The clinical burden of spleen-stomach disorders is substantial. While large language models (LLMs) offer new potential for medical applications, they face three major challenges in the context of integrative Chinese and Western medicine (ICWM): a lack of high-quality data, the absence of models capable of effectively integrating the reasoning logic of traditional Chinese medicine (TCM) syndrome differentiation with that of Western medical (WM) disease diagnosis, and the shortage of a standardized evaluation benchmark. To address these interrelated challenges, we propose DongYuan, an ICWM spleen-stomach diagnostic framework. Specifically, three ICWM datasets (SSDF-Syndrome, SSDF-Dialogue, and SSDF-PD) were curated to fill the gap in high-quality data for spleen-stomach disorders. We then developed SSDF-Core, a core diagnostic LLM that acquires robust ICWM reasoning capabilities through a two-stage training regimen of supervised fine-tuning. tuning (SFT) and direct preference optimization (DPO), and complemented it with SSDF-Navigator, a pluggable consultation navigation model designed to optimize clinical inquiry strategies. Additionally, we established SSDF-Bench, a comprehensive evaluation benchmark focused on ICWM diagnosis of spleen-stomach disorders. Experimental results demonstrate that SSDF-Core significantly outperforms 12 mainstream baselines on SSDF-Bench. DongYuan lays a solid methodological foundation and provides practical technical references for the future development of intelligent ICWM diagnostic systems.

Background:Infectious diseases remain a major source of health loss, yet long-term, cause-comparable assessments that jointly integrate incidence, disability-adjusted life-years (DALYs), age patterns, and risk attribution for China are fragmented and not comprehensively synthesized within a unified framework. We aimed to evaluate the trends in incidence and DALYs of infectious diseases in China from 2000 to 2023 to inform current priority setting by clarifying long-term trends and changes observed during the COVID-19 pandemic. Methods:Data on the number and rate of incidence and DALYs of five major infectious-disease cause groups (HIV/AIDS and sexually transmitted infections, respiratory infections and tuberculosis, enteric infections, neglected tropical diseases and malaria, and other infectious diseases) and 40 specific infectious diseases in China were obtained from the Global Burden of Disease Study (GBD) 2023. Estimated annual percentage changes (EAPCs) in age-standardized incidence rates (ASIRs) and age-standardized DALY rates (ASDRs) were calculated overall and by sex and age group to quantify temporal trends from 2000 to 2023. We conducted comparative risk assessment using population attributable fractions (PAFs) to quantify risk-attributable DALYs, and decomposition analyses to assess the contributions of population growth, population ageing and epidemiologic change to changes in incident cases, deaths, and DALYs. Sensitivity analyses were also conducted to examine the robustness of the results. Findings:From 2000 to 2023, age-standardized DALY rates declined in four of the five major infectious disease cause groups in China. The largest decrease was observed for enteric infections, with DALY rates falling from 365.4 to 43.8 per 100,000 (EAPC -11.0% [95% CI -12.2 to -9.7]). Among all groups, respiratory infections and tuberculosis had the highest DALY burden throughout the study period, while their DALY rates decreased from 2559.5 to 688.1 per 100,000 (EAPC -8.1% [-9.2 to -7.0]). It's estimated that HIV/AIDS and sexually transmitted infections were the only cause group with an increasing DALY rate, rising from 79.7 to 103.7 per 100,000 (EAPC 1.0% [0.6 to 1.4]). The increase was greatest among adults aged 20-54 years, in whom DALY rates rose from 71.1 to 152.0 per 100,000 (EAPC 3.3% [2.7 to 3.9]). Risk attribution varied by cause and age group. Unsafe sex accounted for most DALYs for sexually transmitted infections excluding HIV and more than 70% of HIV/AIDS DALYs. For lower respiratory infections, child and maternal malnutrition accounted for 42.2% of DALYs in those aged < 20 years, whereas tobacco use accounted for 38.1% in adults aged 20-54 years. Enteric infection DALYs were mainly attributable to unsafe water, sanitation, and handwashing, accounting for more than 70% across age groups. Decomposition analysis showed that epidemiologic change was the main driver of changes in infectious disease burden from 2000 to 2023. Interpretation:China's infectious disease DALY burden declined from 2000 to 2023, but progress was uneven across causes and increasingly concentrated in older adults. Persistent incidence-DALY discordance and heterogeneous age-specific risk profiles indicate that current priority setting should move beyond incidence-led targets towards reducing preventable disability and premature mortality through age- and mechanism-tailored prevention and care. Funding:Prevention and Control of Emerging and Major Infectious Diseases-National Science and Technology Major Project.

Mpox resurged globally in 2022 after five decades of African confinement. Pooling 70 studies (1970–2025) and WHO surveillance, we show hospitalisation and case-fatality fell from 33% and 3.3% pre-2022 to 10% and 0.7% during the clade-IIb pandemic. Clade-Ia remained high deadlier (≈80% admission, 9% death). Since 2024, Central Africa has again driven cases and deaths. High-income regions recorded milder, larger outbreaks; Africa continues to bear excess mortality. Targeted clade-Ia surveillance, vaccination and health-system support in Central Africa are urgently needed to curb mpox’s persistent global burden.

Objective:Traditional disease prevention strategies that rely on fixed parameters and macro-level models struggle to capture the diversity of individual behaviors and environmental complexities. Indoor spaces with high population densities and poor ventilation, such as schools and hospitals, are particularly vulnerable to pathogen transmission. The coronavirus disease (COVID-19) pandemic highlighted the need for precise intervention strategies. Methods:We developed a spatial-individual agent-based model that integrates fine-grained spatiotemporal dynamics, where transmission risk is quantified by the exact distance and duration of contact. This model was applied to a high-resolution case study of a university dormitory floor to evaluate various testing frequencies, scopes, and isolation intensities. Results:Simulations showed that a dormitory-wide isolation policy outperformed individual restrictions by protecting uninfected rooms. Counter-intuitively, every-three-day class-based testing lowered infection risks compared to daily class-based testing by minimizing high-density interactions. In spatially constrained environments, stricter isolation reduces the overall outbreak duration but increases the contact transmission rate among individuals sharing the same enclosed space. Conclusion:Epidemic control in high-density environments requires balancing testing frequency and isolation stringency based on spatial constraints. Under strict isolation, frequent testing is vital for breaking transmission chains. In less restrictive settings, moderately reducing the testing frequency minimizes unnecessary contact. These findings provide data-driven guidance for optimizing public health policies on campuses.

Knowledge graph completion (KGC) addresses the issue of incomplete knowledge graphs by inferring missing triples, which is crucial for information retrieval, question answering, and recommender systems. Existing KGC methods generally focus on either exploiting the graph’s structural topology or leveraging the semantic information from entity descriptions. However, existing approaches often overlook the synergy between structure information and semantic information. To address the existing shortcomings, we propose StrucSem, a novel model that leverages both structural and semantic information to boost KGC performance. Our model: 1) encodes the graph’s structural information by aggregating neighborhood data around the query entity using an attention mechanism; and 2) combines this with the encoding of textual descriptions, facilitating the integration of both types of information. Additionally, we extend contrastive learning to incorporate multiple positive samples, improving the model’s ability to represent diverse relational patterns. Our approach significantly enhances KGC performance, as demonstrated through extensive evaluations on standard benchmark datasets. The results highlight the superiority of combining structural and semantic information, offering new insights into improving KGC tasks.

The accurate prediction of infectious disease transmission patterns is a cornerstone of computational epidemiology, with profound implications for public health planning and intervention strategies. Traditional epidemiological and statistical models, along with recent deep learning architectures, typically operate under discrete-time assumptions, inherently limiting their capacity to represent the continuous and nonlinear nature of epidemic processes. Here, we present a novel Spectral-Temporal Fusion Neural Ordinary Differential Equation (STF-NODE) framework, which significantly enhances the predictive accuracy of epidemic models by integrating temporal information with frequency-domain features extracted via adaptive spectral analysis. By decomposing historical epidemic data into distinct frequency components, our approach effectively captures transient, high-frequency events such as super-spreading phenomena and simultaneously stabilizes long-term trend predictions through low-frequency regularization. Comprehensive experiments conducted on real-world datasets—including hand, foot, and mouth disease and COVID-19—demonstrate that STFNODE consistently outperforms established baseline methods across multiple forecast horizons.

This study investigates the evolving dynamics of public attention toward multiple influenza subtypes in China from 2019 to early 2025, using Baidu Search Index (BSI) as a proxy for collective health awareness. Drawing upon official surveillance data from the Chinese Center for Disease Control and Prevention (CDC), we examine the temporal and directional relationships between public search behavior and influenza epidemic indicators across multiple time scales and platforms. Using correlation analysis, lagged cross-correlation, and Granger causality modeling, we identify both synchronous co-movements and predictive lead-lag structures between BSI and surveillance-based indicators. Influenza A, Swine, and Avian Influenza consistently emerge as early signals of rising public concern, with Swine and Avian strains often showing directional influence on broader influenza-related search trends. In contrast, Influenza B remains peripheral, with weaker and less interconnected public attention. Our findings highlight a post-pandemic amplification of influenza awareness during the 2023-2024 season, marked by historically high search volumes and the re-establishment of predictive lag structures, particularly at Lag = 2-3 weeks. Regional disparities are evident: southern provinces exhibited more resilient BSI-CDC coherence during the pandemic, while northern regions showed delayed recovery. These results underscore the utility of search engine data for real-time infodemiological surveillance and early warning. From a policy perspective, integrating BSI into national influenza monitoring frameworks can enhance the timeliness and responsiveness of public health interventions. This study provides actionable insights for designing adaptive, multi-scale digital surveillance systems in the post-pandemic era.

Large language models (LLMs) represent a novel technological species in the realm of general intelligence. Their problem-solving approach is not based on “first principles” (logocentrism) but rather on empirical learning from observed data. LLMs possess the ability to extract intuitive knowledge from vast amounts of data, enabling them to offer flexible and effective solutions in the face of complex and dynamic scenarios. The general intelligence characteristics of LLMs are mainly reflected in three aspects: technologically, they exhibit common sense, deep reasoning, strong generalization, and natural human-computer interaction; in terms of intelligence, they demonstrate memory-driven core features, powerful data-driven learning capabilities, and exceptional generalization abilities; in terms of thought, they possess highly humanlike cognitive traits such as contextual understanding, analogy, and intuitive reasoning. These capabilities collectively suggest that LLMs can adapt to a wide range of complex, open-ended scenarios, presenting a stark contrast to traditional models that emphasize formal logic, quantitative analysis, and narrowly defined problem structures. As such, the rise of LLMs is likely to drive significant shifts in AI theory and application, potentially redefining how intelligent systems approach decision-making, strategic reasoning, and contextual understanding in uncertain and dynamic environments.

The accelerating global spread of infectious diseases under urbanization highlights critical gaps in conventional epidemic forecasting. While traditional models struggle with complex spatiotemporal dynamics - especially in high-density, mobile regions like Beijing's Haidian District - our proposed ST-LLM4Epi framework (SpatioTemporal Large Language Model for Epidemiology) demonstrates how Large Language Models can transform prediction capabilities. Leveraging Large Language Models (LLMs), this study explores their potential in epidemic prediction by integrating spatial and temporal data. ST-LLM4Epi's ability to infer adjacency information and generalize from sparse datasets enhances robustness compared to conventional methods. The research aims to improve early warning systems and support public health decision-making, ultimately strengthening preparedness for future outbreaks.

Joint extraction of entities and relations has attracted increasing attention in information extraction, yet most existing methods struggle to handle discontinuous entities—a structure frequently observed in real-world texts. In this work, we present a novel approach that formulates joint entity and relation extraction as a character-level table-filling task. To handle discontinuous structures, we introduce a unified label space that encodes both entity spans and relation links between character pairs. On top of this, we design a neural architecture combining BERT-BiLSTM embeddings, multi-dilation convolutions for capturing multi-scale context, and a co-predictor integrating biaffine and MLP classifiers. To support evaluation, we construct a large-scale dataset from COVID-19 epidemiological reports, annotated with 12 entity types and 8 relation types, including a substantial proportion of discontinuous entities. Experimental results on both this dataset and the SciERC benchmark show that our model achieves strong performance and generalizes well across domains.

The spread of infectious diseases is inherently linked to human social behavior, characterized by complexity, diversity, and openness. Intelligent agents in computer science provide a powerful framework for capturing such dynamics, enabling complex epidemic patterns to emerge from simple local rules. These agents exhibit self-organization, adaptability, and self-optimization, making them well suited for individual-level modeling. Agent-based models (ABMs) have shown promising results in epidemic simulation and policy evaluation. However, current implementations often suffer from simplistic behavioral assumptions and rigid interaction mechanisms, limiting their realism and flexibility. This paper first reviews the current landscape of epidemic modeling approaches. It then analyzes the underlying mechanisms of advanced intelligent agents, highlighting their modeling capabilities. The study focuses on four key advantages of intelligent agent-based modeling and elaborates on three critical roles these agents play in evaluating and optimizing intervention strategies.

Objective:Infectious diseases controlling system is indispensable for weaken the damage to the people's life and property security caused by infectious diseases. An effective infectious diseases controlling system must incorporate an early warning mechanism designed to detect abnormal rising trends (outbreak) in spatial-temporal series. However, existing anomaly detection methods are often constrained by the quality and quantity of available data in specific application scenarios, particularly in infectious diseases early warning scenarios. Methods:The emergence of generative pre-trained large time series models-hereafter referred to as large time series models-may provide a solution to this challenge. Based on these models, we propose an effective early warning framework. Results:We compared the framework with statistic and deep learning methods on real-world infectious diseases datasets and related derived datasets. Our framework has a better performance and requires less data. Conclusion:We propose a readily deployable early warning framework characterized by strong generalization ability and exceptional performance, which would enlighten the epidemic modeling researchers.

The increasing frequency and complexity of infectious disease outbreaks, exemplified by the COVID-19 pandemic, underscore the urgent need for accurate and adaptive epidemic forecasting models. Recently, spatio-temporal graph neural networks have shown potential in modeling infectious disease spread, as they effectively capture the interplay between spatial and temporal dependencies. However, existing STGNN-based approaches often treat disease transmission as a single, unified process, overlooking the distinct mechanisms underlying local and cross-regional spread. In this study, we propose D-STEM, a novel framework that decouples local and spillover transmission dynamics to achieve more precise predictions. Our approach integrates a hybrid local evolution module, combining GRU and self-attention mechanisms to model intra-regional transmission, and employs dynamic spatio-temporal convolution alongside population mobility networks to capture cross-regional transmission. Crucially, we introduce physical constraints to guide the disentanglement of these two mechanisms, ensuring that each module learns distinct features even in the absence of direct observations. Extensive experiments on real-world datasets, including US-state and Japan-prefecture COVID-19 data, demonstrate that D-STEM consistently outperforms all baselines. Our framework advances epidemic forecasting and provides actionable insights for public health interventions.

Decentralized trading schemes involving energy prosumers have prevailed in recent years. Such schemes provide a pathway for increased energy efficiency and can be enhanced by the use of blockchain technology to address security concerns in decentralized trading. To improve transaction security and privacy protection while ensuring desirable social governance, this article proposes a novel two-stage blockchain-based operation and trading mechanism to enhance energy hubs connected with integrated energy systems (IESs). This mechanism includes multienergy aggregators (MAGs) that use a consortium blockchain and its enabled proof-of-work (PoW) to transfer and audit transaction records, with social governance principles for guiding prosumers' decision-making in the peer-to-peer (P2P) transaction management process. The uncertain nature of renewable generation and load demand are adequately modeled in the two-stage Wasserstein-based distributionally robust optimization (DRO). The practicality of the proposed mechanism is illustrated by several case studies that jointly show its ability to handle an increased renewable generation capacity, achieve a 16.7% saving in the audit cost, and facilitate 2.4% more P2P interactions. Overall, the proposed two-stage blockchain-based trading mechanism provides a practical trading scheme and can reduce redundant trading amounts by 6.5%, leading to a further reduction of the overall operation cost. Compared to the state-of-the-art benchmark methods, our mechanism exhibits significant operation cost reduction and ensures social governance and transaction security for IES and energy hubs.

As the world continues to transition towards cleaner and more efficient energy sources, the intricate interplay between water and energy in power systems has emerged as an essential and multifaceted relationship with profound implications for sustainable energy planning. This comprehensive exploration considers a diverse range of academic databases and synthesizes relevant research to systematically investigate the current state of knowledge on the water-energy nexus. By distilling key findings and concepts related to the water-energy nexus in power systems, this work underscores the pivotal role of water in power generation and the energy required for water treatment and distribution. Additionally, this exploration brings into focus the challenges that the water-energy nexus faces, including the far-reaching impacts of climate change and the potential of renewable energy solutions. The complex policy and regulatory frameworks that govern the water-energy nexus in power systems are also examined, highlighting the crucial need for integrated approaches in energy and water management. By identifying key areas for further research and emphasizing the urgency for innovative solutions, this exploration stresses the need to prioritize sustainable management of water and energy resources in an effective, efficient, and resilient manner.

The implementation of early warning systems is an indispensable strategy in the containment of infectious diseases. Prior research has extensively employed the Long Short-Term Memory (LSTM) or Transformer network, equipped with a prediction or reconstruction error. Despite these advancements, the majority of existing early warning systems continue to rely on traditional models devoid of artificial neural networks. A comparative analysis of various models is conducted, ranging from traditional methods to advanced machine learning networks. The findings revealed that in certain instances, the performance of complex models did not significantly surpass that of simpler methods. This insight is particularly beneficial for the development of cost-effective early warning systems.

Since the start of the COVID-19 pandemic, many firms have been shifting their supply chains away from countries with stringent control measures to mitigate supply-chain disruption. Nowadays, the global economy has reopened from the COVID-19 pandemic at various paces in different countries. Understanding how the global supply network evolves during and after the pandemic is necessary for determining the timing and speed of reopening. By harnessing the real-world and real-time global human movement and the latest macroeconomic data, we propose an evolutionary economic-epidemiological model to explore the evolutionary dynamics of the global supply network under various global reopening scenarios. We find that, for highly restrictive countries, the delay in reopening has limited public health benefits in the long run but leads to significant supply-chain loss. A longer duration of stringent control measures substantially hurts the profitability of firms in highly restrictive countries, leading to slower supply-chain recovery in 5 years. This research presents the first data-driven evidence of supply chain loss due to the timing and speed of reopening and sheds light on the post-pandemic supply-chain reformation and recovery. Insights learned from COVID-19 will also be a valuable policymaking reference for combating future infectious disease epidemics and geopolitical changes.

This study is based on baseline data and 2-year follow-up information from 145 heart failure patients in Guangxi, China, combined with a publicly available scientific dataset on Chinese heart failure patients. Multiple datasets of varying scales were constructed, and traditional Cox proportional hazards models, along with Logistic Regression, Random Forest, Support Vector Machine, XGRoost, Random Survival Forest, and Survival Support Vector Machine algorithms were employed to develop heart failure mortality risk prediction models. These models were used to identify and quantitatively evaluate both risk factors and protective factors for HF mortality. In terms of the outcome, XGBoost demonstrated superior performance on high-dimensional datasets with missing values, whereas Support Vector Machine exhibited stronger predictive capability on similarly scaled datasets without missing values. The results based on XGBoost model and evaluation based on SHAP values further confirmed that Glomerular Filtration Rate, Height and Glucose are critical predictors. For survival time analysis, the Cox models generally outperformed Random Survival Forest and Survival Support Vector Machine algorithms. The mutual validation of different modeling approaches can enhance the robustness and effectiveness of heart failure mortality risk prediction, better supporting clinical prevention and treatment decision-making.

Coronary artery disease (CAD) remains a major global health concern, significantly contributing to morbidity and mortality. This study aimed to investigate the co-occurrence patterns of diagnoses and comorbidities in CAD patients using a network-based approach. A retrospective analysis was conducted on 195 hospitalized CAD patients from a single hospital in Guangxi, China, with data collected on age, sex, and comorbidities. Network analysis, supported by sensitivity analysis, revealed key diagnostic clusters and comorbidity hubs, with hypertension emerging as the central node in the co-occurrence network. Unstable angina and myocardial infarction were identified as central diagnoses, frequently co-occurring with metabolic conditions such as diabetes. The results also highlighted significant age- and sex-specific differences in CAD diagnoses and comorbidities. Sensitivity analysis confirmed the robustness of the network structure and identified clusters, despite the limitations of sample size and data source. Modularity analysis uncovered distinct clusters, illustrating the complex interplay between cardiovascular and metabolic disorders. These findings provide valuable insights into the relationships between CAD and its comorbidities, emphasizing the importance of integrated, personalized management strategies. Future studies with larger, multi-center datasets and longitudinal designs are needed to validate these results and explore the temporal dynamics of CAD progression.