Current shipping,tourism,and resource development requirements call for more accurate predictions of the Arctic sea-ice concentration(SIC).However,due to the complex physical processes involved,predicting the spatiote...Current shipping,tourism,and resource development requirements call for more accurate predictions of the Arctic sea-ice concentration(SIC).However,due to the complex physical processes involved,predicting the spatiotemporal distribution of Arctic SIC is more challenging than predicting its total extent.In this study,spatiotemporal prediction models for monthly Arctic SIC at 1-to 3-month leads are developed based on U-Net-an effective convolutional deep-learning approach.Based on explicit Arctic sea-ice-atmosphere interactions,11 variables associated with Arctic sea-ice variations are selected as predictors,including observed Arctic SIC,atmospheric,oceanic,and heat flux variables at 1-to 3-month leads.The prediction skills for the monthly Arctic SIC of the test set(from January 2018 to December 2022)are evaluated by examining the mean absolute error(MAE)and binary accuracy(BA).Results showed that the U-Net model had lower MAE and higher BA for Arctic SIC compared to two dynamic climate prediction systems(CFSv2 and NorCPM).By analyzing the relative importance of each predictor,the prediction accuracy relies more on the SIC at the 1-month lead,but on the surface net solar radiation flux at 2-to 3-month leads.However,dynamic models show limited prediction skills for surface net solar radiation flux and other physical processes,especially in autumn.Therefore,the U-Net model can be used to capture the connections among these key physical processes associated with Arctic sea ice and thus offers a significant advantage in predicting Arctic SIC.展开更多
Sporadic E(Es)layers in the ionosphere are characterized by intense plasma irregularities in the E region at altitudes of 90-130 km.Because they can significantly influence radio communications and navigation systems,...Sporadic E(Es)layers in the ionosphere are characterized by intense plasma irregularities in the E region at altitudes of 90-130 km.Because they can significantly influence radio communications and navigation systems,accurate forecasting of Es layers is crucial for ensuring the precision and dependability of navigation satellite systems.In this study,we present Es predictions made by an empirical model and by a deep learning model,and analyze their differences comprehensively by comparing the model predictions to satellite RO measurements and ground-based ionosonde observations.The deep learning model exhibited significantly better performance,as indicated by its high coefficient of correlation(r=0.87)with RO observations and predictions,than did the empirical model(r=0.53).This study highlights the importance of integrating artificial intelligence technology into ionosphere modelling generally,and into predicting Es layer occurrences and characteristics,in particular.展开更多
A coupled thermal-hydro-mechanical cohesive phase-field model for hydraulic fracturing in deep coal seams is presented.Heat exchange between the cold fluid and the hot rock is considered,and the thermal contribution t...A coupled thermal-hydro-mechanical cohesive phase-field model for hydraulic fracturing in deep coal seams is presented.Heat exchange between the cold fluid and the hot rock is considered,and the thermal contribution terms between the cold fluid and the hot rock are derived.Heat transfer obeys Fourier's law,and porosity is used to relate the thermodynamic parameters of the fracture and matrix domains.The net pressure difference between the fracture and the matrix is neglected,and thus the fluid flow is modeled by the unified fluid-governing equations.The evolution equations of porosity and Biot's coefficient during hydraulic fracturing are derived from their definitions.The effect of coal cleats is considered and modeled by Voronoi polygons,and this approach is shown to have high accuracy.The accuracy of the proposed model is verified by two sets of fracturing experiments in multilayer coal seams.Subsequently,the differences in fracture morphology,fluid pressure response,and fluid pressure distribution between direct fracturing of coal seams and indirect fracturing of shale interlayers are explored,and the effects of the cluster number and cluster spacing on fracture morphology for multi-cluster fracturing are also examined.The numerical results show that the proposed model is expected to be a powerful tool for the fracturing design and optimization of deep coalbed methane.展开更多
Rock residual strength,as an important input parameter,plays an indispensable role in proposing the reasonable and scientific scheme about stope design,underground tunnel excavation and stability evaluation of deep ch...Rock residual strength,as an important input parameter,plays an indispensable role in proposing the reasonable and scientific scheme about stope design,underground tunnel excavation and stability evaluation of deep chambers.Therefore,previous residual strength models of rocks established were reviewed.And corresponding related problems were stated.Subsequently,starting from the effects of bedding and whole life-cycle evolution process,series of triaxial mechanical tests of deep bedded sandstone with five bedding angles were conducted under different confining pressures.Then,six residual strength models considering the effects of bedding and whole life-cycle evolution process were established and evaluated.Finally,a cohesion loss model for determining residual strength of deep bedded sandstone was verified.The results showed that the effects of bedding and whole life-cycle evolution process had both significant influences on the evolution characteristic of residual strength of deep bedded sandstone.Additionally,residual strength parameters:residual cohesion and residual internal friction angle of deep bedded sandstone were not constant,which both significantly changed with increasing bedding angle.Besides,the cohesion loss model was the most suitable for determining and estimating the residual strength of bedded rocks,which could provide more accurate theoretical guidance for the stability control of deep chambers.展开更多
Taking the second member of the Xujiahe Formation of the Upper Triassic in the Xinchang structural belt as an example,based on data such as logging,production,seismic interpretation and test,a systematic analysis was ...Taking the second member of the Xujiahe Formation of the Upper Triassic in the Xinchang structural belt as an example,based on data such as logging,production,seismic interpretation and test,a systematic analysis was conducted on the structural characteristics and evolution,reservoir diagenesis and densification processes,and types and stages of faults/fractures,and revealing the multi-stage and multi-factor dynamic coupled enrichment mechanisms of tight gas reservoirs.(1)In the early Yanshan period,the paleo-structural traps were formed with low-medium maturity hydrocarbons accumulating in structural highs driven by buoyancy since reservoirs were not fully densified in this stage,demonstrating paleo-structure control on traps and early hydrocarbon accumulation.(2)In the middle-late Yanshan period,the source rocks became mature to generate and expel a large quantity of hydrocarbons.Grain size and type of sandstone controlled the time of reservoir densification,which restricted the scale of hydrocarbon charging,allowing for only a small-scale migration through sand bodies near the fault/fracture or less-densified matrix reservoirs.(3)During the Himalayan period,the source rocks reached overmaturity,and the residual oil cracking gas was efficiently transported along the late-stage faults/fractures.Wells with high production capacity were mainly located in Type I and II fault/fracture zones comprising the late-stage north-south trending fourth-order faults and the late-stage fractures.The productivity of the wells was controlled by the transformation of the late-stage faults/fractures.(4)The Xinchang structural belt underwent three stages of tectonic evolution,two stages of reservoir formation,and three stages of fault/fractures development.Hydrocarbons mainly accumulated in the paleo-structure highs.After reservoir densification and late fault/fracture adjustment,a complex gas-water distribution pattern was formed.Thus,it is summarized as the model of“near-source and low-abundance hydrocarbon charging in the early stage,and differential enrichment of natural gas under the joint control of fault-fold-fracture complex,high-quality reservoirs and structural highs in the late stage”.Faults/fractures with well-coupled fault-fold-fracture-pore are favorable exploration targets with high exploration effectiveness.展开更多
BACKGROUND Gastrointestinal diseases have complex etiologies and clinical presentations.An accurate diagnosis requires physicians to integrate diverse information,including medical history,laboratory test results,and ...BACKGROUND Gastrointestinal diseases have complex etiologies and clinical presentations.An accurate diagnosis requires physicians to integrate diverse information,including medical history,laboratory test results,and imaging findings.Existing artificial intelligence-assisted diagnostic tools are limited to single-modality information,resulting in recommendations that are often incomplete and may be associated with clinical or legal risks.AIM To develop and evaluate a collaborative multimodal large language model(LLM)framework for clinical decision-making in digestive diseases.METHODS In this observational study,DeepGut,a multimodal LLM collaborative diagnostic framework,was developed to integrate four distinct large models into a four-tiered structure.The framework sequentially accomplishes multimodal infor-mation extraction,logical“chain”construction,diagnostic and treatment suggestion generation,and risk analysis.The model was evaluated using objective metrics,which assess the reliability and comprehensiveness of model-generated results,and subjective expert opinions,which examine the effectiveness of the framework in assisting physicians.RESULTS The diagnostic and treatment recommendations generated by the DeepGut framework achieved exceptional performance,with a diagnostic accuracy of 97.8%,diagnostic completeness of 93.9%,treatment plan accuracy of 95.2%,and treatment plan completeness of 98.0%,significantly surpassing the capabilities of single-modal LLM-based diagnostic tools.Experts evaluating the framework commended the completeness,relevance,and logical coherence of its outputs.However,the collaborative multimodal LLM approach resulted in increased input and output token counts,leading to higher computational costs and extended diagnostic times.CONCLUSION The framework achieves successful integration of multimodal diagnostic data,demonstrating enhanced performance enabled by multimodal LLM collaboration,which opens new horizons for the clinical application of artificial intelligence-assisted technology.展开更多
Angiosperms experienced one of the most remarkable radiations of land plants and are now the dominant autotrophs in terrestrial ecosystems.Recent phylogenomic studies based on large-scale data from plastid,mitochondri...Angiosperms experienced one of the most remarkable radiations of land plants and are now the dominant autotrophs in terrestrial ecosystems.Recent phylogenomic studies based on large-scale data from plastid,mitochondrial,or nuclear transcriptomes/genomes and increased taxon sampling have provided unprecedent resolution into the phylogeny of flowering plants.However,owing to ancient rapid radiations,the interrelationships among the five lineages of Mesangiospermae,the vast majority of angiosperms,remain contentious.Here we show that,although plastid and mitochondrial genomes lack sufficient phylogenetic signal for resolving deeper phylogeny,the relationships among five mesangiosperm lineages can be confidently resolved under better-fitting models using genome-scale data.According to our Bayesian cross-validation and model test in a maximum likelihood framework,siteheterogeneous models(e.g.,CAT-GTR+G4,LG+C20+F+G)outperform site-homogeneous or partition models often used in previous studies.Under site-heterogeneous models,the approximately unbiased test favored our preferred tree recovered from various datasets:Ceratophyllales(coontails)are robustly recovered as sister to monocots,and they together are sister to the clade comprising magnoliids,Chloranthales,and eudicots.Our phylogenomic analyses resolve the last enigma of the deeper phylogeny of angiosperms and emphasize the efficacy of modeling compositional heterogeneity in resolving rapid radiations of plants.展开更多
The rapid advancement of artificial intelligence technology is driving transformative changes in medical diagnosis,treatment,and management systems through large-scale deep learning models-a process that brings both g...The rapid advancement of artificial intelligence technology is driving transformative changes in medical diagnosis,treatment,and management systems through large-scale deep learning models-a process that brings both groundbreaking opportunities and multifaceted challenges.This study focuses on the medical and healthcare applications of large-scale deep learning architectures,conducting a comprehensive survey to categorize and analyze their diverse uses.The survey results reveal that current applications of large models in healthcare encompass medical data management,healthcare services,medical devices,and preventive medicine,among others.Concurrently,large models demonstrate significant advantages in the medical domain,especially in high-precision diagnosis and prediction,data analysis and knowledge discovery,and enhancing operational efficiency.Nevertheless,we identify several challenges that need urgent attention,including improving the interpretability of large models,strengthening privacy protection,and addressing issues related to handling incomplete data.This research is dedicated to systematically elucidating the deep collaborative mechanisms between artificial intelligence and the healthcare field,providing theoretical references and practical guidance for both academia and industry.展开更多
Accurate retrieval of casting 3D models is crucial for process reuse.Current methods primarily focus on shape similarity,neglecting process design features,which compromises reusability.In this study,a novel deep lear...Accurate retrieval of casting 3D models is crucial for process reuse.Current methods primarily focus on shape similarity,neglecting process design features,which compromises reusability.In this study,a novel deep learning retrieval method for process reuse was proposed,which integrates process design features into the retrieval of casting 3D models.This method leverages the comparative language-image pretraining(CLIP)model to extract shape features from the three views and sectional views of the casting model and combines them with process design features such as modulus,main wall thickness,symmetry,and length-to-height ratio to enhance process reusability.A database of 230 production casting models was established for model validation.Results indicate that incorporating process design features improves model accuracy by 6.09%,reaching 97.82%,and increases process similarity by 30.25%.The reusability of the process was further verified using the casting simulation software EasyCast.The results show that the process retrieved after integrating process design features produces the least shrinkage in the target model,demonstrating this method’s superior ability for process reuse.This approach does not require a large dataset for training and optimization,making it highly applicable to casting process design and related manufacturing processes.展开更多
Porosity is an important attribute for evaluating the petrophysical properties of reservoirs, and has guiding significance for the exploration and development of oil and gas. The seismic inversion is a key method for ...Porosity is an important attribute for evaluating the petrophysical properties of reservoirs, and has guiding significance for the exploration and development of oil and gas. The seismic inversion is a key method for comprehensively obtaining the porosity. Deep learning methods provide an intelligent approach to suppress the ambiguity of the conventional inversion method. However, under the trace-bytrace inversion strategy, there is a lack of constraints from geological structural information, resulting in poor lateral continuity of prediction results. In addition, the heterogeneity and the sedimentary variability of subsurface media also lead to uncertainty in intelligent prediction. To achieve fine prediction of porosity, we consider the lateral continuity and variability and propose an improved structural modeling deep learning porosity prediction method. First, we combine well data, waveform attributes, and structural information as constraints to model geophysical parameters, constructing a high-quality training dataset with sedimentary facies-controlled significance. Subsequently, we introduce a gated axial attention mechanism to enhance the features of dataset and design a bidirectional closed-loop network system constrained by inversion and forward processes. The constraint coefficient is adaptively adjusted by the petrophysical information contained between the porosity and impedance in the study area. We demonstrate the effectiveness of the adaptive coefficient through numerical experiments.Finally, we compare the performance differences between the proposed method and conventional deep learning methods using data from two study areas. The proposed method achieves better consistency with the logging porosity, demonstrating the superiority of the proposed method.展开更多
The El Niño-Southern Oscillation(ENSO)is a naturally recurring interannual climate fluctuation that affects the global climate system.The advent of deep learning-based approaches has led to transformative changes...The El Niño-Southern Oscillation(ENSO)is a naturally recurring interannual climate fluctuation that affects the global climate system.The advent of deep learning-based approaches has led to transformative changes in ENSO forecasts,resulting in significant progress.Most deep learning-based ENSO prediction models which primarily rely solely on reanalysis data may lead to challenges in intensity underestimation in long-term forecasts,reducing the forecasting skills.To this end,we propose a deep residual-coupled model prediction(Res-CMP)model,which integrates historical reanalysis data and coupled model forecast data for multiyear ENSO prediction.The Res-CMP model is designed as a lightweight model that leverages only short-term reanalysis data and nudging assimilation prediction results of the Community Earth System Model(CESM)for effective prediction of the Niño 3.4 index.We also developed a transfer learning strategy for this model to overcome the limitations of inadequate forecast data.After determining the optimal configuration,which included selecting a suitable transfer learning rate during training,along with input variables and CESM forecast lengths,Res-CMP demonstrated a high correlation ability for 19-month lead time predictions(correlation coefficients exceeding 0.5).The Res-CMP model also alleviated the spring predictability barrier(SPB).When validated against actual ENSO events,Res-CMP successfully captured the temporal evolution of the Niño 3.4 index during La Niña events(1998/99 and 2020/21)and El Niño events(2009/10 and 2015/16).Our proposed model has the potential to further enhance ENSO prediction performance by using coupled models to assist deep learning methods.展开更多
The Sawayaerdun gold deposit is the only extremely large deposit occurring in the giant gold metallogenic belt of the Central Asian Tianshan Orogenic Belt.Breakthroughs in deep and peripheral exploration are of great ...The Sawayaerdun gold deposit is the only extremely large deposit occurring in the giant gold metallogenic belt of the Central Asian Tianshan Orogenic Belt.Breakthroughs in deep and peripheral exploration are of great significance for the sustainable development of the region.In this study,a comprehensive prospecting and exploration model for gold exploration in the high-cold and high-altitude areas of the southwestern Tianshan Mountains was developed by systematically evaluating the metallogenic geological conditions of the deposit and integrating geological,geochemical,and geophysical exploration multi-source data.Both the shallow metamorphic carbonaceous fine clastic rock and ductile shear zones controlled the Au-mineralized zone.Moreover,mineralized alteration,Au-Sb-As element chemical anomalies,high magnetic anomalies,and medium-high values from low-wave impedance were also important components of the model.Three-dimensional(3-D)geological attribute models of various ore-controlling factors were established,revealing the spatial distribution patterns of the deep structural frameworks and hidden mineralization.Geostatistical methods were used to delineate the three edges and two deep-mineral exploration areas.A comprehensive evaluation showed that anomalous variables were well confirmed in geological understanding,highlighting the effectiveness and accuracy of the predictions.These results are significant for deep-edge exploration.展开更多
Predicting cross-immunity between viral strains is vital for public health surveillance and vaccine development.Traditional neural network methods,such as BiLSTM,could be ineffective due to the lack of lab data for mo...Predicting cross-immunity between viral strains is vital for public health surveillance and vaccine development.Traditional neural network methods,such as BiLSTM,could be ineffective due to the lack of lab data for model training and the overshadowing of crucial features within sequence concatenation.The current work proposes a less data-consuming model incorporating a pre-trained gene sequence model and a mutual information inference operator.Our methodology utilizes gene alignment and deduplication algorithms to preprocess gene sequences,enhancing the model’s capacity to discern and focus on distinctions among input gene pairs.The model,i.e.,DNA Pretrained Cross-Immunity Protection Inference model(DPCIPI),outperforms state-of-theart(SOTA)models in predicting hemagglutination inhibition titer from influenza viral gene sequences only.Improvement in binary cross-immunity prediction is 1.58%in F1,2.34%in precision,1.57%in recall,and 1.57%in Accuracy.For multilevel cross-immunity improvements,the improvement is 2.12%in F1,3.50%in precision,2.19%in recall,and 2.19%in Accuracy.Our study showcases the potential of pre-trained gene models to improve predictions of antigenic variation and cross-immunity.With expanding gene data and advancements in pre-trained models,this approach promises significant impacts on vaccine development and public health.展开更多
DeepSeek,a Chinese artificial intelligence(AI)startup,has released their V3 and R1 series models,which attracted global attention due to their low cost,high performance,and open-source advantages.This paper begins by ...DeepSeek,a Chinese artificial intelligence(AI)startup,has released their V3 and R1 series models,which attracted global attention due to their low cost,high performance,and open-source advantages.This paper begins by reviewing the evolution of large AI models focusing on paradigm shifts,the mainstream large language model(LLM)paradigm,and the DeepSeek paradigm.Subsequently,the paper highlights novel algorithms introduced by DeepSeek,including multi-head latent attention(MLA),mixture-of-experts(MoE),multi-token prediction(MTP),and group relative policy optimization(GRPO).The paper then explores DeepSeek's engineering breakthroughs in LLM scaling,training,inference,and system-level optimization architecture.Moreover,the impact of DeepSeek models on the competitive AI landscape is analyzed,comparing them to mainstream LLMs across various fields.Finally,the paper reflects on the insights gained from DeepSeek's innovations and discusses future trends in the technical and engineering development of large AI models,particularly in data,training,and reasoning.展开更多
Deep neural networks(DNNs)have found extensive applications in safety-critical artificial intelligence systems,such as autonomous driving and facial recognition systems.However,recent research has revealed their susce...Deep neural networks(DNNs)have found extensive applications in safety-critical artificial intelligence systems,such as autonomous driving and facial recognition systems.However,recent research has revealed their susceptibility to backdoors maliciously injected by adversaries.This vulnerability arises due to the intricate architecture and opacity of DNNs,resulting in numerous redundant neurons embedded within the models.Adversaries exploit these vulnerabilities to conceal malicious backdoor information within DNNs,thereby causing erroneous outputs and posing substantial threats to the efficacy of DNN-based applications.This article presents a comprehensive survey of backdoor attacks against DNNs and the countermeasure methods employed to mitigate them.Initially,we trace the evolution of the concept from traditional backdoor attacks to backdoor attacks against DNNs,highlighting the feasibility and practicality of generating backdoor attacks against DNNs.Subsequently,we provide an overview of notable works encompassing various attack and defense strategies,facilitating a comparative analysis of their approaches.Through these discussions,we offer constructive insights aimed at refining these techniques.Finally,we extend our research perspective to the domain of large language models(LLMs)and synthesize the characteristics and developmental trends of backdoor attacks and defense methods targeting LLMs.Through a systematic review of existing studies on backdoor vulnerabilities in LLMs,we identify critical open challenges in this field and propose actionable directions for future research.展开更多
The Yishu fault zone in Shandong Province,China,exhibits favorable conditions for medium-and high-temperature geothermal storage.However,its geothermal occurrence patterns and heat storage model remain unclear.Recogni...The Yishu fault zone in Shandong Province,China,exhibits favorable conditions for medium-and high-temperature geothermal storage.However,its geothermal occurrence patterns and heat storage model remain unclear.Recognizing the importance of deep heat sources and heat channels in unraveling the formation mechanisms of medium-and high-temperature geothermal systems,we conducted a magnetotelluric survey along the Yishu fault zone,which contained 10 survey lines and 119 usable points.Using two-dimensional and three-dimensional nonlinear conjugate gradient inversion,the deep electrical structure was obtained,and the three-dimensional geothermal reservoir model in the study area was constructed for the first time.The results show that the deep low-resistance anomaly in the Yishu fault zone moves upward through the channel with slightly higher resistivity,forming a relatively low-resistance layer and a low-resistance layer in the shallow part,corresponding to the heat source,thermal conduction fault,heat reservoir,and overburden layer,respectively.The fault structure primarily controls regional geothermal anomalies,influencing atmospheric precipitation,surface water infiltration,and geothermal water migration pathways.The results of this study have a certain guiding significance for the study of the formation mechanism and distribution law of geothermal resources in the middle and deep strata of the Yishu fault zone.Additionally,the results provide valuable insights for the exploration and assessment of geothermal resources in the area.展开更多
Objective:To systematically evaluate prediction models for postoperative deep vein thrombosis(DVT)in elderly hip fracture patients and assess their methodological quality and predictive performance.Methods:Following P...Objective:To systematically evaluate prediction models for postoperative deep vein thrombosis(DVT)in elderly hip fracture patients and assess their methodological quality and predictive performance.Methods:Following PRISMA guidelines,we searched eight databases(PubMed,Embase,Cochrane Library,Web of Science,CINAHL,CNKI,Wanfang,VIP)from inception to May 2025.Studies developing or validating DVT prediction models in elderly hip fracture patients were included.Two reviewers independently screened studies,extracted data,and assessed risk of bias and applicability using the PROBAST tool.Results:Eleven studies were included,all conducted in China between 2021 and 2025.Sample sizes ranged from 101 to 504 patients(total n=3,286).Models incorporated 3 to 9 predictors,with D-dimer,age,and time from injury to surgery being most common.All 11 studies(100%)were rated as high risk of bias,primarily due to small sample sizes,lack of validation,and inadequate missing data handling.Applicability concerns were low in 8 studies(72.7%).AUC values ranged from 0.648 to 0.967,with 10 studies(90.9%)reporting AUC>0.7.Meta-analysis identified time from injury to surgery(OR=4.63,95%CI:2.58–6.68),age(OR=1.99),D-dimer(OR=1.51),and Caprini score(OR=1.75)as significant predictors.Conclusion:Current DVT prediction models for elderly hip fracture patients demonstrate acceptable discrimination but are limited by high risk of bias and lack of external validation.Prospective,multicenter studies with rigorous validation are needed to develop clinically applicable models.展开更多
Artificial intelligence(AI)is rapidly transforming the landscape of hepatology by enabling automated data interpretation,early disease detection,and individualized treatment strategies.Chronic liver diseases,including...Artificial intelligence(AI)is rapidly transforming the landscape of hepatology by enabling automated data interpretation,early disease detection,and individualized treatment strategies.Chronic liver diseases,including non-alcoholic fatty liver disease,cirrhosis,and hepatocellular carcinoma,often progress silently and pose diagnostic challenges due to reliance on invasive biopsies and operatordependent imaging.This review explores the integration of AI across key domains such as big data analytics,deep learning-based image analysis,histopathological interpretation,biomarker discovery,and clinical prediction modeling.AI algorithms have demonstrated high accuracy in liver fibrosis staging,hepatocellular carcinoma detection,and non-alcoholic fatty liver disease risk stratification,while also enhancing survival prediction and treatment response assessment.For instance,convolutional neural networks trained on portal venous-phase computed tomography have achieved area under the curves up to 0.92 for significant fibrosis(F2-F4)and 0.89 for advanced fibrosis,with magnetic resonance imaging-based models reporting comparable performance.Advanced methodologies such as federated learning preserve patient privacy during cross-center model training,and explainable AI techniques promote transparency and clinician trust.Despite these advancements,clinical adoption remains limited by challenges including data heterogeneity,algorithmic bias,regulatory uncertainty,and lack of real-time integration into electronic health records.Looking forward,the convergence of multi-omics,imaging,and clinical data through interpretable and validated AI frameworks holds great promise for precision liver care.Continued efforts in model standardization,ethical oversight,and clinician-centered deployment will be essential to realize the full potential of AI in hepatopathy diagnosis and treatment.展开更多
基金supported by the National Key Research and Development Program of China[grant number 2022YFE0106800]an Innovation Group Project of the Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai)[grant number 311024001]+3 种基金a project supported by the Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai)[grant number SML2023SP209]a Research Council of Norway funded project(MAPARC)[grant number 328943]a Nansen Center´s basic institutional funding[grant number 342624]the high-performance computing support from the School of Atmospheric Science at Sun Yat-sen University。
文摘Current shipping,tourism,and resource development requirements call for more accurate predictions of the Arctic sea-ice concentration(SIC).However,due to the complex physical processes involved,predicting the spatiotemporal distribution of Arctic SIC is more challenging than predicting its total extent.In this study,spatiotemporal prediction models for monthly Arctic SIC at 1-to 3-month leads are developed based on U-Net-an effective convolutional deep-learning approach.Based on explicit Arctic sea-ice-atmosphere interactions,11 variables associated with Arctic sea-ice variations are selected as predictors,including observed Arctic SIC,atmospheric,oceanic,and heat flux variables at 1-to 3-month leads.The prediction skills for the monthly Arctic SIC of the test set(from January 2018 to December 2022)are evaluated by examining the mean absolute error(MAE)and binary accuracy(BA).Results showed that the U-Net model had lower MAE and higher BA for Arctic SIC compared to two dynamic climate prediction systems(CFSv2 and NorCPM).By analyzing the relative importance of each predictor,the prediction accuracy relies more on the SIC at the 1-month lead,but on the surface net solar radiation flux at 2-to 3-month leads.However,dynamic models show limited prediction skills for surface net solar radiation flux and other physical processes,especially in autumn.Therefore,the U-Net model can be used to capture the connections among these key physical processes associated with Arctic sea ice and thus offers a significant advantage in predicting Arctic SIC.
基金supported by the Project of Stable Support for Youth Team in Basic Research Field,CAS(grant No.YSBR-018)the National Natural Science Foundation of China(grant Nos.42188101,42130204)+4 种基金the B-type Strategic Priority Program of CAS(grant no.XDB41000000)the National Natural Science Foundation of China(NSFC)Distinguished Overseas Young Talents Program,Innovation Program for Quantum Science and Technology(2021ZD0300301)the Open Research Project of Large Research Infrastructures of CAS-“Study on the interaction between low/mid-latitude atmosphere and ionosphere based on the Chinese Meridian Project”.The project was supported also by the National Key Laboratory of Deep Space Exploration(Grant No.NKLDSE2023A002)the Open Fund of Anhui Provincial Key Laboratory of Intelligent Underground Detection(Grant No.APKLIUD23KF01)the China National Space Administration(CNSA)pre-research Project on Civil Aerospace Technologies No.D010305,D010301.
文摘Sporadic E(Es)layers in the ionosphere are characterized by intense plasma irregularities in the E region at altitudes of 90-130 km.Because they can significantly influence radio communications and navigation systems,accurate forecasting of Es layers is crucial for ensuring the precision and dependability of navigation satellite systems.In this study,we present Es predictions made by an empirical model and by a deep learning model,and analyze their differences comprehensively by comparing the model predictions to satellite RO measurements and ground-based ionosonde observations.The deep learning model exhibited significantly better performance,as indicated by its high coefficient of correlation(r=0.87)with RO observations and predictions,than did the empirical model(r=0.53).This study highlights the importance of integrating artificial intelligence technology into ionosphere modelling generally,and into predicting Es layer occurrences and characteristics,in particular.
基金Project supported by the National Natural Science Foundation of China(No.42202314)。
文摘A coupled thermal-hydro-mechanical cohesive phase-field model for hydraulic fracturing in deep coal seams is presented.Heat exchange between the cold fluid and the hot rock is considered,and the thermal contribution terms between the cold fluid and the hot rock are derived.Heat transfer obeys Fourier's law,and porosity is used to relate the thermodynamic parameters of the fracture and matrix domains.The net pressure difference between the fracture and the matrix is neglected,and thus the fluid flow is modeled by the unified fluid-governing equations.The evolution equations of porosity and Biot's coefficient during hydraulic fracturing are derived from their definitions.The effect of coal cleats is considered and modeled by Voronoi polygons,and this approach is shown to have high accuracy.The accuracy of the proposed model is verified by two sets of fracturing experiments in multilayer coal seams.Subsequently,the differences in fracture morphology,fluid pressure response,and fluid pressure distribution between direct fracturing of coal seams and indirect fracturing of shale interlayers are explored,and the effects of the cluster number and cluster spacing on fracture morphology for multi-cluster fracturing are also examined.The numerical results show that the proposed model is expected to be a powerful tool for the fracturing design and optimization of deep coalbed methane.
基金Projects(2024YFC3013801,2022YFC3004602)supported by the National Key R&D Program of ChinaProjects(U23B2093,52034009)supported by the National Natural Science Foundation of China。
文摘Rock residual strength,as an important input parameter,plays an indispensable role in proposing the reasonable and scientific scheme about stope design,underground tunnel excavation and stability evaluation of deep chambers.Therefore,previous residual strength models of rocks established were reviewed.And corresponding related problems were stated.Subsequently,starting from the effects of bedding and whole life-cycle evolution process,series of triaxial mechanical tests of deep bedded sandstone with five bedding angles were conducted under different confining pressures.Then,six residual strength models considering the effects of bedding and whole life-cycle evolution process were established and evaluated.Finally,a cohesion loss model for determining residual strength of deep bedded sandstone was verified.The results showed that the effects of bedding and whole life-cycle evolution process had both significant influences on the evolution characteristic of residual strength of deep bedded sandstone.Additionally,residual strength parameters:residual cohesion and residual internal friction angle of deep bedded sandstone were not constant,which both significantly changed with increasing bedding angle.Besides,the cohesion loss model was the most suitable for determining and estimating the residual strength of bedded rocks,which could provide more accurate theoretical guidance for the stability control of deep chambers.
基金Supported by the National Natural Science Foundation of China(42302141).
文摘Taking the second member of the Xujiahe Formation of the Upper Triassic in the Xinchang structural belt as an example,based on data such as logging,production,seismic interpretation and test,a systematic analysis was conducted on the structural characteristics and evolution,reservoir diagenesis and densification processes,and types and stages of faults/fractures,and revealing the multi-stage and multi-factor dynamic coupled enrichment mechanisms of tight gas reservoirs.(1)In the early Yanshan period,the paleo-structural traps were formed with low-medium maturity hydrocarbons accumulating in structural highs driven by buoyancy since reservoirs were not fully densified in this stage,demonstrating paleo-structure control on traps and early hydrocarbon accumulation.(2)In the middle-late Yanshan period,the source rocks became mature to generate and expel a large quantity of hydrocarbons.Grain size and type of sandstone controlled the time of reservoir densification,which restricted the scale of hydrocarbon charging,allowing for only a small-scale migration through sand bodies near the fault/fracture or less-densified matrix reservoirs.(3)During the Himalayan period,the source rocks reached overmaturity,and the residual oil cracking gas was efficiently transported along the late-stage faults/fractures.Wells with high production capacity were mainly located in Type I and II fault/fracture zones comprising the late-stage north-south trending fourth-order faults and the late-stage fractures.The productivity of the wells was controlled by the transformation of the late-stage faults/fractures.(4)The Xinchang structural belt underwent three stages of tectonic evolution,two stages of reservoir formation,and three stages of fault/fractures development.Hydrocarbons mainly accumulated in the paleo-structure highs.After reservoir densification and late fault/fracture adjustment,a complex gas-water distribution pattern was formed.Thus,it is summarized as the model of“near-source and low-abundance hydrocarbon charging in the early stage,and differential enrichment of natural gas under the joint control of fault-fold-fracture complex,high-quality reservoirs and structural highs in the late stage”.Faults/fractures with well-coupled fault-fold-fracture-pore are favorable exploration targets with high exploration effectiveness.
基金Supported by China Health Promotion Foundation Young Doctors’Research Foundation for Inflammatory Bowel DiseaseTaishan Scholars Program of Shandong Province,China,NO.tsqn202306343National Natural Science Foundation of China,No.82270580,No.82070552,No.82270578,and No.82300599.
文摘BACKGROUND Gastrointestinal diseases have complex etiologies and clinical presentations.An accurate diagnosis requires physicians to integrate diverse information,including medical history,laboratory test results,and imaging findings.Existing artificial intelligence-assisted diagnostic tools are limited to single-modality information,resulting in recommendations that are often incomplete and may be associated with clinical or legal risks.AIM To develop and evaluate a collaborative multimodal large language model(LLM)framework for clinical decision-making in digestive diseases.METHODS In this observational study,DeepGut,a multimodal LLM collaborative diagnostic framework,was developed to integrate four distinct large models into a four-tiered structure.The framework sequentially accomplishes multimodal infor-mation extraction,logical“chain”construction,diagnostic and treatment suggestion generation,and risk analysis.The model was evaluated using objective metrics,which assess the reliability and comprehensiveness of model-generated results,and subjective expert opinions,which examine the effectiveness of the framework in assisting physicians.RESULTS The diagnostic and treatment recommendations generated by the DeepGut framework achieved exceptional performance,with a diagnostic accuracy of 97.8%,diagnostic completeness of 93.9%,treatment plan accuracy of 95.2%,and treatment plan completeness of 98.0%,significantly surpassing the capabilities of single-modal LLM-based diagnostic tools.Experts evaluating the framework commended the completeness,relevance,and logical coherence of its outputs.However,the collaborative multimodal LLM approach resulted in increased input and output token counts,leading to higher computational costs and extended diagnostic times.CONCLUSION The framework achieves successful integration of multimodal diagnostic data,demonstrating enhanced performance enabled by multimodal LLM collaboration,which opens new horizons for the clinical application of artificial intelligence-assisted technology.
基金supported by the National Natural Science Foundation of China(42222201,42288201)。
文摘Angiosperms experienced one of the most remarkable radiations of land plants and are now the dominant autotrophs in terrestrial ecosystems.Recent phylogenomic studies based on large-scale data from plastid,mitochondrial,or nuclear transcriptomes/genomes and increased taxon sampling have provided unprecedent resolution into the phylogeny of flowering plants.However,owing to ancient rapid radiations,the interrelationships among the five lineages of Mesangiospermae,the vast majority of angiosperms,remain contentious.Here we show that,although plastid and mitochondrial genomes lack sufficient phylogenetic signal for resolving deeper phylogeny,the relationships among five mesangiosperm lineages can be confidently resolved under better-fitting models using genome-scale data.According to our Bayesian cross-validation and model test in a maximum likelihood framework,siteheterogeneous models(e.g.,CAT-GTR+G4,LG+C20+F+G)outperform site-homogeneous or partition models often used in previous studies.Under site-heterogeneous models,the approximately unbiased test favored our preferred tree recovered from various datasets:Ceratophyllales(coontails)are robustly recovered as sister to monocots,and they together are sister to the clade comprising magnoliids,Chloranthales,and eudicots.Our phylogenomic analyses resolve the last enigma of the deeper phylogeny of angiosperms and emphasize the efficacy of modeling compositional heterogeneity in resolving rapid radiations of plants.
基金funded by the National Natural Science Foundation of China(Grant No.62272236)the Natural Science Foundation of Jiangsu Province(Grant No.BK20201136).
文摘The rapid advancement of artificial intelligence technology is driving transformative changes in medical diagnosis,treatment,and management systems through large-scale deep learning models-a process that brings both groundbreaking opportunities and multifaceted challenges.This study focuses on the medical and healthcare applications of large-scale deep learning architectures,conducting a comprehensive survey to categorize and analyze their diverse uses.The survey results reveal that current applications of large models in healthcare encompass medical data management,healthcare services,medical devices,and preventive medicine,among others.Concurrently,large models demonstrate significant advantages in the medical domain,especially in high-precision diagnosis and prediction,data analysis and knowledge discovery,and enhancing operational efficiency.Nevertheless,we identify several challenges that need urgent attention,including improving the interpretability of large models,strengthening privacy protection,and addressing issues related to handling incomplete data.This research is dedicated to systematically elucidating the deep collaborative mechanisms between artificial intelligence and the healthcare field,providing theoretical references and practical guidance for both academia and industry.
基金supported by the National Natural Science Foundation of China(Nos.52074246,52275390,52375394)the National Defense Basic Scientific Research Program of China(No.JCKY2020408B002)the Key R&D Program of Shanxi Province(No.202102050201011).
文摘Accurate retrieval of casting 3D models is crucial for process reuse.Current methods primarily focus on shape similarity,neglecting process design features,which compromises reusability.In this study,a novel deep learning retrieval method for process reuse was proposed,which integrates process design features into the retrieval of casting 3D models.This method leverages the comparative language-image pretraining(CLIP)model to extract shape features from the three views and sectional views of the casting model and combines them with process design features such as modulus,main wall thickness,symmetry,and length-to-height ratio to enhance process reusability.A database of 230 production casting models was established for model validation.Results indicate that incorporating process design features improves model accuracy by 6.09%,reaching 97.82%,and increases process similarity by 30.25%.The reusability of the process was further verified using the casting simulation software EasyCast.The results show that the process retrieved after integrating process design features produces the least shrinkage in the target model,demonstrating this method’s superior ability for process reuse.This approach does not require a large dataset for training and optimization,making it highly applicable to casting process design and related manufacturing processes.
基金the support of Research Program of Fine Exploration and Surrounding Rock Classification Technology for Deep Buried Long Tunnels Driven by Horizontal Directional Drilling and Magnetotelluric Methods Based on Deep Learning under Grant E202408010the Sichuan Science and Technology Program under Grant 2024NSFSC1984 and Grant 2024NSFSC1990。
文摘Porosity is an important attribute for evaluating the petrophysical properties of reservoirs, and has guiding significance for the exploration and development of oil and gas. The seismic inversion is a key method for comprehensively obtaining the porosity. Deep learning methods provide an intelligent approach to suppress the ambiguity of the conventional inversion method. However, under the trace-bytrace inversion strategy, there is a lack of constraints from geological structural information, resulting in poor lateral continuity of prediction results. In addition, the heterogeneity and the sedimentary variability of subsurface media also lead to uncertainty in intelligent prediction. To achieve fine prediction of porosity, we consider the lateral continuity and variability and propose an improved structural modeling deep learning porosity prediction method. First, we combine well data, waveform attributes, and structural information as constraints to model geophysical parameters, constructing a high-quality training dataset with sedimentary facies-controlled significance. Subsequently, we introduce a gated axial attention mechanism to enhance the features of dataset and design a bidirectional closed-loop network system constrained by inversion and forward processes. The constraint coefficient is adaptively adjusted by the petrophysical information contained between the porosity and impedance in the study area. We demonstrate the effectiveness of the adaptive coefficient through numerical experiments.Finally, we compare the performance differences between the proposed method and conventional deep learning methods using data from two study areas. The proposed method achieves better consistency with the logging porosity, demonstrating the superiority of the proposed method.
基金The National Key Research and Development Program of China under contract Nos 2024YFF0808900,2023YFF0805300,and 2020YFA0608804the Civilian Space Programme of China under contract No.D040305.
文摘The El Niño-Southern Oscillation(ENSO)is a naturally recurring interannual climate fluctuation that affects the global climate system.The advent of deep learning-based approaches has led to transformative changes in ENSO forecasts,resulting in significant progress.Most deep learning-based ENSO prediction models which primarily rely solely on reanalysis data may lead to challenges in intensity underestimation in long-term forecasts,reducing the forecasting skills.To this end,we propose a deep residual-coupled model prediction(Res-CMP)model,which integrates historical reanalysis data and coupled model forecast data for multiyear ENSO prediction.The Res-CMP model is designed as a lightweight model that leverages only short-term reanalysis data and nudging assimilation prediction results of the Community Earth System Model(CESM)for effective prediction of the Niño 3.4 index.We also developed a transfer learning strategy for this model to overcome the limitations of inadequate forecast data.After determining the optimal configuration,which included selecting a suitable transfer learning rate during training,along with input variables and CESM forecast lengths,Res-CMP demonstrated a high correlation ability for 19-month lead time predictions(correlation coefficients exceeding 0.5).The Res-CMP model also alleviated the spring predictability barrier(SPB).When validated against actual ENSO events,Res-CMP successfully captured the temporal evolution of the Niño 3.4 index during La Niña events(1998/99 and 2020/21)and El Niño events(2009/10 and 2015/16).Our proposed model has the potential to further enhance ENSO prediction performance by using coupled models to assist deep learning methods.
基金supported by the Science&Technology Fundamental Resources Investigation Program(2022FY101705)China geological survey project(DD20240204805)market prospecting exploration project(CG-2024-HX-S012)。
文摘The Sawayaerdun gold deposit is the only extremely large deposit occurring in the giant gold metallogenic belt of the Central Asian Tianshan Orogenic Belt.Breakthroughs in deep and peripheral exploration are of great significance for the sustainable development of the region.In this study,a comprehensive prospecting and exploration model for gold exploration in the high-cold and high-altitude areas of the southwestern Tianshan Mountains was developed by systematically evaluating the metallogenic geological conditions of the deposit and integrating geological,geochemical,and geophysical exploration multi-source data.Both the shallow metamorphic carbonaceous fine clastic rock and ductile shear zones controlled the Au-mineralized zone.Moreover,mineralized alteration,Au-Sb-As element chemical anomalies,high magnetic anomalies,and medium-high values from low-wave impedance were also important components of the model.Three-dimensional(3-D)geological attribute models of various ore-controlling factors were established,revealing the spatial distribution patterns of the deep structural frameworks and hidden mineralization.Geostatistical methods were used to delineate the three edges and two deep-mineral exploration areas.A comprehensive evaluation showed that anomalous variables were well confirmed in geological understanding,highlighting the effectiveness and accuracy of the predictions.These results are significant for deep-edge exploration.
基金supported by the Bill & Melinda Gates Foundation and the Minderoo Foundation
文摘Predicting cross-immunity between viral strains is vital for public health surveillance and vaccine development.Traditional neural network methods,such as BiLSTM,could be ineffective due to the lack of lab data for model training and the overshadowing of crucial features within sequence concatenation.The current work proposes a less data-consuming model incorporating a pre-trained gene sequence model and a mutual information inference operator.Our methodology utilizes gene alignment and deduplication algorithms to preprocess gene sequences,enhancing the model’s capacity to discern and focus on distinctions among input gene pairs.The model,i.e.,DNA Pretrained Cross-Immunity Protection Inference model(DPCIPI),outperforms state-of-theart(SOTA)models in predicting hemagglutination inhibition titer from influenza viral gene sequences only.Improvement in binary cross-immunity prediction is 1.58%in F1,2.34%in precision,1.57%in recall,and 1.57%in Accuracy.For multilevel cross-immunity improvements,the improvement is 2.12%in F1,3.50%in precision,2.19%in recall,and 2.19%in Accuracy.Our study showcases the potential of pre-trained gene models to improve predictions of antigenic variation and cross-immunity.With expanding gene data and advancements in pre-trained models,this approach promises significant impacts on vaccine development and public health.
基金supported by the National Natural Science Foundation of China(62233005,62293502,U2441245,62176185,U23B2057,62306112)the STCSM Science and Technology Innovation Action Plan Computational Biology Program(24JS2830400)+2 种基金the State Key Laboratory of Industrial Control Technology,China(ICT2024A22)the Shanghai Sailing Program(23YF1409400)the National Science and Technology Major Project(2024ZD0532403).
文摘DeepSeek,a Chinese artificial intelligence(AI)startup,has released their V3 and R1 series models,which attracted global attention due to their low cost,high performance,and open-source advantages.This paper begins by reviewing the evolution of large AI models focusing on paradigm shifts,the mainstream large language model(LLM)paradigm,and the DeepSeek paradigm.Subsequently,the paper highlights novel algorithms introduced by DeepSeek,including multi-head latent attention(MLA),mixture-of-experts(MoE),multi-token prediction(MTP),and group relative policy optimization(GRPO).The paper then explores DeepSeek's engineering breakthroughs in LLM scaling,training,inference,and system-level optimization architecture.Moreover,the impact of DeepSeek models on the competitive AI landscape is analyzed,comparing them to mainstream LLMs across various fields.Finally,the paper reflects on the insights gained from DeepSeek's innovations and discusses future trends in the technical and engineering development of large AI models,particularly in data,training,and reasoning.
基金supported in part by the National Natural Science Foundation of China under Grants No.62372087 and No.62072076the Research Fund of State Key Laboratory of Processors under Grant No.CLQ202310the CSC scholarship.
文摘Deep neural networks(DNNs)have found extensive applications in safety-critical artificial intelligence systems,such as autonomous driving and facial recognition systems.However,recent research has revealed their susceptibility to backdoors maliciously injected by adversaries.This vulnerability arises due to the intricate architecture and opacity of DNNs,resulting in numerous redundant neurons embedded within the models.Adversaries exploit these vulnerabilities to conceal malicious backdoor information within DNNs,thereby causing erroneous outputs and posing substantial threats to the efficacy of DNN-based applications.This article presents a comprehensive survey of backdoor attacks against DNNs and the countermeasure methods employed to mitigate them.Initially,we trace the evolution of the concept from traditional backdoor attacks to backdoor attacks against DNNs,highlighting the feasibility and practicality of generating backdoor attacks against DNNs.Subsequently,we provide an overview of notable works encompassing various attack and defense strategies,facilitating a comparative analysis of their approaches.Through these discussions,we offer constructive insights aimed at refining these techniques.Finally,we extend our research perspective to the domain of large language models(LLMs)and synthesize the characteristics and developmental trends of backdoor attacks and defense methods targeting LLMs.Through a systematic review of existing studies on backdoor vulnerabilities in LLMs,we identify critical open challenges in this field and propose actionable directions for future research.
文摘The Yishu fault zone in Shandong Province,China,exhibits favorable conditions for medium-and high-temperature geothermal storage.However,its geothermal occurrence patterns and heat storage model remain unclear.Recognizing the importance of deep heat sources and heat channels in unraveling the formation mechanisms of medium-and high-temperature geothermal systems,we conducted a magnetotelluric survey along the Yishu fault zone,which contained 10 survey lines and 119 usable points.Using two-dimensional and three-dimensional nonlinear conjugate gradient inversion,the deep electrical structure was obtained,and the three-dimensional geothermal reservoir model in the study area was constructed for the first time.The results show that the deep low-resistance anomaly in the Yishu fault zone moves upward through the channel with slightly higher resistivity,forming a relatively low-resistance layer and a low-resistance layer in the shallow part,corresponding to the heat source,thermal conduction fault,heat reservoir,and overburden layer,respectively.The fault structure primarily controls regional geothermal anomalies,influencing atmospheric precipitation,surface water infiltration,and geothermal water migration pathways.The results of this study have a certain guiding significance for the study of the formation mechanism and distribution law of geothermal resources in the middle and deep strata of the Yishu fault zone.Additionally,the results provide valuable insights for the exploration and assessment of geothermal resources in the area.
文摘Objective:To systematically evaluate prediction models for postoperative deep vein thrombosis(DVT)in elderly hip fracture patients and assess their methodological quality and predictive performance.Methods:Following PRISMA guidelines,we searched eight databases(PubMed,Embase,Cochrane Library,Web of Science,CINAHL,CNKI,Wanfang,VIP)from inception to May 2025.Studies developing or validating DVT prediction models in elderly hip fracture patients were included.Two reviewers independently screened studies,extracted data,and assessed risk of bias and applicability using the PROBAST tool.Results:Eleven studies were included,all conducted in China between 2021 and 2025.Sample sizes ranged from 101 to 504 patients(total n=3,286).Models incorporated 3 to 9 predictors,with D-dimer,age,and time from injury to surgery being most common.All 11 studies(100%)were rated as high risk of bias,primarily due to small sample sizes,lack of validation,and inadequate missing data handling.Applicability concerns were low in 8 studies(72.7%).AUC values ranged from 0.648 to 0.967,with 10 studies(90.9%)reporting AUC>0.7.Meta-analysis identified time from injury to surgery(OR=4.63,95%CI:2.58–6.68),age(OR=1.99),D-dimer(OR=1.51),and Caprini score(OR=1.75)as significant predictors.Conclusion:Current DVT prediction models for elderly hip fracture patients demonstrate acceptable discrimination but are limited by high risk of bias and lack of external validation.Prospective,multicenter studies with rigorous validation are needed to develop clinically applicable models.
基金Supported by the Science Planning Project of Liaoning Province,No.2019JH2/10300031-05the National Natural Science Foundation of China,No.12171074.
文摘Artificial intelligence(AI)is rapidly transforming the landscape of hepatology by enabling automated data interpretation,early disease detection,and individualized treatment strategies.Chronic liver diseases,including non-alcoholic fatty liver disease,cirrhosis,and hepatocellular carcinoma,often progress silently and pose diagnostic challenges due to reliance on invasive biopsies and operatordependent imaging.This review explores the integration of AI across key domains such as big data analytics,deep learning-based image analysis,histopathological interpretation,biomarker discovery,and clinical prediction modeling.AI algorithms have demonstrated high accuracy in liver fibrosis staging,hepatocellular carcinoma detection,and non-alcoholic fatty liver disease risk stratification,while also enhancing survival prediction and treatment response assessment.For instance,convolutional neural networks trained on portal venous-phase computed tomography have achieved area under the curves up to 0.92 for significant fibrosis(F2-F4)and 0.89 for advanced fibrosis,with magnetic resonance imaging-based models reporting comparable performance.Advanced methodologies such as federated learning preserve patient privacy during cross-center model training,and explainable AI techniques promote transparency and clinician trust.Despite these advancements,clinical adoption remains limited by challenges including data heterogeneity,algorithmic bias,regulatory uncertainty,and lack of real-time integration into electronic health records.Looking forward,the convergence of multi-omics,imaging,and clinical data through interpretable and validated AI frameworks holds great promise for precision liver care.Continued efforts in model standardization,ethical oversight,and clinician-centered deployment will be essential to realize the full potential of AI in hepatopathy diagnosis and treatment.
