded reservoirs connected end to end are located in the upper Yangtze River, and obtains the following key research progress: For the study reach (between Sanduizi and Xiangjiaba Stations), during the period of 1966–2023, the change rates of annual incoming and outgoing runoff were 2.88 × 10⁸ m³·yr⁻¹ and −0.186 × 10⁸ m³·yr⁻¹, respectively, accounting for 0.017% and 0.013% of the annual average runoff. The changing trends were not significant. During the same period, the change rates of Suspended Sediment Load (SSL) at the inlet and outlet of this river reach were −8.0 × 10⁵ t·yr⁻¹ and −46 × 10⁵ t·yr⁻¹, respectively, accounting for 1.25% and 2.45% of their respective annual average sediment discharge. The SSL showed a significant decreasing trend, which was particularly characterized by a sharp reduction at the outlet. The massive sediment retention and multi-mode operation of cascaded reservoirs are the fundamental reasons for the variation in the water–sediment relationship and the sharp decrease in annual SSL in this reach, and they also lead to an obvious adjustment of water and sediment in the river basin that “cuts peaks and fills valleys” within a year. Climate change and other human activities have reduced the sediment input in the study reach. Looking forward to the next few decades, climate factors will remain the dominant factor affecting the inter-annual variation in runoff in the study area. In contrast, human activities such as reservoir operation will continue to fully control the sediment output of the river reach and also restrict the annual distribution of water and sediment. The results of this study can provide a reference for predicting the changing trends of water and sediment in similar river reaches with cascaded reservoir groups and formulating effective river management measures. Full article

Crisis management in smart cities demands coherent, interoperable, and reusable semantic models to represent complex systems, their risks, crisis situations, interdependencies, and decision-making processes across all lifecycle phases, i.e., prevention, preparedness, response, and recovery. This paper presents the TERMINUS (TERritorial Management and INfrastructures ontology for institutional and industrial USage) ontology, a BFO (Basic Formal Ontology)-aligned conceptual model based on semantic patterns for the crisis management lifecycle operationalized as both ontology design patterns (ODPs) to structure the ontology and ontology query patterns (OQPs) to use it in specific contexts. ODPs capture reusable conceptual structures for modeling domains, while OQPs provide SPARQL (SPARQL Protocol and RDF Query Language)-based templates to retrieve and reason over knowledge graph instances derived from these model chunks. The approach ensures semantic continuity from conceptual modeling to operational applications, enabling automated scenario generation, cascading risk analysis, and participatory decision-making. We position the patterns within the crisis management lifecycle and demonstrate their use through real-world case studies, covering semantic spatio-temporal risk assessment, interdependent infrastructure risk cascades, creative emergency scenario design, and recovery planning. Evaluation results highlight the ontology’s ability to support domain experts in generating plausible context-specific models, fostering collaborative validation, and enhancing preparedness and resilience. TERMINUS thus provides a versatile and interoperable semantic infrastructure for integrating ontologies and knowledge graphs into urban crisis management workflows. Full article