Traditional inverse neural network(INN)approaches for inverse design typically require auxiliary feedforward networks,leading to increased computational complexity and architectural dependencies.This study introduces ...Traditional inverse neural network(INN)approaches for inverse design typically require auxiliary feedforward networks,leading to increased computational complexity and architectural dependencies.This study introduces a standalone INN methodology that eliminates the need for feedforward networks while maintaining high reconstruction accuracy.The approach integrates Principal Component Analysis(PCA)and Partial Least Squares(PLS)for optimized feature space learning,enabling the standalone INN to effectively capture bidirectionalmappings between geometric parameters and mechanical properties.Validation using established numerical datasets demonstrates that the standalone INN architecture achieves reconstruction accuracy equal or better than traditional tandem approaches while completely eliminating the workload and training time required for Feedforward Neural Networks(FNN).These findings contribute to AI methodology development by proving that standalone invertible architectures can achieve comparable performance to complex hybrid systems with significantly improved computational efficiency.展开更多
Structural colors based on metasurfaces have very promising applications in areas such as optical image encryption and color printing.Herein,we propose a deep learning-enabled reverse design of polarization-selective ...Structural colors based on metasurfaces have very promising applications in areas such as optical image encryption and color printing.Herein,we propose a deep learning-enabled reverse design of polarization-selective structural color based on coding metasurface.In this study,the long short-term memory(LSTM)neural network is presented to enable the forward and inverse mapping between coding metasurface structure and corresponding color.The results show that the method can achieve 98%accuracy for the forward prediction of color and 93%accuracy for the inverse design of the structure.Moreover,a cascaded architecture is adopted to train the inverse neural network model,which can solve the nonuniqueness problem of the polarization-selective color reverse design.This study provides a new path for the application and development of structural colors.展开更多
As one of the most seismically active regions,Sichuan Basin is a key area of seismological studies in China.This study applies a neural network model with attention mechanisms,simultaneously picking the P-wave arrival...As one of the most seismically active regions,Sichuan Basin is a key area of seismological studies in China.This study applies a neural network model with attention mechanisms,simultaneously picking the P-wave arrival times and determining the first-motion polarity.The polarity information is subsequently used to derive source focal mechanisms.The model is trained and tested using small to moderate earthquake data from June to December 2019 in Sichuan.We apply the trained model to predict first-motion polarity directions of earthquake recordings in Sichuan from January to May 2019,and then derive focal mechanism solutions using HASH algorithm with predicted results.Compared with the source mechanism solutions obtained by manual processing,the deep learning method picks more polarities from smaller events,resulting in more focal mechanism solutions.The catalog documents focal mechanism solutions of 22 events(M_(L) 2.6–4.8)from analysts during this period,whereas we obtain focal mechanism solutions of 53 events(M_(L) 1.9–4.8)through the deep learning method.The derived focal mechanism solutions for the same events are consistent with the manual solutions.This method provides an efficient way for the source mechanism inversion of small to moderate earthquakes in Sichuan region,with high stability and reliability.展开更多
Forecasting landslide deformation is challenging due to influence of various internal and external factors on the occurrence of systemic and localized heterogeneities.Despite the potential to improve landslide predict...Forecasting landslide deformation is challenging due to influence of various internal and external factors on the occurrence of systemic and localized heterogeneities.Despite the potential to improve landslide predictability,deep learning has yet to be sufficiently explored for complex deformation patterns associated with landslides and is inherently opaque.Herein,we developed a holistic landslide deformation forecasting method that considers spatiotemporal correlations of landslide deformation by integrating domain knowledge into interpretable deep learning.By spatially capturing the interconnections between multiple deformations from different observation points,our method contributes to the understanding and forecasting of landslide systematic behavior.By integrating specific domain knowledge relevant to each observation point and merging internal properties with external variables,the local heterogeneity is considered in our method,identifying deformation temporal patterns in different landslide zones.Case studies involving reservoir-induced landslides and creeping landslides demonstrated that our approach(1)enhances the accuracy of landslide deformation forecasting,(2)identifies significant contributing factors and their influence on spatiotemporal deformation characteristics,and(3)demonstrates how identifying these factors and patterns facilitates landslide forecasting.Our research offers a promising and pragmatic pathway toward a deeper understanding and forecasting of complex landslide behaviors.展开更多
Microseismic monitoring is essential for understanding subsurface dynamics and optimizing oil and gas pro-duction.However,traditional methods for the automatic detection of microseismic events rely heavily on characte...Microseismic monitoring is essential for understanding subsurface dynamics and optimizing oil and gas pro-duction.However,traditional methods for the automatic detection of microseismic events rely heavily on characteristic functions and human intervention,often resulting in suboptimal performance when dealing with complex and noisy data.In this study,we propose a novel approach that leverages deep learning frame to extract multiscale features from microseismic data using a TransUNet neural network.Our model integrates the ad-vantages of Transformer and UNet architectures to achieve high accuracy in multivariate image segmentation and precise picking of P-wave and S-wave first arrivals simultaneously.We validate our approach using both synthetic and field microseismic datasets recorded from gas storage monitoring and roof fracturing in a coal seam.The robustness of the proposed method has been verified in the testing of synthetic data with various levels of Gaussian and real background noises extracted from field data.The comparisons of the proposed method with UNet and SwinUNet in terms of the model architecture and classification performance demonstrate the Tran-sUNet achieves the optimal balance in its architecture and inference speed.With relatively low inference time and network complexity,it operates effectively in high-precision microseismic phase pickings.This advancement holds significant promise for enhancing microseismic monitoring technology in hydraulic fracturing and reser-voir monitoring applications.展开更多
Brain tumor classification is crucial for personalized treatment planning.Although deep learning-based Artificial Intelligence(AI)models can automatically analyze tumor images,fine details of small tumor regions may b...Brain tumor classification is crucial for personalized treatment planning.Although deep learning-based Artificial Intelligence(AI)models can automatically analyze tumor images,fine details of small tumor regions may be overlooked during global feature extraction.Therefore,we propose a brain tumor Magnetic Resonance Imaging(MRI)classification model based on a global-local parallel dual-branch structure.The global branch employs ResNet50 with a Multi-Head Self-Attention(MHSA)to capture global contextual information from whole brain images,while the local branch utilizes VGG16 to extract fine-grained features from segmented brain tumor regions.The features from both branches are processed through designed attention-enhanced feature fusion module to filter and integrate important features.Additionally,to address sample imbalance in the dataset,we introduce a category attention block to improve the recognition of minority classes.Experimental results indicate that our method achieved a classification accuracy of 98.04%and a micro-average Area Under the Curve(AUC)of 0.989 in the classification of three types of brain tumors,surpassing several existing pre-trained Convolutional Neural Network(CNN)models.Additionally,feature interpretability analysis validated the effectiveness of the proposed model.This suggests that the method holds significant potential for brain tumor image classification.展开更多
Marine heatwave(MHW)events refer to periods of significantly elevated sea surface temperatures(SST),persisting from days to months,with significant impacts on marine ecosystems,including increased mortality among mari...Marine heatwave(MHW)events refer to periods of significantly elevated sea surface temperatures(SST),persisting from days to months,with significant impacts on marine ecosystems,including increased mortality among marine life and coral bleaching.Forecasting MHW events are crucial to mitigate their harmful effects.This study presents a twostep forecasting process:short-term SST prediction followed by MHW event detection based on the forecasted SST.Firstly,we developed the“SST-MHW-DL”model using the ConvLSTM architecture,which incorporates an attention mechanism to enhance both SST forecasting and MHW event detection.The model utilizes SST data from the preceding 60 d to forecast SST and detect MHW events for the subsequent 15 d.Verification results for SST forecasting demonstrate a root mean square error(RMSE)of 0.64℃,a mean absolute percentage error(MAPE)of 2.05%,and a coefficient of determination(R^(2))of 0.85,indicating the model’s ability to accurately predict future temperatures by leveraging historical sea temperature information.For MHW event detection using forecasted SST,the evaluation metrics of“accuracy”,“precision”,and“recall”achieved values of 0.77,0.73,and 0.43,respectively,demonstrating the model’s capability to capture the occurrence of MHW events accurately.Furthermore,the attention-enhanced mechanism reveals that recent SST variations within the past 10 days have the most significant impact on forecasting accuracy,while variations in deep-sea regions and along the Taiwan Strait significantly contribute to the model’s efficacy in capturing spatial characteristics.Additionally,the proposed model and temporal mechanism were applied to detect MHWs in the Atlantic Ocean.By inputting 30 d of SST data,the model predicted SST with an RMSE of 1.02℃and an R^(2)of 0.94.The accuracy,precision,and recall for MHW detection were 0.79,0.78,and 0.62,respectively,further demonstrating the model’s robustness and usability.展开更多
Human activity recognition is a significant area of research in artificial intelligence for surveillance,healthcare,sports,and human-computer interaction applications.The article benchmarks the performance of You Only...Human activity recognition is a significant area of research in artificial intelligence for surveillance,healthcare,sports,and human-computer interaction applications.The article benchmarks the performance of You Only Look Once version 11-based(YOLOv11-based)architecture for multi-class human activity recognition.The article benchmarks the performance of You Only Look Once version 11-based(YOLOv11-based)architecture for multi-class human activity recognition.The dataset consists of 14,186 images across 19 activity classes,from dynamic activities such as running and swimming to static activities such as sitting and sleeping.Preprocessing included resizing all images to 512512 pixels,annotating them in YOLO’s bounding box format,and applying data augmentation methods such as flipping,rotation,and cropping to enhance model generalization.The proposed model was trained for 100 epochs with adaptive learning rate methods and hyperparameter optimization for performance improvement,with a mAP@0.5 of 74.93%and a mAP@0.5-0.95 of 64.11%,outperforming previous versions of YOLO(v10,v9,and v8)and general-purpose architectures like ResNet50 and EfficientNet.It exhibited improved precision and recall for all activity classes with high precision values of 0.76 for running,0.79 for swimming,0.80 for sitting,and 0.81 for sleeping,and was tested for real-time deployment with an inference time of 8.9 ms per image,being computationally light.Proposed YOLOv11’s improvements are attributed to architectural advancements like a more complex feature extraction process,better attention modules,and an anchor-free detection mechanism.While YOLOv10 was extremely stable in static activity recognition,YOLOv9 performed well in dynamic environments but suffered from overfitting,and YOLOv8,while being a decent baseline,failed to differentiate between overlapping static activities.The experimental results determine proposed YOLOv11 to be the most appropriate model,providing an ideal balance between accuracy,computational efficiency,and robustness for real-world deployment.Nevertheless,there exist certain issues to be addressed,particularly in discriminating against visually similar activities and the use of publicly available datasets.Future research will entail the inclusion of 3D data and multimodal sensor inputs,such as depth and motion information,for enhancing recognition accuracy and generalizability to challenging real-world environments.展开更多
To develop emerging electrode materials and improve the performances of batteries,the machine learning techniques can provide insights to discover,design and develop battery new materials in high-throughput way.In thi...To develop emerging electrode materials and improve the performances of batteries,the machine learning techniques can provide insights to discover,design and develop battery new materials in high-throughput way.In this paper,two deep learning models are developed and trained with two feature groups extracted from the Materials Project datasets to predict the battery electrochemical performances including average voltage,specific capacity and specific energy.The deep learning models are trained with the multilayer perceptron as the core.The Bayesian optimization and Monte Carlo methods are applied to improve the prediction accuracy of models.Based on 10 types of ion batteries,the correlation coefficients are maintained above 0.9 compared to DFT calculation results and the mean absolute error of the prediction results for voltages of two models can reach 0.41 V and 0.20 V,respectively.The electrochemical performance prediction times for the two trained models on thousands of batteries are only 72.9 ms and 75.7 ms.Besides,the two deep learning models are applied to approach the screening of emerging electrode materials for sodium-ion and potassium-ion batteries.This work can contribute to a high-throughput computational method to accelerate the rational and fast materials discovery and design.展开更多
The traditional teaching methods of one-way cultivation of students can no longer meet the requirements of talent cultivation at this stage.The issue of how to promote students from passive acceptance to the independe...The traditional teaching methods of one-way cultivation of students can no longer meet the requirements of talent cultivation at this stage.The issue of how to promote students from passive acceptance to the independent cognitive understanding stage(i.e.deep learning)has become the focus of geography teaching.Therefore,under the guidance of deep learning theory,this paper takes the“landforms”knowledge unit of the Humanistic Education Edition as an example,improves the classroom teaching means through the unit teaching mode,reconstructs the“landforms”teaching unit,and explores the specific teaching of high school geography unit based on deep learning.This study provides a good example and guidelines for high school geography teaching and learning.展开更多
Accurate soil moisture(SM)prediction is critical for understanding hydrological processes.Physics-based(PB)models exhibit large uncertainties in SM predictions arising from uncertain parameterizations and insufficient...Accurate soil moisture(SM)prediction is critical for understanding hydrological processes.Physics-based(PB)models exhibit large uncertainties in SM predictions arising from uncertain parameterizations and insufficient representation of land-surface processes.In addition to PB models,deep learning(DL)models have been widely used in SM predictions recently.However,few pure DL models have notably high success rates due to lacking physical information.Thus,we developed hybrid models to effectively integrate the outputs of PB models into DL models to improve SM predictions.To this end,we first developed a hybrid model based on the attention mechanism to take advantage of PB models at each forecast time scale(attention model).We further built an ensemble model that combined the advantages of different hybrid schemes(ensemble model).We utilized SM forecasts from the Global Forecast System to enhance the convolutional long short-term memory(ConvLSTM)model for 1–16 days of SM predictions.The performances of the proposed hybrid models were investigated and compared with two existing hybrid models.The results showed that the attention model could leverage benefits of PB models and achieved the best predictability of drought events among the different hybrid models.Moreover,the ensemble model performed best among all hybrid models at all forecast time scales and different soil conditions.It is highlighted that the ensemble model outperformed the pure DL model over 79.5%of in situ stations for 16-day predictions.These findings suggest that our proposed hybrid models can adequately exploit the benefits of PB model outputs to aid DL models in making SM predictions.展开更多
The Ultra-Wideband(UWB)Location-Based Service is receiving more and more attention due to its high ranging accuracy and good time resolution.However,the None-Line-of-Sight(NLOS)propagation may reduce the ranging accur...The Ultra-Wideband(UWB)Location-Based Service is receiving more and more attention due to its high ranging accuracy and good time resolution.However,the None-Line-of-Sight(NLOS)propagation may reduce the ranging accuracy for UWB localization system in indoor environment.So it is important to identify LOS and NLOS propagations before taking proper measures to improve the UWB localization accuracy.In this paper,a deep learning-based UWB NLOS/LOS classification algorithm called FCN-Attention is proposed.The proposed FCN-Attention algorithm utilizes a Fully Convolution Network(FCN)for improving feature extraction ability and a self-attention mechanism for enhancing feature description from the data to improve the classification accuracy.The proposed algorithm is evaluated using an open-source dataset,a local collected dataset and a mixed dataset created from these two datasets.The experiment result shows that the proposed FCN-Attention algorithm achieves classification accuracy of 88.24%on the open-source dataset,100%on the local collected dataset and 92.01%on the mixed dataset,which is better than the results from other evaluated NLOS/LOS classification algorithms in most scenarios in this paper.展开更多
The prediction of the mechanical properties of hot-rolled strips is a very complex,highly dimensional and nonlinear problem,and the published models might lack reliability,practicability and generalization.Thus,a new ...The prediction of the mechanical properties of hot-rolled strips is a very complex,highly dimensional and nonlinear problem,and the published models might lack reliability,practicability and generalization.Thus,a new model was proposed for predicting the mechanical properties of hot-rolled strips by deep learning.First,the one-dimensional numerical data were transformed into two-dimensional data for expressing the complex interaction between the influencing factors.Subsequently,a new convolutional network was proposed to establish the prediction model of tensile strength of hot-rolled strips,and an improved inception module was introduced into this network to abstract features from different scales.Many comparative experiments were carried out to find the optimal network structure and its hyperparameters.Finally,the prediction experiments were carried out on different models to evaluate the performance of the new convolutional network,which includes the stepwise regression,ridge regression,support vector machine,random forest,shallow neural network,Bayesian neural network,deep feed-forward network and improved LeNet-5 convolutional neural network.The results show that the proposed convolutional network has better prediction accuracy of the mechanical properties of hot-rolled strips compared with other models.展开更多
Landfill leaks pose a serious threat to environmental health,risking the contamination of both groundwater and soil resources.Accurate investigation of these sites is essential for implementing effective prevention an...Landfill leaks pose a serious threat to environmental health,risking the contamination of both groundwater and soil resources.Accurate investigation of these sites is essential for implementing effective prevention and control measures.The self-potential(SP)stands out for its sensitivity to contamination plumes,offering a solution for monitoring and detecting the movement and seepage of subsurface pollutants.However,traditional SP inversion techniques heavily rely on precise subsurface resistivity information.In this study,we propose the Attention U-Net deep learning network for rapid SP inversion.By incorporating an attention mechanism,this algorithm effectively learns the relationship between array-style SP data and the location and extent of subsurface contaminated sources.We designed a synthetic landfill model with a heterogeneous resistivity structure to assess the performance of Attention U-Net deep learning network.Additionally,we conducted further validation using a laboratory model to assess its practical applicability.The results demonstrate that the algorithm is not solely dependent on resistivity information,enabling effective locating of the source distribution,even in models with intricate subsurface structures.Our work provides a promising tool for SP data processing,enhancing the applicability of this method in the field of near-subsurface environmental monitoring.展开更多
Mangroves are indispensable to coastlines,maintaining biodiversity,and mitigating climate change.Therefore,improving the accuracy of mangrove information identification is crucial for their ecological protection.Aimin...Mangroves are indispensable to coastlines,maintaining biodiversity,and mitigating climate change.Therefore,improving the accuracy of mangrove information identification is crucial for their ecological protection.Aiming at the limited morphological information of synthetic aperture radar(SAR)images,which is greatly interfered by noise,and the susceptibility of optical images to weather and lighting conditions,this paper proposes a pixel-level weighted fusion method for SAR and optical images.Image fusion enhanced the target features and made mangrove monitoring more comprehensive and accurate.To address the problem of high similarity between mangrove forests and other forests,this paper is based on the U-Net convolutional neural network,and an attention mechanism is added in the feature extraction stage to make the model pay more attention to the mangrove vegetation area in the image.In order to accelerate the convergence and normalize the input,batch normalization(BN)layer and Dropout layer are added after each convolutional layer.Since mangroves are a minority class in the image,an improved cross-entropy loss function is introduced in this paper to improve the model’s ability to recognize mangroves.The AttU-Net model for mangrove recognition in high similarity environments is thus constructed based on the fused images.Through comparison experiments,the overall accuracy of the improved U-Net model trained from the fused images to recognize the predicted regions is significantly improved.Based on the fused images,the recognition results of the AttU-Net model proposed in this paper are compared with its benchmark model,U-Net,and the Dense-Net,Res-Net,and Seg-Net methods.The AttU-Net model captured mangroves’complex structures and textural features in images more effectively.The average OA,F1-score,and Kappa coefficient in the four tested regions were 94.406%,90.006%,and 84.045%,which were significantly higher than several other methods.This method can provide some technical support for the monitoring and protection of mangrove ecosystems.展开更多
Financial time series prediction,whether for classification or regression,has been a heated research topic over the last decade.While traditional machine learning algorithms have experienced mediocre results,deep lear...Financial time series prediction,whether for classification or regression,has been a heated research topic over the last decade.While traditional machine learning algorithms have experienced mediocre results,deep learning has largely contributed to the elevation of the prediction performance.Currently,the most up-to-date review of advanced machine learning techniques for financial time series prediction is still lacking,making it challenging for finance domain experts and relevant practitioners to determine which model potentially performs better,what techniques and components are involved,and how themodel can be designed and implemented.This review article provides an overview of techniques,components and frameworks for financial time series prediction,with an emphasis on state-of-the-art deep learning models in the literature from2015 to 2023,including standalonemodels like convolutional neural networks(CNN)that are capable of extracting spatial dependencies within data,and long short-term memory(LSTM)that is designed for handling temporal dependencies;and hybrid models integrating CNN,LSTM,attention mechanism(AM)and other techniques.For illustration and comparison purposes,models proposed in recent studies are mapped to relevant elements of a generalized framework comprised of input,output,feature extraction,prediction,and related processes.Among the state-of-the-artmodels,hybrid models like CNNLSTMand CNN-LSTM-AM in general have been reported superior in performance to stand-alone models like the CNN-only model.Some remaining challenges have been discussed,including non-friendliness for finance domain experts,delayed prediction,domain knowledge negligence,lack of standards,and inability of real-time and highfrequency predictions.The principal contributions of this paper are to provide a one-stop guide for both academia and industry to review,compare and summarize technologies and recent advances in this area,to facilitate smooth and informed implementation,and to highlight future research directions.展开更多
Acoustic source localization(ASL)and sound event detection(SED)are two widely pursued independent research fields.In recent years,in order to achieve a more complete spatial and temporal representation of sound field,...Acoustic source localization(ASL)and sound event detection(SED)are two widely pursued independent research fields.In recent years,in order to achieve a more complete spatial and temporal representation of sound field,sound event localization and detection(SELD)has become a very active research topic.This paper presents a deep learning-based multioverlapping sound event localization and detection algorithm in three-dimensional space.Log-Mel spectrum and generalized cross-correlation spectrum are joined together in channel dimension as input features.These features are classified and regressed in parallel after training by a neural network to obtain sound recognition and localization results respectively.The channel attention mechanism is also introduced in the network to selectively enhance the features containing essential information and suppress the useless features.Finally,a thourough comparison confirms the efficiency and effectiveness of the proposed SELD algorithm.Field experiments show that the proposed algorithm is robust to reverberation and environment and can achieve higher recognition and localization accuracy compared with the baseline method.展开更多
Higher requirements for the accuracy of relevant models are put throughout the transformation and upgrade of the iron and steel sector to intelligent production.It has been difficult to meet the needs of the field wit...Higher requirements for the accuracy of relevant models are put throughout the transformation and upgrade of the iron and steel sector to intelligent production.It has been difficult to meet the needs of the field with the usual prediction model of mechanical properties of hotrolled strip.Insufficient data and difficult parameter adjustment limit deep learning models based on multi-layer networks in practical applications;besides,the limited discrete process parameters used make it impossible to effectively depict the actual strip processing process.In order to solve these problems,this research proposed a new sampling approach for mechanical characteristics input data of hot-rolled strip based on the multi-grained cascade forest(gcForest)framework.According to the characteristics of complex process flow and abnormal sensitivity of process path and parameters to product quality in the hot-rolled strip production,a three-dimensional continuous time series process data sampling method based on time-temperature-deformation was designed.The basic information of strip steel(chemical composition and typical process parameters)is fused with the local process information collected by multi-grained scanning,so that the next link’s input has both local and global features.Furthermore,in the multi-grained scanning structure,a sub sampling scheme with a variable window was designed,so that input data with different dimensions can get output characteristics of the same dimension after passing through the multi-grained scanning structure,allowing the cascade forest structure to be trained normally.Finally,actual production data of three steel grades was used to conduct the experimental evaluation.The results revealed that the gcForest-based mechanical property prediction model outperforms the competition in terms of comprehensive performance,ease of parameter adjustment,and ability to sustain high prediction accuracy with fewer samples.展开更多
In recent years, there has been a revolution in the way that we transmit information through optical communication systems, allowing for fast and high-capacity data transmission using optical communication systems. Du...In recent years, there has been a revolution in the way that we transmit information through optical communication systems, allowing for fast and high-capacity data transmission using optical communication systems. Due to the growing demand for higher-capacity and faster networks, traditional optical communication systems are reaching their limits due to the increasing demand for faster and higher-capacity networks. The advent of machine learning and deep learning approaches has led to the emergence of powerful tools that can dramatically enhance the performance of optical communication systems with significant efficiency improvements. In this paper, we provide an overview of the role that machine learning (ML) and deep learning can play in enhancing the performance of various aspects of optical communication systems, including modulation techniques, channel modelling, equalization, and system optimization methods. The paper discusses the advantages of these approaches, such as improved spectral efficiency, reduced latency, and improved robustness to impairments in the channel, such as spectrum degradation. Additionally, a discussion is made regarding the potential challenges and limitations associated with using machine learning and deep learning in optical communication systems as well as their potential benefits. The purpose of this paper is to provide insight and highlight the potential of these approaches to improve optical communication in the future.展开更多
Total transmission plays an important role in efficiency improvement and wavefront control,and has made great progress in many applications,such as the optical film and signal transmission.Therefore,many traditional p...Total transmission plays an important role in efficiency improvement and wavefront control,and has made great progress in many applications,such as the optical film and signal transmission.Therefore,many traditional physical methods represented by transformation optics have been studied to achieve total transmission.However,these methods have strict limitations on the size of the photonic structure,and the calculation is complex.Here,we exploit deep learning to achieve this goal.In deep learning,the data-driven prediction and design are carried out by artificial neural networks(ANNs),which provide a convenient architecture for large dataset problems.By taking the transmission characteristic of the multi-layer stacks as an example,we demonstrate how optical materials can be designed by using ANNs.The trained network directly establishes the mapping from optical materials to transmission spectra,and enables the forward spectral prediction and inverse material design of total transmission in the given parameter space.Our work paves the way for the optical material design with special properties based on deep learning.展开更多
基金funding by the Deutsche Forschungsgemeinschaft under Germany’s Excellence Strategy within the Cluster of Excellence PhoenixD(EXC 2122,Project ID 390833453).
文摘Traditional inverse neural network(INN)approaches for inverse design typically require auxiliary feedforward networks,leading to increased computational complexity and architectural dependencies.This study introduces a standalone INN methodology that eliminates the need for feedforward networks while maintaining high reconstruction accuracy.The approach integrates Principal Component Analysis(PCA)and Partial Least Squares(PLS)for optimized feature space learning,enabling the standalone INN to effectively capture bidirectionalmappings between geometric parameters and mechanical properties.Validation using established numerical datasets demonstrates that the standalone INN architecture achieves reconstruction accuracy equal or better than traditional tandem approaches while completely eliminating the workload and training time required for Feedforward Neural Networks(FNN).These findings contribute to AI methodology development by proving that standalone invertible architectures can achieve comparable performance to complex hybrid systems with significantly improved computational efficiency.
基金supported by the National Natural Science Foundation of China(Grant Nos.62375137 and 62175114).
文摘Structural colors based on metasurfaces have very promising applications in areas such as optical image encryption and color printing.Herein,we propose a deep learning-enabled reverse design of polarization-selective structural color based on coding metasurface.In this study,the long short-term memory(LSTM)neural network is presented to enable the forward and inverse mapping between coding metasurface structure and corresponding color.The results show that the method can achieve 98%accuracy for the forward prediction of color and 93%accuracy for the inverse design of the structure.Moreover,a cascaded architecture is adopted to train the inverse neural network model,which can solve the nonuniqueness problem of the polarization-selective color reverse design.This study provides a new path for the application and development of structural colors.
基金the National Key R&D Program of China(2021YFC3000701)for the financial support。
文摘As one of the most seismically active regions,Sichuan Basin is a key area of seismological studies in China.This study applies a neural network model with attention mechanisms,simultaneously picking the P-wave arrival times and determining the first-motion polarity.The polarity information is subsequently used to derive source focal mechanisms.The model is trained and tested using small to moderate earthquake data from June to December 2019 in Sichuan.We apply the trained model to predict first-motion polarity directions of earthquake recordings in Sichuan from January to May 2019,and then derive focal mechanism solutions using HASH algorithm with predicted results.Compared with the source mechanism solutions obtained by manual processing,the deep learning method picks more polarities from smaller events,resulting in more focal mechanism solutions.The catalog documents focal mechanism solutions of 22 events(M_(L) 2.6–4.8)from analysts during this period,whereas we obtain focal mechanism solutions of 53 events(M_(L) 1.9–4.8)through the deep learning method.The derived focal mechanism solutions for the same events are consistent with the manual solutions.This method provides an efficient way for the source mechanism inversion of small to moderate earthquakes in Sichuan region,with high stability and reliability.
基金supported by the Postdoctoral Fellowship Program of CPSF(Grant No.GZB20230685)the National Science Foundation of China(Grant No.42277161).
文摘Forecasting landslide deformation is challenging due to influence of various internal and external factors on the occurrence of systemic and localized heterogeneities.Despite the potential to improve landslide predictability,deep learning has yet to be sufficiently explored for complex deformation patterns associated with landslides and is inherently opaque.Herein,we developed a holistic landslide deformation forecasting method that considers spatiotemporal correlations of landslide deformation by integrating domain knowledge into interpretable deep learning.By spatially capturing the interconnections between multiple deformations from different observation points,our method contributes to the understanding and forecasting of landslide systematic behavior.By integrating specific domain knowledge relevant to each observation point and merging internal properties with external variables,the local heterogeneity is considered in our method,identifying deformation temporal patterns in different landslide zones.Case studies involving reservoir-induced landslides and creeping landslides demonstrated that our approach(1)enhances the accuracy of landslide deformation forecasting,(2)identifies significant contributing factors and their influence on spatiotemporal deformation characteristics,and(3)demonstrates how identifying these factors and patterns facilitates landslide forecasting.Our research offers a promising and pragmatic pathway toward a deeper understanding and forecasting of complex landslide behaviors.
基金supported by a National Natural Science Foundation of China(Grant number 41974150 and 42174158)Natural Science Basic Research Program of Shaanxi(2023-JC-YB-220).
文摘Microseismic monitoring is essential for understanding subsurface dynamics and optimizing oil and gas pro-duction.However,traditional methods for the automatic detection of microseismic events rely heavily on characteristic functions and human intervention,often resulting in suboptimal performance when dealing with complex and noisy data.In this study,we propose a novel approach that leverages deep learning frame to extract multiscale features from microseismic data using a TransUNet neural network.Our model integrates the ad-vantages of Transformer and UNet architectures to achieve high accuracy in multivariate image segmentation and precise picking of P-wave and S-wave first arrivals simultaneously.We validate our approach using both synthetic and field microseismic datasets recorded from gas storage monitoring and roof fracturing in a coal seam.The robustness of the proposed method has been verified in the testing of synthetic data with various levels of Gaussian and real background noises extracted from field data.The comparisons of the proposed method with UNet and SwinUNet in terms of the model architecture and classification performance demonstrate the Tran-sUNet achieves the optimal balance in its architecture and inference speed.With relatively low inference time and network complexity,it operates effectively in high-precision microseismic phase pickings.This advancement holds significant promise for enhancing microseismic monitoring technology in hydraulic fracturing and reser-voir monitoring applications.
文摘Brain tumor classification is crucial for personalized treatment planning.Although deep learning-based Artificial Intelligence(AI)models can automatically analyze tumor images,fine details of small tumor regions may be overlooked during global feature extraction.Therefore,we propose a brain tumor Magnetic Resonance Imaging(MRI)classification model based on a global-local parallel dual-branch structure.The global branch employs ResNet50 with a Multi-Head Self-Attention(MHSA)to capture global contextual information from whole brain images,while the local branch utilizes VGG16 to extract fine-grained features from segmented brain tumor regions.The features from both branches are processed through designed attention-enhanced feature fusion module to filter and integrate important features.Additionally,to address sample imbalance in the dataset,we introduce a category attention block to improve the recognition of minority classes.Experimental results indicate that our method achieved a classification accuracy of 98.04%and a micro-average Area Under the Curve(AUC)of 0.989 in the classification of three types of brain tumors,surpassing several existing pre-trained Convolutional Neural Network(CNN)models.Additionally,feature interpretability analysis validated the effectiveness of the proposed model.This suggests that the method holds significant potential for brain tumor image classification.
基金The National Natural Science Foundation of China under contract Nos 42376175,42090044 and U2006211。
文摘Marine heatwave(MHW)events refer to periods of significantly elevated sea surface temperatures(SST),persisting from days to months,with significant impacts on marine ecosystems,including increased mortality among marine life and coral bleaching.Forecasting MHW events are crucial to mitigate their harmful effects.This study presents a twostep forecasting process:short-term SST prediction followed by MHW event detection based on the forecasted SST.Firstly,we developed the“SST-MHW-DL”model using the ConvLSTM architecture,which incorporates an attention mechanism to enhance both SST forecasting and MHW event detection.The model utilizes SST data from the preceding 60 d to forecast SST and detect MHW events for the subsequent 15 d.Verification results for SST forecasting demonstrate a root mean square error(RMSE)of 0.64℃,a mean absolute percentage error(MAPE)of 2.05%,and a coefficient of determination(R^(2))of 0.85,indicating the model’s ability to accurately predict future temperatures by leveraging historical sea temperature information.For MHW event detection using forecasted SST,the evaluation metrics of“accuracy”,“precision”,and“recall”achieved values of 0.77,0.73,and 0.43,respectively,demonstrating the model’s capability to capture the occurrence of MHW events accurately.Furthermore,the attention-enhanced mechanism reveals that recent SST variations within the past 10 days have the most significant impact on forecasting accuracy,while variations in deep-sea regions and along the Taiwan Strait significantly contribute to the model’s efficacy in capturing spatial characteristics.Additionally,the proposed model and temporal mechanism were applied to detect MHWs in the Atlantic Ocean.By inputting 30 d of SST data,the model predicted SST with an RMSE of 1.02℃and an R^(2)of 0.94.The accuracy,precision,and recall for MHW detection were 0.79,0.78,and 0.62,respectively,further demonstrating the model’s robustness and usability.
基金supported by King Saud University,Riyadh,Saudi Arabia,under Ongoing Research Funding Program(ORF-2025-951).
文摘Human activity recognition is a significant area of research in artificial intelligence for surveillance,healthcare,sports,and human-computer interaction applications.The article benchmarks the performance of You Only Look Once version 11-based(YOLOv11-based)architecture for multi-class human activity recognition.The article benchmarks the performance of You Only Look Once version 11-based(YOLOv11-based)architecture for multi-class human activity recognition.The dataset consists of 14,186 images across 19 activity classes,from dynamic activities such as running and swimming to static activities such as sitting and sleeping.Preprocessing included resizing all images to 512512 pixels,annotating them in YOLO’s bounding box format,and applying data augmentation methods such as flipping,rotation,and cropping to enhance model generalization.The proposed model was trained for 100 epochs with adaptive learning rate methods and hyperparameter optimization for performance improvement,with a mAP@0.5 of 74.93%and a mAP@0.5-0.95 of 64.11%,outperforming previous versions of YOLO(v10,v9,and v8)and general-purpose architectures like ResNet50 and EfficientNet.It exhibited improved precision and recall for all activity classes with high precision values of 0.76 for running,0.79 for swimming,0.80 for sitting,and 0.81 for sleeping,and was tested for real-time deployment with an inference time of 8.9 ms per image,being computationally light.Proposed YOLOv11’s improvements are attributed to architectural advancements like a more complex feature extraction process,better attention modules,and an anchor-free detection mechanism.While YOLOv10 was extremely stable in static activity recognition,YOLOv9 performed well in dynamic environments but suffered from overfitting,and YOLOv8,while being a decent baseline,failed to differentiate between overlapping static activities.The experimental results determine proposed YOLOv11 to be the most appropriate model,providing an ideal balance between accuracy,computational efficiency,and robustness for real-world deployment.Nevertheless,there exist certain issues to be addressed,particularly in discriminating against visually similar activities and the use of publicly available datasets.Future research will entail the inclusion of 3D data and multimodal sensor inputs,such as depth and motion information,for enhancing recognition accuracy and generalizability to challenging real-world environments.
基金supported by the National Natural Science Foundation of China(No.52102470).
文摘To develop emerging electrode materials and improve the performances of batteries,the machine learning techniques can provide insights to discover,design and develop battery new materials in high-throughput way.In this paper,two deep learning models are developed and trained with two feature groups extracted from the Materials Project datasets to predict the battery electrochemical performances including average voltage,specific capacity and specific energy.The deep learning models are trained with the multilayer perceptron as the core.The Bayesian optimization and Monte Carlo methods are applied to improve the prediction accuracy of models.Based on 10 types of ion batteries,the correlation coefficients are maintained above 0.9 compared to DFT calculation results and the mean absolute error of the prediction results for voltages of two models can reach 0.41 V and 0.20 V,respectively.The electrochemical performance prediction times for the two trained models on thousands of batteries are only 72.9 ms and 75.7 ms.Besides,the two deep learning models are applied to approach the screening of emerging electrode materials for sodium-ion and potassium-ion batteries.This work can contribute to a high-throughput computational method to accelerate the rational and fast materials discovery and design.
文摘The traditional teaching methods of one-way cultivation of students can no longer meet the requirements of talent cultivation at this stage.The issue of how to promote students from passive acceptance to the independent cognitive understanding stage(i.e.deep learning)has become the focus of geography teaching.Therefore,under the guidance of deep learning theory,this paper takes the“landforms”knowledge unit of the Humanistic Education Edition as an example,improves the classroom teaching means through the unit teaching mode,reconstructs the“landforms”teaching unit,and explores the specific teaching of high school geography unit based on deep learning.This study provides a good example and guidelines for high school geography teaching and learning.
基金supported by the Natural Science Foundation of China(Grant Nos.42088101 and 42205149)Zhongwang WEI was supported by the Natural Science Foundation of China(Grant No.42075158)+1 种基金Wei SHANGGUAN was supported by the Natural Science Foundation of China(Grant No.41975122)Yonggen ZHANG was supported by the National Natural Science Foundation of Tianjin(Grant No.20JCQNJC01660).
文摘Accurate soil moisture(SM)prediction is critical for understanding hydrological processes.Physics-based(PB)models exhibit large uncertainties in SM predictions arising from uncertain parameterizations and insufficient representation of land-surface processes.In addition to PB models,deep learning(DL)models have been widely used in SM predictions recently.However,few pure DL models have notably high success rates due to lacking physical information.Thus,we developed hybrid models to effectively integrate the outputs of PB models into DL models to improve SM predictions.To this end,we first developed a hybrid model based on the attention mechanism to take advantage of PB models at each forecast time scale(attention model).We further built an ensemble model that combined the advantages of different hybrid schemes(ensemble model).We utilized SM forecasts from the Global Forecast System to enhance the convolutional long short-term memory(ConvLSTM)model for 1–16 days of SM predictions.The performances of the proposed hybrid models were investigated and compared with two existing hybrid models.The results showed that the attention model could leverage benefits of PB models and achieved the best predictability of drought events among the different hybrid models.Moreover,the ensemble model performed best among all hybrid models at all forecast time scales and different soil conditions.It is highlighted that the ensemble model outperformed the pure DL model over 79.5%of in situ stations for 16-day predictions.These findings suggest that our proposed hybrid models can adequately exploit the benefits of PB model outputs to aid DL models in making SM predictions.
基金supported by the National Key Research and Development Program of China[grant No.2016YF B0502200]the Postdoctoral Research Foundation of China[grant No.2020M682480]the Fundamental Research Funds for the Central Universities[grant No.2042021kf0009]。
文摘The Ultra-Wideband(UWB)Location-Based Service is receiving more and more attention due to its high ranging accuracy and good time resolution.However,the None-Line-of-Sight(NLOS)propagation may reduce the ranging accuracy for UWB localization system in indoor environment.So it is important to identify LOS and NLOS propagations before taking proper measures to improve the UWB localization accuracy.In this paper,a deep learning-based UWB NLOS/LOS classification algorithm called FCN-Attention is proposed.The proposed FCN-Attention algorithm utilizes a Fully Convolution Network(FCN)for improving feature extraction ability and a self-attention mechanism for enhancing feature description from the data to improve the classification accuracy.The proposed algorithm is evaluated using an open-source dataset,a local collected dataset and a mixed dataset created from these two datasets.The experiment result shows that the proposed FCN-Attention algorithm achieves classification accuracy of 88.24%on the open-source dataset,100%on the local collected dataset and 92.01%on the mixed dataset,which is better than the results from other evaluated NLOS/LOS classification algorithms in most scenarios in this paper.
基金This research is supported by National Natural Science Foundation of China(51774219).
文摘The prediction of the mechanical properties of hot-rolled strips is a very complex,highly dimensional and nonlinear problem,and the published models might lack reliability,practicability and generalization.Thus,a new model was proposed for predicting the mechanical properties of hot-rolled strips by deep learning.First,the one-dimensional numerical data were transformed into two-dimensional data for expressing the complex interaction between the influencing factors.Subsequently,a new convolutional network was proposed to establish the prediction model of tensile strength of hot-rolled strips,and an improved inception module was introduced into this network to abstract features from different scales.Many comparative experiments were carried out to find the optimal network structure and its hyperparameters.Finally,the prediction experiments were carried out on different models to evaluate the performance of the new convolutional network,which includes the stepwise regression,ridge regression,support vector machine,random forest,shallow neural network,Bayesian neural network,deep feed-forward network and improved LeNet-5 convolutional neural network.The results show that the proposed convolutional network has better prediction accuracy of the mechanical properties of hot-rolled strips compared with other models.
基金Projects(42174170,41874145,72088101)supported by the National Natural Science Foundation of ChinaProject(CX20200228)supported by the Hunan Provincial Innovation Foundation for Postgraduate,China。
文摘Landfill leaks pose a serious threat to environmental health,risking the contamination of both groundwater and soil resources.Accurate investigation of these sites is essential for implementing effective prevention and control measures.The self-potential(SP)stands out for its sensitivity to contamination plumes,offering a solution for monitoring and detecting the movement and seepage of subsurface pollutants.However,traditional SP inversion techniques heavily rely on precise subsurface resistivity information.In this study,we propose the Attention U-Net deep learning network for rapid SP inversion.By incorporating an attention mechanism,this algorithm effectively learns the relationship between array-style SP data and the location and extent of subsurface contaminated sources.We designed a synthetic landfill model with a heterogeneous resistivity structure to assess the performance of Attention U-Net deep learning network.Additionally,we conducted further validation using a laboratory model to assess its practical applicability.The results demonstrate that the algorithm is not solely dependent on resistivity information,enabling effective locating of the source distribution,even in models with intricate subsurface structures.Our work provides a promising tool for SP data processing,enhancing the applicability of this method in the field of near-subsurface environmental monitoring.
基金The Key R&D Project of Hainan Province under contract No.ZDYF2023SHFZ097the National Natural Science Foundation of China under contract No.42376180。
文摘Mangroves are indispensable to coastlines,maintaining biodiversity,and mitigating climate change.Therefore,improving the accuracy of mangrove information identification is crucial for their ecological protection.Aiming at the limited morphological information of synthetic aperture radar(SAR)images,which is greatly interfered by noise,and the susceptibility of optical images to weather and lighting conditions,this paper proposes a pixel-level weighted fusion method for SAR and optical images.Image fusion enhanced the target features and made mangrove monitoring more comprehensive and accurate.To address the problem of high similarity between mangrove forests and other forests,this paper is based on the U-Net convolutional neural network,and an attention mechanism is added in the feature extraction stage to make the model pay more attention to the mangrove vegetation area in the image.In order to accelerate the convergence and normalize the input,batch normalization(BN)layer and Dropout layer are added after each convolutional layer.Since mangroves are a minority class in the image,an improved cross-entropy loss function is introduced in this paper to improve the model’s ability to recognize mangroves.The AttU-Net model for mangrove recognition in high similarity environments is thus constructed based on the fused images.Through comparison experiments,the overall accuracy of the improved U-Net model trained from the fused images to recognize the predicted regions is significantly improved.Based on the fused images,the recognition results of the AttU-Net model proposed in this paper are compared with its benchmark model,U-Net,and the Dense-Net,Res-Net,and Seg-Net methods.The AttU-Net model captured mangroves’complex structures and textural features in images more effectively.The average OA,F1-score,and Kappa coefficient in the four tested regions were 94.406%,90.006%,and 84.045%,which were significantly higher than several other methods.This method can provide some technical support for the monitoring and protection of mangrove ecosystems.
基金funded by the Natural Science Foundation of Fujian Province,China (Grant No.2022J05291)Xiamen Scientific Research Funding for Overseas Chinese Scholars.
文摘Financial time series prediction,whether for classification or regression,has been a heated research topic over the last decade.While traditional machine learning algorithms have experienced mediocre results,deep learning has largely contributed to the elevation of the prediction performance.Currently,the most up-to-date review of advanced machine learning techniques for financial time series prediction is still lacking,making it challenging for finance domain experts and relevant practitioners to determine which model potentially performs better,what techniques and components are involved,and how themodel can be designed and implemented.This review article provides an overview of techniques,components and frameworks for financial time series prediction,with an emphasis on state-of-the-art deep learning models in the literature from2015 to 2023,including standalonemodels like convolutional neural networks(CNN)that are capable of extracting spatial dependencies within data,and long short-term memory(LSTM)that is designed for handling temporal dependencies;and hybrid models integrating CNN,LSTM,attention mechanism(AM)and other techniques.For illustration and comparison purposes,models proposed in recent studies are mapped to relevant elements of a generalized framework comprised of input,output,feature extraction,prediction,and related processes.Among the state-of-the-artmodels,hybrid models like CNNLSTMand CNN-LSTM-AM in general have been reported superior in performance to stand-alone models like the CNN-only model.Some remaining challenges have been discussed,including non-friendliness for finance domain experts,delayed prediction,domain knowledge negligence,lack of standards,and inability of real-time and highfrequency predictions.The principal contributions of this paper are to provide a one-stop guide for both academia and industry to review,compare and summarize technologies and recent advances in this area,to facilitate smooth and informed implementation,and to highlight future research directions.
基金supported by the National Natural Science Foundation of China(61877067)the Foundation of Science and Technology on Near-Surface Detection Laboratory(TCGZ2019A002,TCGZ2021C003,6142414200511)the Natural Science Basic Research Program of Shaanxi(2021JZ-19)。
文摘Acoustic source localization(ASL)and sound event detection(SED)are two widely pursued independent research fields.In recent years,in order to achieve a more complete spatial and temporal representation of sound field,sound event localization and detection(SELD)has become a very active research topic.This paper presents a deep learning-based multioverlapping sound event localization and detection algorithm in three-dimensional space.Log-Mel spectrum and generalized cross-correlation spectrum are joined together in channel dimension as input features.These features are classified and regressed in parallel after training by a neural network to obtain sound recognition and localization results respectively.The channel attention mechanism is also introduced in the network to selectively enhance the features containing essential information and suppress the useless features.Finally,a thourough comparison confirms the efficiency and effectiveness of the proposed SELD algorithm.Field experiments show that the proposed algorithm is robust to reverberation and environment and can achieve higher recognition and localization accuracy compared with the baseline method.
基金financially supported by the National Natural Science Foundation of China(No.52004029)the Fundamental Research Funds for the Central Universities,China(No.FRF-TT-20-06).
文摘Higher requirements for the accuracy of relevant models are put throughout the transformation and upgrade of the iron and steel sector to intelligent production.It has been difficult to meet the needs of the field with the usual prediction model of mechanical properties of hotrolled strip.Insufficient data and difficult parameter adjustment limit deep learning models based on multi-layer networks in practical applications;besides,the limited discrete process parameters used make it impossible to effectively depict the actual strip processing process.In order to solve these problems,this research proposed a new sampling approach for mechanical characteristics input data of hot-rolled strip based on the multi-grained cascade forest(gcForest)framework.According to the characteristics of complex process flow and abnormal sensitivity of process path and parameters to product quality in the hot-rolled strip production,a three-dimensional continuous time series process data sampling method based on time-temperature-deformation was designed.The basic information of strip steel(chemical composition and typical process parameters)is fused with the local process information collected by multi-grained scanning,so that the next link’s input has both local and global features.Furthermore,in the multi-grained scanning structure,a sub sampling scheme with a variable window was designed,so that input data with different dimensions can get output characteristics of the same dimension after passing through the multi-grained scanning structure,allowing the cascade forest structure to be trained normally.Finally,actual production data of three steel grades was used to conduct the experimental evaluation.The results revealed that the gcForest-based mechanical property prediction model outperforms the competition in terms of comprehensive performance,ease of parameter adjustment,and ability to sustain high prediction accuracy with fewer samples.
文摘In recent years, there has been a revolution in the way that we transmit information through optical communication systems, allowing for fast and high-capacity data transmission using optical communication systems. Due to the growing demand for higher-capacity and faster networks, traditional optical communication systems are reaching their limits due to the increasing demand for faster and higher-capacity networks. The advent of machine learning and deep learning approaches has led to the emergence of powerful tools that can dramatically enhance the performance of optical communication systems with significant efficiency improvements. In this paper, we provide an overview of the role that machine learning (ML) and deep learning can play in enhancing the performance of various aspects of optical communication systems, including modulation techniques, channel modelling, equalization, and system optimization methods. The paper discusses the advantages of these approaches, such as improved spectral efficiency, reduced latency, and improved robustness to impairments in the channel, such as spectrum degradation. Additionally, a discussion is made regarding the potential challenges and limitations associated with using machine learning and deep learning in optical communication systems as well as their potential benefits. The purpose of this paper is to provide insight and highlight the potential of these approaches to improve optical communication in the future.
基金supported by the National Key Research and Development Program of China under Grant No.2020YFA0710100the National Natural Science Foundation of China under Grants No.92050102,No.11874311,and No.11504306the Fundamental Research Funds for the Central Universities under Grant No.20720200074。
文摘Total transmission plays an important role in efficiency improvement and wavefront control,and has made great progress in many applications,such as the optical film and signal transmission.Therefore,many traditional physical methods represented by transformation optics have been studied to achieve total transmission.However,these methods have strict limitations on the size of the photonic structure,and the calculation is complex.Here,we exploit deep learning to achieve this goal.In deep learning,the data-driven prediction and design are carried out by artificial neural networks(ANNs),which provide a convenient architecture for large dataset problems.By taking the transmission characteristic of the multi-layer stacks as an example,we demonstrate how optical materials can be designed by using ANNs.The trained network directly establishes the mapping from optical materials to transmission spectra,and enables the forward spectral prediction and inverse material design of total transmission in the given parameter space.Our work paves the way for the optical material design with special properties based on deep learning.
