Predicting molecular properties is essential for advancing for advancing drug discovery and design. Recently, Graph Neural Networks (GNNs) have gained prominence due to their ability to capture the complex structural ...Predicting molecular properties is essential for advancing for advancing drug discovery and design. Recently, Graph Neural Networks (GNNs) have gained prominence due to their ability to capture the complex structural and relational information inherent in molecular graphs. Despite their effectiveness, the “black-box” nature of GNNs remains a significant obstacle to their widespread adoption in chemistry, as it hinders interpretability and trust. In this context, several explanation methods based on factual reasoning have emerged. These methods aim to interpret the predictions made by GNNs by analyzing the key features contributing to the prediction. However, these approaches fail to answer critical questions: “How to ensure that the structure-property mapping learned by GNNs is consistent with established domain knowledge”. In this paper, we propose MMGCF, a novel counterfactual explanation framework designed specifically for the prediction of GNN-based molecular properties. MMGCF constructs a hierarchical tree structure on molecular motifs, enabling the systematic generation of counterfactuals through motif perturbations. This framework identifies causally significant motifs and elucidates their impact on model predictions, offering insights into the relationship between structural modifications and predicted properties. Our method demonstrates its effectiveness through comprehensive quantitative and qualitative evaluations of four real-world molecular datasets.展开更多
Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the sett...Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the settlement caused by tunneling.However,well-performing ML models are usually less interpretable.Irrelevant input features decrease the performance and interpretability of an ML model.Nonetheless,feature selection,a critical step in the ML pipeline,is usually ignored in most studies that focused on predicting tunneling-induced settlement.This study applies four techniques,i.e.Pearson correlation method,sequential forward selection(SFS),sequential backward selection(SBS)and Boruta algorithm,to investigate the effect of feature selection on the model’s performance when predicting the tunneling-induced maximum surface settlement(S_(max)).The data set used in this study was compiled from two metro tunnel projects excavated in Hangzhou,China using earth pressure balance(EPB)shields and consists of 14 input features and a single output(i.e.S_(max)).The ML model that is trained on features selected from the Boruta algorithm demonstrates the best performance in both the training and testing phases.The relevant features chosen from the Boruta algorithm further indicate that tunneling-induced settlement is affected by parameters related to tunnel geometry,geological conditions and shield operation.The recently proposed Shapley additive explanations(SHAP)method explores how the input features contribute to the output of a complex ML model.It is observed that the larger settlements are induced during shield tunneling in silty clay.Moreover,the SHAP analysis reveals that the low magnitudes of face pressure at the top of the shield increase the model’s output。展开更多
This paper takes a microanalytic perspective on the speech and gestures used by one teacher of ESL （English as a Second Language） in an intensive English program classroom. Videotaped excerpts from her intermediate-...This paper takes a microanalytic perspective on the speech and gestures used by one teacher of ESL （English as a Second Language） in an intensive English program classroom. Videotaped excerpts from her intermediate-level grammar course were transcribed to represent the speech, gesture and other non-verbal behavior that accompanied unplanned explanations of vocabulary that arose during three focus-on-form lessons. The gesture classification system of McNeill （1992）, which delineates different types of hand movements （iconics metaphorics, deictics, beats）, was used to understand the role the gestures played in these explanations. Results suggest that gestures and other non-verbal behavior are forms of input to classroom second language learners that must be considered a salient factor in classroom-based SLA （Second Language Acquisition） research展开更多
In this paper I examine the following claims by William Eaton in his monograph Boyle on Fire： （i） that Boyle＇s religious convictions led him to believe that the world was not completely explicable, and this shows ...In this paper I examine the following claims by William Eaton in his monograph Boyle on Fire： （i） that Boyle＇s religious convictions led him to believe that the world was not completely explicable, and this shows that there is a shortcoming in the power of mechanical explanations; （ii） that mechanical explanations offer only sufficient, not necessary explanations, and this too was taken by Boyle to be a limit in the explanatory power of mechanical explanations; （iii） that the mature Boyle thought that there could be more intelligible explanatory models than mechanism; and （iv） that what Boyle says at any point in his career is incompatible with the statement of Maria Boas-Hall, i.e., that the mechanical hypothesis can explicate all natural phenomena. Since all four of these claims are part of Eaton＇s developmental argument, my rejection of them will not only show how the particular developmental story Eaton diagnoses is inaccurate, but will also explain what limits there actually are in Boyle＇s account of the intelligibility of mechanical explanations. My account will also show why important philosophers like Locke and Leibniz should be interested in Boyle＇s philosophical work.展开更多
Wind power forecasting(WPF)is important for safe,stable,and reliable integration of new energy technologies into power systems.Machine learning(ML)algorithms have recently attracted increasing attention in the field o...Wind power forecasting(WPF)is important for safe,stable,and reliable integration of new energy technologies into power systems.Machine learning(ML)algorithms have recently attracted increasing attention in the field of WPF.However,opaque decisions and lack of trustworthiness of black-box models for WPF could cause scheduling risks.This study develops a method for identifying risky models in practical applications and avoiding the risks.First,a local interpretable model-agnostic explanations algorithm is introduced and improved for WPF model analysis.On that basis,a novel index is presented to quantify the level at which neural networks or other black-box models can trust features involved in training.Then,by revealing the operational mechanism for local samples,human interpretability of the black-box model is examined under different accuracies,time horizons,and seasons.This interpretability provides a basis for several technical routes for WPF from the viewpoint of the forecasting model.Moreover,further improvements in accuracy of WPF are explored by evaluating possibilities of using interpretable ML models that use multi-horizons global trust modeling and multi-seasons interpretable feature selection methods.Experimental results from a wind farm in China show that error can be robustly reduced.展开更多
Contemporary demands necessitate the swift and accurate detection of cracks in critical infrastructures,including tunnels and pavements.This study proposed a transfer learning-based encoder-decoder method with visual ...Contemporary demands necessitate the swift and accurate detection of cracks in critical infrastructures,including tunnels and pavements.This study proposed a transfer learning-based encoder-decoder method with visual explanations for infrastructure crack segmentation.Firstly,a vast dataset containing 7089 images was developed,comprising diverse conditions—simple and complex crack patterns as well as clean and rough backgrounds.Secondly,leveraging transfer learning,an encoder-decoder model with visual explanations was formulated,utilizing varied pre-trained convolutional neural network(CNN)as the encoder.Visual explanations were achieved through gradient-weighted class activation mapping(Grad-CAM)to interpret the CNN segmentation model.Thirdly,accuracy,complexity(computation and model),and memory usage assessed CNN feasibility in practical engineering.Model performance was gauged via prediction and visual explanation.The investigation encompassed hyperparameters,data augmentation,deep learning from scratch vs.transfer learning,segmentation model architectures,segmentation model encoders,and encoder pre-training strategies.Results underscored transfer learning’s potency in enhancing CNN accuracy for crack segmentation,surpassing deep learning from scratch.Notably,encoder classification accuracy bore no significant correlation with CNN segmentation accuracy.Among all tested models,UNet-EfficientNet_B7 excelled in crack segmentation,harmonizing accuracy,complexity,memory usage,prediction,and visual explanation.展开更多
BACKGROUND Severe dengue children with critical complications have been attributed to high mortality rates,varying from approximately 1%to over 20%.To date,there is a lack of data on machine-learning-based algorithms ...BACKGROUND Severe dengue children with critical complications have been attributed to high mortality rates,varying from approximately 1%to over 20%.To date,there is a lack of data on machine-learning-based algorithms for predicting the risk of inhospital mortality in children with dengue shock syndrome(DSS).AIM To develop machine-learning models to estimate the risk of death in hospitalized children with DSS.METHODS This single-center retrospective study was conducted at tertiary Children’s Hospital No.2 in Viet Nam,between 2013 and 2022.The primary outcome was the in-hospital mortality rate in children with DSS admitted to the pediatric intensive care unit(PICU).Nine significant features were predetermined for further analysis using machine learning models.An oversampling method was used to enhance the model performance.Supervised models,including logistic regression,Naïve Bayes,Random Forest(RF),K-nearest neighbors,Decision Tree and Extreme Gradient Boosting(XGBoost),were employed to develop predictive models.The Shapley Additive Explanation was used to determine the degree of contribution of the features.RESULTS In total,1278 PICU-admitted children with complete data were included in the analysis.The median patient age was 8.1 years(interquartile range:5.4-10.7).Thirty-nine patients(3%)died.The RF and XGboost models demonstrated the highest performance.The Shapley Addictive Explanations model revealed that the most important predictive features included younger age,female patients,presence of underlying diseases,severe transaminitis,severe bleeding,low platelet counts requiring platelet transfusion,elevated levels of international normalized ratio,blood lactate and serum creatinine,large volume of resuscitation fluid and a high vasoactive inotropic score(>30).CONCLUSION We developed robust machine learning-based models to estimate the risk of death in hospitalized children with DSS.The study findings are applicable to the design of management schemes to enhance survival outcomes of patients with DSS.展开更多
Sinter is the core raw material for blast furnaces.Flue pressure,which is an important state parameter,affects sinter quality.In this paper,flue pressure prediction and optimization were studied based on the shapley a...Sinter is the core raw material for blast furnaces.Flue pressure,which is an important state parameter,affects sinter quality.In this paper,flue pressure prediction and optimization were studied based on the shapley additive explanation(SHAP)to predict the flue pressure and take targeted adjustment measures.First,the sintering process data were collected and processed.A flue pressure prediction model was then constructed after comparing different feature selection methods and model algorithms using SHAP+extremely random-ized trees(ET).The prediction accuracy of the model within the error range of±0.25 kPa was 92.63%.SHAP analysis was employed to improve the interpretability of the prediction model.The effects of various sintering operation parameters on flue pressure,the relation-ship between the numerical range of key operation parameters and flue pressure,the effect of operation parameter combinations on flue pressure,and the prediction process of the flue pressure prediction model on a single sample were analyzed.A flue pressure optimization module was also constructed and analyzed when the prediction satisfied the judgment conditions.The operating parameter combination was then pushed.The flue pressure was increased by 5.87%during the verification process,achieving a good optimization effect.展开更多
BACKGROUND Colorectal polyps are precancerous diseases of colorectal cancer.Early detection and resection of colorectal polyps can effectively reduce the mortality of colorectal cancer.Endoscopic mucosal resection(EMR...BACKGROUND Colorectal polyps are precancerous diseases of colorectal cancer.Early detection and resection of colorectal polyps can effectively reduce the mortality of colorectal cancer.Endoscopic mucosal resection(EMR)is a common polypectomy proce-dure in clinical practice,but it has a high postoperative recurrence rate.Currently,there is no predictive model for the recurrence of colorectal polyps after EMR.AIM To construct and validate a machine learning(ML)model for predicting the risk of colorectal polyp recurrence one year after EMR.METHODS This study retrospectively collected data from 1694 patients at three medical centers in Xuzhou.Additionally,a total of 166 patients were collected to form a prospective validation set.Feature variable screening was conducted using uni-variate and multivariate logistic regression analyses,and five ML algorithms were used to construct the predictive models.The optimal models were evaluated based on different performance metrics.Decision curve analysis(DCA)and SHapley Additive exPlanation(SHAP)analysis were performed to assess clinical applicability and predictor importance.RESULTS Multivariate logistic regression analysis identified 8 independent risk factors for colorectal polyp recurrence one year after EMR(P<0.05).Among the models,eXtreme Gradient Boosting(XGBoost)demonstrated the highest area under the curve(AUC)in the training set,internal validation set,and prospective validation set,with AUCs of 0.909(95%CI:0.89-0.92),0.921(95%CI:0.90-0.94),and 0.963(95%CI:0.94-0.99),respectively.DCA indicated favorable clinical utility for the XGBoost model.SHAP analysis identified smoking history,family history,and age as the top three most important predictors in the model.CONCLUSION The XGBoost model has the best predictive performance and can assist clinicians in providing individualized colonoscopy follow-up recommendations.展开更多
Colletotrichum kahawae(Coffee Berry Disease)spreads through spores that can be carried by wind,rain,and insects affecting coffee plantations,and causes 80%yield losses and poor-quality coffee beans.The deadly disease ...Colletotrichum kahawae(Coffee Berry Disease)spreads through spores that can be carried by wind,rain,and insects affecting coffee plantations,and causes 80%yield losses and poor-quality coffee beans.The deadly disease is hard to control because wind,rain,and insects carry spores.Colombian researchers utilized a deep learning system to identify CBD in coffee cherries at three growth stages and classify photographs of infected and uninfected cherries with 93%accuracy using a random forest method.If the dataset is too small and noisy,the algorithm may not learn data patterns and generate accurate predictions.To overcome the existing challenge,early detection of Colletotrichum Kahawae disease in coffee cherries requires automated processes,prompt recognition,and accurate classifications.The proposed methodology selects CBD image datasets through four different stages for training and testing.XGBoost to train a model on datasets of coffee berries,with each image labeled as healthy or diseased.Once themodel is trained,SHAP algorithmto figure out which features were essential formaking predictions with the proposed model.Some of these characteristics were the cherry’s colour,whether it had spots or other damage,and how big the Lesions were.Virtual inception is important for classification to virtualize the relationship between the colour of the berry is correlated with the presence of disease.To evaluate themodel’s performance andmitigate excess fitting,a 10-fold cross-validation approach is employed.This involves partitioning the dataset into ten subsets,training the model on each subset,and evaluating its performance.In comparison to other contemporary methodologies,the model put forth achieved an accuracy of 98.56%.展开更多
Boosting algorithms have been widely utilized in the development of landslide susceptibility mapping(LSM)studies.However,these algorithms possess distinct computational strategies and hyperparameters,making it challen...Boosting algorithms have been widely utilized in the development of landslide susceptibility mapping(LSM)studies.However,these algorithms possess distinct computational strategies and hyperparameters,making it challenging to propose an ideal LSM model.To investigate the impact of different boosting algorithms and hyperparameter optimization algorithms on LSM,this study constructed a geospatial database comprising 12 conditioning factors,such as elevation,stratum,and annual average rainfall.The XGBoost(XGB),LightGBM(LGBM),and CatBoost(CB)algorithms were employed to construct the LSM model.Furthermore,the Bayesian optimization(BO),particle swarm optimization(PSO),and Hyperband optimization(HO)algorithms were applied to optimizing the LSM model.The boosting algorithms exhibited varying performances,with CB demonstrating the highest precision,followed by LGBM,and XGB showing poorer precision.Additionally,the hyperparameter optimization algorithms displayed different performances,with HO outperforming PSO and BO showing poorer performance.The HO-CB model achieved the highest precision,boasting an accuracy of 0.764,an F1-score of 0.777,an area under the curve(AUC)value of 0.837 for the training set,and an AUC value of 0.863 for the test set.The model was interpreted using SHapley Additive exPlanations(SHAP),revealing that slope,curvature,topographic wetness index(TWI),degree of relief,and elevation significantly influenced landslides in the study area.This study offers a scientific reference for LSM and disaster prevention research.This study examines the utilization of various boosting algorithms and hyperparameter optimization algorithms in Wanzhou District.It proposes the HO-CB-SHAP framework as an effective approach to accurately forecast landslide disasters and interpret LSM models.However,limitations exist concerning the generalizability of the model and the data processing,which require further exploration in subsequent studies.展开更多
Today,urban traffic,growing populations,and dense transportation networks are contributing to an increase in traffic incidents.These incidents include traffic accidents,vehicle breakdowns,fires,and traffic disputes,re...Today,urban traffic,growing populations,and dense transportation networks are contributing to an increase in traffic incidents.These incidents include traffic accidents,vehicle breakdowns,fires,and traffic disputes,resulting in long waiting times,high carbon emissions,and other undesirable situations.It is vital to estimate incident response times quickly and accurately after traffic incidents occur for the success of incident-related planning and response activities.This study presents a model for forecasting the traffic incident duration of traffic events with high precision.The proposed model goes through a 4-stage process using various features to predict the duration of four different traffic events and presents a feature reduction approach to enable real-time data collection and prediction.In the first stage,the dataset consisting of 24,431 data points and 75 variables is prepared by data collection,merging,missing data processing and data cleaning.In the second stage,models such as Decision Trees(DT),K-Nearest Neighbour(KNN),Random Forest(RF)and Support Vector Machines(SVM)are used and hyperparameter optimisation is performed with GridSearchCV.In the third stage,feature selection and reduction are performed and real-time data are used.In the last stage,model performance with 14 variables is evaluated with metrics such as accuracy,precision,recall,F1-score,MCC,confusion matrix and SHAP.The RF model outperforms other models with an accuracy of 98.5%.The study’s prediction results demonstrate that the proposed dynamic prediction model can achieve a high level of success.展开更多
With the advancements in artificial intelligence(AI)technology,attackers are increasingly using sophisticated techniques,including ChatGPT.Endpoint Detection&Response(EDR)is a system that detects and responds to s...With the advancements in artificial intelligence(AI)technology,attackers are increasingly using sophisticated techniques,including ChatGPT.Endpoint Detection&Response(EDR)is a system that detects and responds to strange activities or security threats occurring on computers or endpoint devices within an organization.Unlike traditional antivirus software,EDR is more about responding to a threat after it has already occurred than blocking it.This study aims to overcome challenges in security control,such as increased log size,emerging security threats,and technical demands faced by control staff.Previous studies have focused on AI detection models,emphasizing detection rates and model performance.However,the underlying reasons behind the detection results were often insufficiently understood,leading to varying outcomes based on the learning model.Additionally,the presence of both structured or unstructured logs,the growth in new security threats,and increasing technical disparities among control staff members pose further challenges for effective security control.This study proposed to improve the problems of the existing EDR system and overcome the limitations of security control.This study analyzed data during the preprocessing stage to identify potential threat factors that influence the detection process and its outcomes.Additionally,eleven commonly-used machine learning(ML)models for malware detection in XAI were tested,with the five models showing the highest performance selected for further analysis.Explainable AI(XAI)techniques are employed to assess the impact of preprocessing on the learning process outcomes.To ensure objectivity and versatility in the analysis,five widely recognized datasets were used.Additionally,eleven commonly-used machine learning models for malware detection in XAI were tested with the five models showing the highest performance selected for further analysis.The results indicate that eXtreme Gradient Boosting(XGBoost)model outperformed others.Moreover,the study conducts an in-depth analysis of the preprocessing phase,tracing backward from the detection result to infer potential threats and classify the primary variables influencing the model’s prediction.This analysis includes the application of SHapley Additive exPlanations(SHAP),an XAI result,which provides insight into the influence of specific features on detection outcomes,and suggests potential breaches by identifying common parameters in malware through file backtracking and providing weights.This study also proposed a counter-detection analysis process to overcome the limitations of existing Deep Learning outcomes,understand the decision-making process of AI,and enhance reliability.These contributions are expected to significantly enhance EDR systems and address existing limitations in security control.展开更多
Objective To investigate the facial spectrum and color characteristics of patients with essen-tial hypertension post administering antihypertensive drugs,establish a classification and evaluation model based on the fa...Objective To investigate the facial spectrum and color characteristics of patients with essen-tial hypertension post administering antihypertensive drugs,establish a classification and evaluation model based on the facial colors of the enrolled patients,and perform in-depth analysis on the important characteristics of their facial spectrum.Methods From September 3,2018,to March 23,2024,participants with essential hyperten-sion(receiving antihypertensive medication treatment,hypertension group)and normal blood pressure(control group)were recruited from the Cardiology Department of Shanghai Hospital of Traditional Chinese Medicine,the Coronary Care Unit of Shanghai Tenth People's Hospital,the Physical Examination Center of Shuguang Hospital Affiliated to Shanghai Uni-versity of Traditional Chinese Medicine,and the Gaohang Community Health Service Center.This study employed the propensity score matching(PSM)method to reduce study partici-pants selection bias.Spectral information in the facial visible light spectrum of the subjects was collected using a flame spectrometer,and the spectral chromaticity values were calculat-ed using the equal-interval wavelength method.The study analyzed the differences in spec-tral reflectance across various facial regions,including the entire face,forehead,glabella,nose,jaw,left and right zygomatic regions,left and right cheek regions as well as differences in parameters within the Lab color space between the two subject groups.Feature selection was conducted using least absolute shrinkage and selection operator(LASSO)regression,fol-lowed by the application of various machine learning algorithms,including logistic regres-sion(LR),support vector machine(SVM),random forest(RF),Naïve Bayes(NB),and eX-treme Gradient Boosting(XGB).The reduced-dimensional dataset was split in a 7:3 ratio to establish a classification and assessment model for facial coloration related to primary hyper-tension.Additionally,model fusion techniques were applied to enhance the predictive power.The performance of the models was evaluated using metrics including the area under the curve(AUC)and accuracy.Shapley Additive exPlanations(SHAP)was used to interpret the outcomes of the models.Results A total of 114 participants were included in both hypertension and control groups.Reflectance analysis across the entire face and eight predefined areas revealed that the hypertensive group exhibited significantly higher reflectance of corresponding color light in the blue-violet region(P<0.05)and a lower reflectance in the red region(P<0.05)compared with control group.Analysis of Lab color space parameters across the entire face and eight predefined areas showed that hypertensive group had significantly lower a and b values than control group(P<0.05).LASSO regression analysis identified a total of 18 facial color features that were highly correlated with hypertension,including the a values of the chin and the right cheek,the reflectance at 380 nm and at 780 nm of the forehead.The results of the multi-mod-el classification showed that the RF classification model was the most effective,with an AUC of 0.74 and an accuracy of 0.77.The combined model of RF+LR+SVM outperformed a single model in their classification performance,achieving an AUC of 0.80 and an accuracy of 0.76.SHAP model visualization results indicated that the top three contributors to ideal prediction results based on the characteristics from the facial spectrum were the reflectance at 380 nm across the entire face and of the nose as well as the a value of the chin.Conclusion Within the same age group,patients with essential hypertension exhibited signif-icant and regular changes in facial color and facial spectral reflectance parameters after the administration of antihypertensive drugs.Furthermore,facial reflectance indicators,such as the overall reflectance at 380 nm and the a value of the chin,could offer valuable references for clinically assessing the drug efficacy and health status of patients with essential hyperten-sion.展开更多
A lightweight malware detection and family classification system for the Internet of Things (IoT) was designed to solve the difficulty of deploying defense models caused by the limited computing and storage resources ...A lightweight malware detection and family classification system for the Internet of Things (IoT) was designed to solve the difficulty of deploying defense models caused by the limited computing and storage resources of IoT devices. By training complex models with IoT software gray-scale images and utilizing the gradient-weighted class-activated mapping technique, the system can identify key codes that influence model decisions. This allows for the reconstruction of gray-scale images to train a lightweight model called LMDNet for malware detection. Additionally, the multi-teacher knowledge distillation method is employed to train KD-LMDNet, which focuses on classifying malware families. The results indicate that the model’s identification speed surpasses that of traditional methods by 23.68%. Moreover, the accuracy achieved on the Malimg dataset for family classification is an impressive 99.07%. Furthermore, with a model size of only 0.45M, it appears to be well-suited for the IoT environment. By training complex models using IoT software gray-scale images and utilizing the gradient-weighted class-activated mapping technique, the system can identify key codes that influence model decisions. This allows for the reconstruction of gray-scale images to train a lightweight model called LMDNet for malware detection. Thus, the presented approach can address the challenges associated with malware detection and family classification in IoT devices.展开更多
Alzheimer’s disease(AD)is a neurological disorder that predominantly affects the brain.In the coming years,it is expected to spread rapidly,with limited progress in diagnostic techniques.Various machine learning(ML)a...Alzheimer’s disease(AD)is a neurological disorder that predominantly affects the brain.In the coming years,it is expected to spread rapidly,with limited progress in diagnostic techniques.Various machine learning(ML)and artificial intelligence(AI)algorithms have been employed to detect AD using single-modality data.However,recent developments in ML have enabled the application of these methods to multiple data sources and input modalities for AD prediction.In this study,we developed a framework that utilizes multimodal data(tabular data,magnetic resonance imaging(MRI)images,and genetic information)to classify AD.As part of the pre-processing phase,we generated a knowledge graph from the tabular data and MRI images.We employed graph neural networks for knowledge graph creation,and region-based convolutional neural network approach for image-to-knowledge graph generation.Additionally,we integrated various explainable AI(XAI)techniques to interpret and elucidate the prediction outcomes derived from multimodal data.Layer-wise relevance propagation was used to explain the layer-wise outcomes in the MRI images.We also incorporated submodular pick local interpretable model-agnostic explanations to interpret the decision-making process based on the tabular data provided.Genetic expression values play a crucial role in AD analysis.We used a graphical gene tree to identify genes associated with the disease.Moreover,a dashboard was designed to display XAI outcomes,enabling experts and medical professionals to easily comprehend the predic-tion results.展开更多
Dear Editor,Scene understanding is an essential task in computer vision.The ultimate objective of scene understanding is to instruct computers to understand and reason about the scenes as humans do.Parallel vision is ...Dear Editor,Scene understanding is an essential task in computer vision.The ultimate objective of scene understanding is to instruct computers to understand and reason about the scenes as humans do.Parallel vision is a research framework that unifies the explanation and perception of dynamic and complex scenes.展开更多
Accurate prediction of molten steel temperature in the ladle furnace(LF)refining process has an important influence on the quality of molten steel and the control of steelmaking cost.Extensive research on establishing...Accurate prediction of molten steel temperature in the ladle furnace(LF)refining process has an important influence on the quality of molten steel and the control of steelmaking cost.Extensive research on establishing models to predict molten steel temperature has been conducted.However,most researchers focus solely on improving the accuracy of the model,neglecting its explainability.The present study aims to develop a high-precision and explainable model with improved reliability and transparency.The eXtreme gradient boosting(XGBoost)and light gradient boosting machine(LGBM)were utilized,along with bayesian optimization and grey wolf optimiz-ation(GWO),to establish the prediction model.Different performance evaluation metrics and graphical representations were applied to compare the optimal XGBoost and LGBM models obtained through varying hyperparameter optimization methods with the other models.The findings indicated that the GWO-LGBM model outperformed other methods in predicting molten steel temperature,with a high pre-diction accuracy of 89.35%within the error range of±5°C.The model’s learning/decision process was revealed,and the influence degree of different variables on the molten steel temperature was clarified using the tree structure visualization and SHapley Additive exPlana-tions(SHAP)analysis.Consequently,the explainability of the optimal GWO-LGBM model was enhanced,providing reliable support for prediction results.展开更多
文摘Predicting molecular properties is essential for advancing for advancing drug discovery and design. Recently, Graph Neural Networks (GNNs) have gained prominence due to their ability to capture the complex structural and relational information inherent in molecular graphs. Despite their effectiveness, the “black-box” nature of GNNs remains a significant obstacle to their widespread adoption in chemistry, as it hinders interpretability and trust. In this context, several explanation methods based on factual reasoning have emerged. These methods aim to interpret the predictions made by GNNs by analyzing the key features contributing to the prediction. However, these approaches fail to answer critical questions: “How to ensure that the structure-property mapping learned by GNNs is consistent with established domain knowledge”. In this paper, we propose MMGCF, a novel counterfactual explanation framework designed specifically for the prediction of GNN-based molecular properties. MMGCF constructs a hierarchical tree structure on molecular motifs, enabling the systematic generation of counterfactuals through motif perturbations. This framework identifies causally significant motifs and elucidates their impact on model predictions, offering insights into the relationship between structural modifications and predicted properties. Our method demonstrates its effectiveness through comprehensive quantitative and qualitative evaluations of four real-world molecular datasets.
基金support provided by The Science and Technology Development Fund,Macao SAR,China(File Nos.0057/2020/AGJ and SKL-IOTSC-2021-2023)Science and Technology Program of Guangdong Province,China(Grant No.2021A0505080009).
文摘Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the settlement caused by tunneling.However,well-performing ML models are usually less interpretable.Irrelevant input features decrease the performance and interpretability of an ML model.Nonetheless,feature selection,a critical step in the ML pipeline,is usually ignored in most studies that focused on predicting tunneling-induced settlement.This study applies four techniques,i.e.Pearson correlation method,sequential forward selection(SFS),sequential backward selection(SBS)and Boruta algorithm,to investigate the effect of feature selection on the model’s performance when predicting the tunneling-induced maximum surface settlement(S_(max)).The data set used in this study was compiled from two metro tunnel projects excavated in Hangzhou,China using earth pressure balance(EPB)shields and consists of 14 input features and a single output(i.e.S_(max)).The ML model that is trained on features selected from the Boruta algorithm demonstrates the best performance in both the training and testing phases.The relevant features chosen from the Boruta algorithm further indicate that tunneling-induced settlement is affected by parameters related to tunnel geometry,geological conditions and shield operation.The recently proposed Shapley additive explanations(SHAP)method explores how the input features contribute to the output of a complex ML model.It is observed that the larger settlements are induced during shield tunneling in silty clay.Moreover,the SHAP analysis reveals that the low magnitudes of face pressure at the top of the shield increase the model’s output。
文摘This paper takes a microanalytic perspective on the speech and gestures used by one teacher of ESL （English as a Second Language） in an intensive English program classroom. Videotaped excerpts from her intermediate-level grammar course were transcribed to represent the speech, gesture and other non-verbal behavior that accompanied unplanned explanations of vocabulary that arose during three focus-on-form lessons. The gesture classification system of McNeill （1992）, which delineates different types of hand movements （iconics metaphorics, deictics, beats）, was used to understand the role the gestures played in these explanations. Results suggest that gestures and other non-verbal behavior are forms of input to classroom second language learners that must be considered a salient factor in classroom-based SLA （Second Language Acquisition） research
文摘In this paper I examine the following claims by William Eaton in his monograph Boyle on Fire： （i） that Boyle＇s religious convictions led him to believe that the world was not completely explicable, and this shows that there is a shortcoming in the power of mechanical explanations; （ii） that mechanical explanations offer only sufficient, not necessary explanations, and this too was taken by Boyle to be a limit in the explanatory power of mechanical explanations; （iii） that the mature Boyle thought that there could be more intelligible explanatory models than mechanism; and （iv） that what Boyle says at any point in his career is incompatible with the statement of Maria Boas-Hall, i.e., that the mechanical hypothesis can explicate all natural phenomena. Since all four of these claims are part of Eaton＇s developmental argument, my rejection of them will not only show how the particular developmental story Eaton diagnoses is inaccurate, but will also explain what limits there actually are in Boyle＇s account of the intelligibility of mechanical explanations. My account will also show why important philosophers like Locke and Leibniz should be interested in Boyle＇s philosophical work.
基金supported by the National Key R&D Program of China(Technology and application of wind power/photovoltaic power prediction for promoting renewable energy consumption)under Grant(2018YFB0904200).
文摘Wind power forecasting(WPF)is important for safe,stable,and reliable integration of new energy technologies into power systems.Machine learning(ML)algorithms have recently attracted increasing attention in the field of WPF.However,opaque decisions and lack of trustworthiness of black-box models for WPF could cause scheduling risks.This study develops a method for identifying risky models in practical applications and avoiding the risks.First,a local interpretable model-agnostic explanations algorithm is introduced and improved for WPF model analysis.On that basis,a novel index is presented to quantify the level at which neural networks or other black-box models can trust features involved in training.Then,by revealing the operational mechanism for local samples,human interpretability of the black-box model is examined under different accuracies,time horizons,and seasons.This interpretability provides a basis for several technical routes for WPF from the viewpoint of the forecasting model.Moreover,further improvements in accuracy of WPF are explored by evaluating possibilities of using interpretable ML models that use multi-horizons global trust modeling and multi-seasons interpretable feature selection methods.Experimental results from a wind farm in China show that error can be robustly reduced.
基金the National Natural Science Foundation of China(Grant Nos.52090083 and 52378405)Key Technology R&D Plan of Yunnan Provincial Department of Science and Technology(Grant No.202303AA080003)for their financial support.
文摘Contemporary demands necessitate the swift and accurate detection of cracks in critical infrastructures,including tunnels and pavements.This study proposed a transfer learning-based encoder-decoder method with visual explanations for infrastructure crack segmentation.Firstly,a vast dataset containing 7089 images was developed,comprising diverse conditions—simple and complex crack patterns as well as clean and rough backgrounds.Secondly,leveraging transfer learning,an encoder-decoder model with visual explanations was formulated,utilizing varied pre-trained convolutional neural network(CNN)as the encoder.Visual explanations were achieved through gradient-weighted class activation mapping(Grad-CAM)to interpret the CNN segmentation model.Thirdly,accuracy,complexity(computation and model),and memory usage assessed CNN feasibility in practical engineering.Model performance was gauged via prediction and visual explanation.The investigation encompassed hyperparameters,data augmentation,deep learning from scratch vs.transfer learning,segmentation model architectures,segmentation model encoders,and encoder pre-training strategies.Results underscored transfer learning’s potency in enhancing CNN accuracy for crack segmentation,surpassing deep learning from scratch.Notably,encoder classification accuracy bore no significant correlation with CNN segmentation accuracy.Among all tested models,UNet-EfficientNet_B7 excelled in crack segmentation,harmonizing accuracy,complexity,memory usage,prediction,and visual explanation.
文摘BACKGROUND Severe dengue children with critical complications have been attributed to high mortality rates,varying from approximately 1%to over 20%.To date,there is a lack of data on machine-learning-based algorithms for predicting the risk of inhospital mortality in children with dengue shock syndrome(DSS).AIM To develop machine-learning models to estimate the risk of death in hospitalized children with DSS.METHODS This single-center retrospective study was conducted at tertiary Children’s Hospital No.2 in Viet Nam,between 2013 and 2022.The primary outcome was the in-hospital mortality rate in children with DSS admitted to the pediatric intensive care unit(PICU).Nine significant features were predetermined for further analysis using machine learning models.An oversampling method was used to enhance the model performance.Supervised models,including logistic regression,Naïve Bayes,Random Forest(RF),K-nearest neighbors,Decision Tree and Extreme Gradient Boosting(XGBoost),were employed to develop predictive models.The Shapley Additive Explanation was used to determine the degree of contribution of the features.RESULTS In total,1278 PICU-admitted children with complete data were included in the analysis.The median patient age was 8.1 years(interquartile range:5.4-10.7).Thirty-nine patients(3%)died.The RF and XGboost models demonstrated the highest performance.The Shapley Addictive Explanations model revealed that the most important predictive features included younger age,female patients,presence of underlying diseases,severe transaminitis,severe bleeding,low platelet counts requiring platelet transfusion,elevated levels of international normalized ratio,blood lactate and serum creatinine,large volume of resuscitation fluid and a high vasoactive inotropic score(>30).CONCLUSION We developed robust machine learning-based models to estimate the risk of death in hospitalized children with DSS.The study findings are applicable to the design of management schemes to enhance survival outcomes of patients with DSS.
基金supported by the General Program of the National Natural Science Foundation of China(No.52274326)the China Baowu Low Carbon Metallurgy Innovation Foundation(No.BWLCF202109)the Seventh Batch of Ten Thousand Talents Plan of China(No.ZX20220553).
文摘Sinter is the core raw material for blast furnaces.Flue pressure,which is an important state parameter,affects sinter quality.In this paper,flue pressure prediction and optimization were studied based on the shapley additive explanation(SHAP)to predict the flue pressure and take targeted adjustment measures.First,the sintering process data were collected and processed.A flue pressure prediction model was then constructed after comparing different feature selection methods and model algorithms using SHAP+extremely random-ized trees(ET).The prediction accuracy of the model within the error range of±0.25 kPa was 92.63%.SHAP analysis was employed to improve the interpretability of the prediction model.The effects of various sintering operation parameters on flue pressure,the relation-ship between the numerical range of key operation parameters and flue pressure,the effect of operation parameter combinations on flue pressure,and the prediction process of the flue pressure prediction model on a single sample were analyzed.A flue pressure optimization module was also constructed and analyzed when the prediction satisfied the judgment conditions.The operating parameter combination was then pushed.The flue pressure was increased by 5.87%during the verification process,achieving a good optimization effect.
文摘BACKGROUND Colorectal polyps are precancerous diseases of colorectal cancer.Early detection and resection of colorectal polyps can effectively reduce the mortality of colorectal cancer.Endoscopic mucosal resection(EMR)is a common polypectomy proce-dure in clinical practice,but it has a high postoperative recurrence rate.Currently,there is no predictive model for the recurrence of colorectal polyps after EMR.AIM To construct and validate a machine learning(ML)model for predicting the risk of colorectal polyp recurrence one year after EMR.METHODS This study retrospectively collected data from 1694 patients at three medical centers in Xuzhou.Additionally,a total of 166 patients were collected to form a prospective validation set.Feature variable screening was conducted using uni-variate and multivariate logistic regression analyses,and five ML algorithms were used to construct the predictive models.The optimal models were evaluated based on different performance metrics.Decision curve analysis(DCA)and SHapley Additive exPlanation(SHAP)analysis were performed to assess clinical applicability and predictor importance.RESULTS Multivariate logistic regression analysis identified 8 independent risk factors for colorectal polyp recurrence one year after EMR(P<0.05).Among the models,eXtreme Gradient Boosting(XGBoost)demonstrated the highest area under the curve(AUC)in the training set,internal validation set,and prospective validation set,with AUCs of 0.909(95%CI:0.89-0.92),0.921(95%CI:0.90-0.94),and 0.963(95%CI:0.94-0.99),respectively.DCA indicated favorable clinical utility for the XGBoost model.SHAP analysis identified smoking history,family history,and age as the top three most important predictors in the model.CONCLUSION The XGBoost model has the best predictive performance and can assist clinicians in providing individualized colonoscopy follow-up recommendations.
基金support from the Deanship for Research&Innovation,Ministry of Education in Saudi Arabia,under the Auspices of Project Number:IFP22UQU4281768DSR122.
文摘Colletotrichum kahawae(Coffee Berry Disease)spreads through spores that can be carried by wind,rain,and insects affecting coffee plantations,and causes 80%yield losses and poor-quality coffee beans.The deadly disease is hard to control because wind,rain,and insects carry spores.Colombian researchers utilized a deep learning system to identify CBD in coffee cherries at three growth stages and classify photographs of infected and uninfected cherries with 93%accuracy using a random forest method.If the dataset is too small and noisy,the algorithm may not learn data patterns and generate accurate predictions.To overcome the existing challenge,early detection of Colletotrichum Kahawae disease in coffee cherries requires automated processes,prompt recognition,and accurate classifications.The proposed methodology selects CBD image datasets through four different stages for training and testing.XGBoost to train a model on datasets of coffee berries,with each image labeled as healthy or diseased.Once themodel is trained,SHAP algorithmto figure out which features were essential formaking predictions with the proposed model.Some of these characteristics were the cherry’s colour,whether it had spots or other damage,and how big the Lesions were.Virtual inception is important for classification to virtualize the relationship between the colour of the berry is correlated with the presence of disease.To evaluate themodel’s performance andmitigate excess fitting,a 10-fold cross-validation approach is employed.This involves partitioning the dataset into ten subsets,training the model on each subset,and evaluating its performance.In comparison to other contemporary methodologies,the model put forth achieved an accuracy of 98.56%.
基金funded by the Natural Science Foundation of Chongqing(Grants No.CSTB2022NSCQ-MSX0594)the Humanities and Social Sciences Research Project of the Ministry of Education(Grants No.16YJCZH061).
文摘Boosting algorithms have been widely utilized in the development of landslide susceptibility mapping(LSM)studies.However,these algorithms possess distinct computational strategies and hyperparameters,making it challenging to propose an ideal LSM model.To investigate the impact of different boosting algorithms and hyperparameter optimization algorithms on LSM,this study constructed a geospatial database comprising 12 conditioning factors,such as elevation,stratum,and annual average rainfall.The XGBoost(XGB),LightGBM(LGBM),and CatBoost(CB)algorithms were employed to construct the LSM model.Furthermore,the Bayesian optimization(BO),particle swarm optimization(PSO),and Hyperband optimization(HO)algorithms were applied to optimizing the LSM model.The boosting algorithms exhibited varying performances,with CB demonstrating the highest precision,followed by LGBM,and XGB showing poorer precision.Additionally,the hyperparameter optimization algorithms displayed different performances,with HO outperforming PSO and BO showing poorer performance.The HO-CB model achieved the highest precision,boasting an accuracy of 0.764,an F1-score of 0.777,an area under the curve(AUC)value of 0.837 for the training set,and an AUC value of 0.863 for the test set.The model was interpreted using SHapley Additive exPlanations(SHAP),revealing that slope,curvature,topographic wetness index(TWI),degree of relief,and elevation significantly influenced landslides in the study area.This study offers a scientific reference for LSM and disaster prevention research.This study examines the utilization of various boosting algorithms and hyperparameter optimization algorithms in Wanzhou District.It proposes the HO-CB-SHAP framework as an effective approach to accurately forecast landslide disasters and interpret LSM models.However,limitations exist concerning the generalizability of the model and the data processing,which require further exploration in subsequent studies.
文摘Today,urban traffic,growing populations,and dense transportation networks are contributing to an increase in traffic incidents.These incidents include traffic accidents,vehicle breakdowns,fires,and traffic disputes,resulting in long waiting times,high carbon emissions,and other undesirable situations.It is vital to estimate incident response times quickly and accurately after traffic incidents occur for the success of incident-related planning and response activities.This study presents a model for forecasting the traffic incident duration of traffic events with high precision.The proposed model goes through a 4-stage process using various features to predict the duration of four different traffic events and presents a feature reduction approach to enable real-time data collection and prediction.In the first stage,the dataset consisting of 24,431 data points and 75 variables is prepared by data collection,merging,missing data processing and data cleaning.In the second stage,models such as Decision Trees(DT),K-Nearest Neighbour(KNN),Random Forest(RF)and Support Vector Machines(SVM)are used and hyperparameter optimisation is performed with GridSearchCV.In the third stage,feature selection and reduction are performed and real-time data are used.In the last stage,model performance with 14 variables is evaluated with metrics such as accuracy,precision,recall,F1-score,MCC,confusion matrix and SHAP.The RF model outperforms other models with an accuracy of 98.5%.The study’s prediction results demonstrate that the proposed dynamic prediction model can achieve a high level of success.
基金supported by Innovative Human Resource Development for Local Intellectualization program through the Institute of Information&Communications Technology Planning&Evaluation(IITP)grant funded by the Korea government(MSIT)(IITP-2024-RS-2022-00156287,50%)supported by the MSIT(Ministry of Science and ICT),Republic of Korea,under the Convergence Security Core Talent Training Business Support Program(IITP-2024-RS-2022-II221203,50%)supervised by the IITP(Institute of Information&Communications Technology Planning&Evaluation).
文摘With the advancements in artificial intelligence(AI)technology,attackers are increasingly using sophisticated techniques,including ChatGPT.Endpoint Detection&Response(EDR)is a system that detects and responds to strange activities or security threats occurring on computers or endpoint devices within an organization.Unlike traditional antivirus software,EDR is more about responding to a threat after it has already occurred than blocking it.This study aims to overcome challenges in security control,such as increased log size,emerging security threats,and technical demands faced by control staff.Previous studies have focused on AI detection models,emphasizing detection rates and model performance.However,the underlying reasons behind the detection results were often insufficiently understood,leading to varying outcomes based on the learning model.Additionally,the presence of both structured or unstructured logs,the growth in new security threats,and increasing technical disparities among control staff members pose further challenges for effective security control.This study proposed to improve the problems of the existing EDR system and overcome the limitations of security control.This study analyzed data during the preprocessing stage to identify potential threat factors that influence the detection process and its outcomes.Additionally,eleven commonly-used machine learning(ML)models for malware detection in XAI were tested,with the five models showing the highest performance selected for further analysis.Explainable AI(XAI)techniques are employed to assess the impact of preprocessing on the learning process outcomes.To ensure objectivity and versatility in the analysis,five widely recognized datasets were used.Additionally,eleven commonly-used machine learning models for malware detection in XAI were tested with the five models showing the highest performance selected for further analysis.The results indicate that eXtreme Gradient Boosting(XGBoost)model outperformed others.Moreover,the study conducts an in-depth analysis of the preprocessing phase,tracing backward from the detection result to infer potential threats and classify the primary variables influencing the model’s prediction.This analysis includes the application of SHapley Additive exPlanations(SHAP),an XAI result,which provides insight into the influence of specific features on detection outcomes,and suggests potential breaches by identifying common parameters in malware through file backtracking and providing weights.This study also proposed a counter-detection analysis process to overcome the limitations of existing Deep Learning outcomes,understand the decision-making process of AI,and enhance reliability.These contributions are expected to significantly enhance EDR systems and address existing limitations in security control.
基金National Natural Science Foundation of China(82104738 and 82004255)Key Discipline Construction Project of High-level Traditional Chinese Medicine of the National Administration of Traditional Chinese Medicine-Traditional Chinese Medical Diagnostics(ZYYZDXK-2023069).
文摘Objective To investigate the facial spectrum and color characteristics of patients with essen-tial hypertension post administering antihypertensive drugs,establish a classification and evaluation model based on the facial colors of the enrolled patients,and perform in-depth analysis on the important characteristics of their facial spectrum.Methods From September 3,2018,to March 23,2024,participants with essential hyperten-sion(receiving antihypertensive medication treatment,hypertension group)and normal blood pressure(control group)were recruited from the Cardiology Department of Shanghai Hospital of Traditional Chinese Medicine,the Coronary Care Unit of Shanghai Tenth People's Hospital,the Physical Examination Center of Shuguang Hospital Affiliated to Shanghai Uni-versity of Traditional Chinese Medicine,and the Gaohang Community Health Service Center.This study employed the propensity score matching(PSM)method to reduce study partici-pants selection bias.Spectral information in the facial visible light spectrum of the subjects was collected using a flame spectrometer,and the spectral chromaticity values were calculat-ed using the equal-interval wavelength method.The study analyzed the differences in spec-tral reflectance across various facial regions,including the entire face,forehead,glabella,nose,jaw,left and right zygomatic regions,left and right cheek regions as well as differences in parameters within the Lab color space between the two subject groups.Feature selection was conducted using least absolute shrinkage and selection operator(LASSO)regression,fol-lowed by the application of various machine learning algorithms,including logistic regres-sion(LR),support vector machine(SVM),random forest(RF),Naïve Bayes(NB),and eX-treme Gradient Boosting(XGB).The reduced-dimensional dataset was split in a 7:3 ratio to establish a classification and assessment model for facial coloration related to primary hyper-tension.Additionally,model fusion techniques were applied to enhance the predictive power.The performance of the models was evaluated using metrics including the area under the curve(AUC)and accuracy.Shapley Additive exPlanations(SHAP)was used to interpret the outcomes of the models.Results A total of 114 participants were included in both hypertension and control groups.Reflectance analysis across the entire face and eight predefined areas revealed that the hypertensive group exhibited significantly higher reflectance of corresponding color light in the blue-violet region(P<0.05)and a lower reflectance in the red region(P<0.05)compared with control group.Analysis of Lab color space parameters across the entire face and eight predefined areas showed that hypertensive group had significantly lower a and b values than control group(P<0.05).LASSO regression analysis identified a total of 18 facial color features that were highly correlated with hypertension,including the a values of the chin and the right cheek,the reflectance at 380 nm and at 780 nm of the forehead.The results of the multi-mod-el classification showed that the RF classification model was the most effective,with an AUC of 0.74 and an accuracy of 0.77.The combined model of RF+LR+SVM outperformed a single model in their classification performance,achieving an AUC of 0.80 and an accuracy of 0.76.SHAP model visualization results indicated that the top three contributors to ideal prediction results based on the characteristics from the facial spectrum were the reflectance at 380 nm across the entire face and of the nose as well as the a value of the chin.Conclusion Within the same age group,patients with essential hypertension exhibited signif-icant and regular changes in facial color and facial spectral reflectance parameters after the administration of antihypertensive drugs.Furthermore,facial reflectance indicators,such as the overall reflectance at 380 nm and the a value of the chin,could offer valuable references for clinically assessing the drug efficacy and health status of patients with essential hyperten-sion.
文摘A lightweight malware detection and family classification system for the Internet of Things (IoT) was designed to solve the difficulty of deploying defense models caused by the limited computing and storage resources of IoT devices. By training complex models with IoT software gray-scale images and utilizing the gradient-weighted class-activated mapping technique, the system can identify key codes that influence model decisions. This allows for the reconstruction of gray-scale images to train a lightweight model called LMDNet for malware detection. Additionally, the multi-teacher knowledge distillation method is employed to train KD-LMDNet, which focuses on classifying malware families. The results indicate that the model’s identification speed surpasses that of traditional methods by 23.68%. Moreover, the accuracy achieved on the Malimg dataset for family classification is an impressive 99.07%. Furthermore, with a model size of only 0.45M, it appears to be well-suited for the IoT environment. By training complex models using IoT software gray-scale images and utilizing the gradient-weighted class-activated mapping technique, the system can identify key codes that influence model decisions. This allows for the reconstruction of gray-scale images to train a lightweight model called LMDNet for malware detection. Thus, the presented approach can address the challenges associated with malware detection and family classification in IoT devices.
文摘Alzheimer’s disease(AD)is a neurological disorder that predominantly affects the brain.In the coming years,it is expected to spread rapidly,with limited progress in diagnostic techniques.Various machine learning(ML)and artificial intelligence(AI)algorithms have been employed to detect AD using single-modality data.However,recent developments in ML have enabled the application of these methods to multiple data sources and input modalities for AD prediction.In this study,we developed a framework that utilizes multimodal data(tabular data,magnetic resonance imaging(MRI)images,and genetic information)to classify AD.As part of the pre-processing phase,we generated a knowledge graph from the tabular data and MRI images.We employed graph neural networks for knowledge graph creation,and region-based convolutional neural network approach for image-to-knowledge graph generation.Additionally,we integrated various explainable AI(XAI)techniques to interpret and elucidate the prediction outcomes derived from multimodal data.Layer-wise relevance propagation was used to explain the layer-wise outcomes in the MRI images.We also incorporated submodular pick local interpretable model-agnostic explanations to interpret the decision-making process based on the tabular data provided.Genetic expression values play a crucial role in AD analysis.We used a graphical gene tree to identify genes associated with the disease.Moreover,a dashboard was designed to display XAI outcomes,enabling experts and medical professionals to easily comprehend the predic-tion results.
基金supported by the Natural Science Foundation for Young Scientists in Shaanxi Province of China (2023-JC-QN-0729)the Fundamental Research Funds for the Central Universities (GK202207008)。
文摘Dear Editor,Scene understanding is an essential task in computer vision.The ultimate objective of scene understanding is to instruct computers to understand and reason about the scenes as humans do.Parallel vision is a research framework that unifies the explanation and perception of dynamic and complex scenes.
基金financially supported by the National Natural Science Foundation of China(Nos.51974023 and 52374321)the funding of State Key Laboratory of Advanced Metallurgy,University of Science and Technology Beijing(No.41621005)the Youth Science and Technology Innovation Fund of Jianlong Group-University of Science and Technology Beijing(No.20231235).
文摘Accurate prediction of molten steel temperature in the ladle furnace(LF)refining process has an important influence on the quality of molten steel and the control of steelmaking cost.Extensive research on establishing models to predict molten steel temperature has been conducted.However,most researchers focus solely on improving the accuracy of the model,neglecting its explainability.The present study aims to develop a high-precision and explainable model with improved reliability and transparency.The eXtreme gradient boosting(XGBoost)and light gradient boosting machine(LGBM)were utilized,along with bayesian optimization and grey wolf optimiz-ation(GWO),to establish the prediction model.Different performance evaluation metrics and graphical representations were applied to compare the optimal XGBoost and LGBM models obtained through varying hyperparameter optimization methods with the other models.The findings indicated that the GWO-LGBM model outperformed other methods in predicting molten steel temperature,with a high pre-diction accuracy of 89.35%within the error range of±5°C.The model’s learning/decision process was revealed,and the influence degree of different variables on the molten steel temperature was clarified using the tree structure visualization and SHapley Additive exPlana-tions(SHAP)analysis.Consequently,the explainability of the optimal GWO-LGBM model was enhanced,providing reliable support for prediction results.
