Search

To explore the design of the English pronunciation quality evaluation system, a design of the English pronunciation quality evaluation system based on a deep learning model is proposed. This method explores the research of English pronunciation quality evaluation by recommending key technical problems and solutions based on information represented by the deep learning model. The research shows that the efficiency of the English pronunciation quality evaluation system based on a deep learning model is about 30% higher than that of traditional methods. Through the experimental verification, the English pronunciation quality evaluation model method is reasonable and reliable. It can give learners timely, accurate, and objective evaluation and feedback guidance, help learners find out the difference between their pronunciation and standard pronunciation, correct pronunciation errors, and improve the efficiency of English spoken language learning.

Abstract Deep learning models have been widely used in electroencephalogram (EEG) analysis and obtained excellent performance. But the adversarial attack and defense for them should be thoroughly studied before putting them into safety-sensitive use. This work exposes an important safety issue in deep-learning-based brain disease diagnostic systems by examining the vulnerability of deep learning models for diagnosing epilepsy with brain electrical activity mappings (BEAMs) to white-box attacks. It proposes two methods, Gradient Perturbations of BEAMs (GPBEAM), and Gradient Perturbations of BEAMs with Differential Evolution (GPBEAM-DE), which generate EEG adversarial samples, for the first time by perturbing BEAMs densely and sparsely respectively, and find that these BEAMs-based adversarial samples can easily mislead deep learning models. The experiments use the EEG data from CHB-MIT dataset and two types of victim models each of which has four different deep neural network (DNN) architectures. It is shown that: (1) these BEAM-based adversarial samples produced by the proposed methods in this paper are aggressive to BEAM-related victim models which use BEAMs as the input to internal DNN architectures, but unaggressive to EEG-related victim models which have raw EEG as the input to internal DNN architectures, with the top success rate of attacking BEAM-related models up to 0.8 while the top success rate of attacking EEG-related models only 0.01; (2) GPBEAM-DE outperforms GPBEAM when they are attacking the same victim model under a same distortion constraint, with the top attack success rate 0.8 for the former and 0.59 for the latter; (3) a simple modification to the GPBEAM/GPBEAM-DE will make it have aggressiveness to both BEAMs-related and EEG-related models (with top attack success rate 0.8 and 0.64), and this capacity enhancement is done without any cost of distortion increment. The goal of this study is not to attack any of EEG medical diagnostic systems, but to raise concerns about the safety of deep learning models and hope to lead to a safer design.

Abstract Water sources for irrigation systems in the Red River Delta are crucial to the socioeconomic growth of the region's communities. Human activities (discharge) have polluted the water source in recent years, and the water source from upstream is limited. Currently, the surface water quality index (WQI), which is calculated from numerous surface water quality parameters (physical, chemical, microbiological, heavy metals, etc.) is frequently used to evaluate the surface water quality of irrigation systems. However, the calculation of the WQI from water quality monitoring parameters remains constrained due to the need for a large number of monitoring parameters and the relative complexity of the calculation. To better serve the assessment of surface water quality in the study area, it is crucial and essential to conduct research to identify an efficient and accurate method of calculating the WQI. This study used machine learning and deep learning algorithms to calculate the WQI with minimal input data (water quality parameters) to reduce the cost of monitoring surface water quality. The study used the Bayes method (BMA) to select important parameters (BOD5, NH4 +, PO4 3−, turbidity, TSS, coliform, and DO). The results indicate that the machine learning model is more effective than the deep learning model, with the gradient boosting model having the most accurate prediction results because it has the highest coefficient of determination R2 (0.96). This is a solid scientific basis and an important result for the application of machine learning and deep learning algorithms to calculate WQI for the research area. The study also demonstrated the potential of artificial intelligence algorithms to improve water quality forecasting compared to traditional methods with minimal cost and time.

As smart and environmentally friendly technologies and equipment are introduced in the sea port industry, electric power consumption is expected to rapidly increase. However, there is a paucity of research on the creation of electric power management plans, specifically in relation to electric power consumption forecasting, in ports. In order to address this gap, this study forecasts future electric power consumption in Busan New Port (South Korea's largest container port) and, comparing this with the current standard electric power supply capacity, investigated the feasibility of maintaining a stable electric power supply in the future. We applied a Long Short-Term Memory (LSTM) model trained using electric power consumption and throughput data of the last 10 years to forecast the future electric power consumption of Busan New Port. According to the results, electric power consumption is expected to increase at an annual average of 4.9 % until 2040, exceeding the predicted annual 4.7 % increase in throughput during the same period. Given these results, the current standard electric power supply capacity is forecast to reach only 35 % of demand in 2040, indicating that additional electrical power supply facilities will be needed for stable port operation in the future.

Proton-exchange membrane fuel cells (PEMFCs) play a crucial role in the transition to sustainable energy systems. Accurately estimating the state of health (SOH) of PEMFCs under dynamic operating conditions is essential for ensuring their reliability and longevity. This study designed dynamic operating conditions for fuel cells and conducted durability tests using both crack-free fuel cells and fuel cells with uniform cracks. Utilizing deep learning methods, we estimated the SOH of PEMFCs under dynamic operating conditions and investigated the performance of long short-term memory networks (LSTM), gated recurrent units (GRU), temporal convolutional networks (TCN), and transformer models for SOH estimation tasks. We also explored the impact of different sampling intervals and training set proportions on the predictive performance of these models. The results indicated that shorter sampling intervals and higher training set proportions significantly improve prediction accuracy. The study also highlighted the challenges posed by the presence of cracks. Cracks cause more frequent and intense voltage fluctuations, making it more difficult for the models to accurately capture the dynamic behavior of PEMFCs, thereby increasing prediction errors. However, under crack-free conditions, due to more stable voltage output, all models showed improved predictive performance. Finally, this study underscores the effectiveness of deep learning models in estimating the SOH of PEMFCs and provides insights into optimizing sampling and training strategies to enhance prediction accuracy. The findings make a significant contribution to the development of more reliable and efficient PEMFC systems for sustainable energy applications.

Objective: This study aimed to assess the impact of artificial intelligence (AI)-driven noise reduction algorithms on metal artifacts and image quality parameters in cone-beam computed tomography (CBCT) images of the oral cavity. Materials and Methods: This retrospective study included 70 patients, 61 of whom were analyzed after excluding those with severe motion artifacts. CBCT scans, performed using a Hyperion X9 PRO 13 × 10 CBCT machine, included images with dental implants, amalgam fillings, orthodontic appliances, root canal fillings, and crowns. Images were processed with the ClariCT.AI deep learning model (DLM) for noise reduction. Objective image quality was assessed using metrics such as the differentiation between voxel values (ΔVVs), the artifact index (AIx), and the contrast-to-noise ratio (CNR). Subjective assessments were performed by two experienced readers, who rated overall image quality and artifact intensity on predefined scales. Results: Compared with native images, DLM reconstructions significantly reduced the AIx and increased the CNR (p < 0.001), indicating improved image clarity and artifact reduction. Subjective assessments also favored DLM images, with higher ratings for overall image quality and lower artifact intensity (p < 0.001). However, the ΔVV values were similar between the native and DLM images, indicating that while the DLM reduced noise, it maintained the overall density distribution. Orthodontic appliances produced the most pronounced artifacts, while implants generated the least. Conclusions: AI-based noise reduction using ClariCT.AI significantly enhances CBCT image quality by reducing noise and metal artifacts, thereby improving diagnostic accuracy and treatment planning. Further research with larger, multicenter cohorts is recommended to validate these findings.

In this study, a tropical cyclone or typhoon rainfall forecast model based on Random Forest is developed to forecast the daily rainfall at 133 weather stations in China. The input factors to the model training process include rainfall observations during 1960–2018, typhoon information (position and intensity), station information (position and altitude), and properties of the western Pacific subtropical high. Model evaluation shows that besides the distance between a station and cyclone, the subtropical high properties are ranked very high in the model’s feature importance, especially the subtropical ridgeline, and intensity. These aspects of the subtropical high influence the location and timing of typhoon landfall. The forecast model has a correlation coefficient of about 0.73, an Index of Agreement of nearly 0.8, and a mean bias of 1.28 mm based on the training dataset. Biases are consistently low, with both positive and negative signs, for target stations in the outer rainband (up to 1000 km, beyond which the model does not forecast) of typhoons. The range of biases is much larger for target stations in the inner-core (0–200 km) region. In this region, the model mostly overestimates (underestimates) the small (large) rain rates. Cases study of Typhoon Doksuri and Talim in 2023, as independent cases, shows the high performance of the model in forecasting the peak rain rates and timing of their occurrence of the two impactful typhoons.

Objectives: This study delves into the cutting-edge field of deep learning techniques, particularly deep convolutional neural networks (DCNNs), which have demonstrated unprecedented potential in assisting radiologists and orthopedic surgeons in precisely identifying meniscal tears. This research aims to evaluate the effectiveness of deep learning models in recognizing, localizing, describing, and categorizing meniscal tears in magnetic resonance images (MRIs). Materials and methods: This systematic review was rigorously conducted, strictly following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Extensive searches were conducted on MEDLINE (PubMed), Web of Science, Cochrane Library, and Google Scholar. All identified articles underwent a comprehensive risk of bias analysis. Predictive performance values were either extracted or calculated for quantitative analysis, including sensitivity and specificity. The meta-analysis was performed for all prediction models that identified the presence and location of meniscus tears. Results: This study’s findings underscore that a range of deep learning models exhibit robust performance in detecting and classifying meniscal tears, in one case surpassing the expertise of musculoskeletal radiologists. Most studies in this review concentrated on identifying tears in the medial or lateral meniscus and even precisely locating tears—whether in the anterior or posterior horn—with exceptional accuracy, as demonstrated by AUC values ranging from 0.83 to 0.94. Conclusions: Based on these findings, deep learning models have showcased significant potential in analyzing knee MR images by learning intricate details within images. They offer precise outcomes across diverse tasks, including segmenting specific anatomical structures and identifying pathological regions. Contributions: This study focused exclusively on DL models for identifying and localizing meniscus tears. It presents a meta-analysis that includes eight studies for detecting the presence of a torn meniscus and a meta-analysis of three studies with low heterogeneity that localize and classify the menisci. Another novelty is the analysis of arthroscopic surgery as ground truth. The quality of the studies was assessed against the CLAIM checklist, and the risk of bias was determined using the QUADAS-2 tool.

Anomaly detection is the process of detecting unusual or unforeseen patterns or events in data. Many factors, such as malfunctioning hardware, malevolent activities, or modifications to the data’s underlying distribution, might cause anomalies. One of the key factors in anomaly detection is balancing the trade-off between sensitivity and specificity. Balancing these trade-offs requires careful tuning of the anomaly detection algorithm and consideration of the specific domain and application. Deep learning techniques’ applications, such as LSTMs (long short-term memory algorithms), which are autoencoders for detecting an anomaly, have garnered increasing attention in recent years. The main goal of this work was to develop an anomaly detection solution for an electrical machine using an LSTM-autoencoder deep learning model. The work focused on detecting anomalies in an electrical motor’s variation vibrations in three axes: axial (X), radial (Y), and tangential (Z), which are indicative of potential faults or failures. The presented model is a combination of the two architectures; LSTM layers were added to the autoencoder in order to leverage the LSTM capacity for handling large amounts of temporal data. To prove the LSTM efficiency, we will create a regular autoencoder model using the Python programming language and the TensorFlow machine learning framework, and compare its performance with our main LSTM-based autoencoder model. The two models will be trained on the same database, and evaluated on three primary points: training time, loss function, and MSE anomalies. Based on the obtained results, it is clear that the LSTM-autoencoder shows significantly smaller loss values and MSE anomalies compared to the regular autoencoder. On the other hand, the regular autoencoder performs better than the LSTM, comparing the training time. It appears then, that the LSTM-autoencoder presents a superior performance although it was slower than the standard autoencoder due to the complexity of the added LSTM layers.

In the rapid development of digitalization, the combination of artificial intelligence technology and environmental art design is expected to gradually change the design process and promote the sustainable development of design. After specifically exploring the application of deep learning technology in the integration of modern art and environmental design, this paper utilizes the generative adversarial network under the deep learning model to design the intelligent color-matching model integrating visual aesthetics and the intelligent environmental art style generating a model based on CycleGAN, respectively and conducts the perceptual intention analysis of the environmental art design works under the intelligent generation to explore its developmental progression. The results of the eye movement experiment show that the correlation coefficients between the perceptual imagery score data of the four groups of environmental art designs are all in the range of 0.433 to 0.599, with a moderate positive correlation. Among them, the correlation coefficients between the total access time, gaze duration, gaze number, and perceptual imagery score are close to 0.6, and the significance coefficients are all less than 0.05, which indicates that the automatic generation of modern art design and environmental design proposed in this paper has a better user aesthetic preference. Introducing AI-assisted sustainable design can ensure the sustainability and effectiveness of the design solutions in practical applications, thus promoting the entire environmental art design industry to develop in a greener and more sustainable direction.

This paper proposes a deep learning method for face recognition in the self-service customs clearance system at border crossings and designs the encoder and face feature mining module in the learning framework. Meanwhile, the loss function is constructed by combining L1 loss and KL scatter. The face recognition technology based on the deep learning model is used to construct the self-service border crossing system, and the research and analysis are conducted from two aspects, namely, the test of the self-service border crossing system and the application situation. The number of outbound self-clearance acceptors has increased by 2957931, and the self-clearance system at border crossings is able to provide more travelers with the convenience brought by self-clearance. This study solves the problem of self-clearance at border crossing with the help of face recognition technology in a deep learning model, which provides technical support and theoretical reference for the optimization and upgrading of self-clearance system at border crossing in the future.

This paper first explores the revitalization of national traditional sports and cultural innovation and identifies the key to the definition of national traditional sports. Secondly, it explores the operation mechanism of sports and cultural co-innovation and proposes pathways for sports and cultural innovation. Then, the relationship between traditional ethnic sports and sports culture innovation was studied using the BERT model in the deep learning model. Finally, a controlled experiment was designed to test the comparison of students’ performance before and after learning about sports culture innovation, and the study showed that the performance was improved by 14.3506 and 13.6275, respectively, and that learning about sports culture innovation was effective in improving students’ deep learning ability. Testing the role of three different algorithms on the revitalization and cultural innovation of national traditional sports, the accuracy of the four indicators of BERT is 0.75, 0.78, 0.88 and 0.98, respectively, which is in the first place, and the accuracy of BERT is higher.

In this paper, according to the characteristics and development trend of traditional opera, based on the deep learning model, we use a convolutional neural network to extract the opera features, combined with an SVM classifier to construct a CNN-SVM classification model. For the two classification algorithm models of logistic regression and deep confidence network, combined with two types of feature parameters of timbre and melody, six groups of experiments are designed to extract the time-frequency features of traditional opera. The CNN-SVM classification model is used to categorize the emotion of traditional opera, which aims to divide the musical features by multi-feature selection. Analyze the timbre feature parameter MFCC to investigate the impact of traditional opera timbre on ethnic vocal singing. For the logistic regression model, the coefficient of MFCCs is 0.5588, and the classification accuracy is only 0.5301 when the feature parameters are selected as melodic features, i.e., gene frequency, resonance peak, and band energy, and 0.6228 when the feature parameters are selected as a combination of timbral and melodic features. The diversity of the traditional opera timbres contributes to the development of ethnic vocal art with the trend of inclusiveness.

In this paper, the design process of cultural and creative product design is first constructed in the context of deep learning and the categories of cultural idea factors are planned. The training of the network for the image extraction classification task is accelerated due to the deep learning model’s fitting ability. Secondly, the 3D virtual intelligence algorithm is applied to use the point cloud as the basis for reconstruction in order to generate a highly reproducible object surface model. Finally, the K-means algorithm is employed to obtain the final extracted color, which constitutes the technical foundation of the product design system. The final results show that the color matching degree reaches 8.91, while the image design level reaches 9.24, which illustrates the effectiveness of the tourism cultural and creative product design system under deep learning.

Wind energy has become an essential part of the energy power source of current power systems since it is eco-friendly and sustainable. To optimize the operations of wind farms with the constraint of satisfying the power demand, it is critical to provide accurate predictions of wind power generated in the future. Although deep learning models have greatly improved prediction accuracy, the overfitting issue limits the application of deep learning models trained under one condition to another. A huge number of data are required to train a new deep learning model for another environment, which is sometimes not practical in some urgent situations with only very little training data on wind power. In this paper, we propose a novel learning method, named meta-reservoir computing (MRC), to address the above issue, combining reservoir computing and meta-learning. The reservoir computing method improves the computational efficiency of training a deep neural network for time series data. On the other hand, meta-learning is used to improve the initial point and other hyperparameters of reservoir computing. The proposed MRC method is validated using an experimental dataset of wind power compared with the traditional training method. The results show that the MRC method can obtain an accurate predictive model of wind power with only a few shots of data.

In order to scientifically allocate rural landscape resources, reasonably plan rural tourism space, and ensure that the local characteristics of the countryside are not homogenized when carrying out rural landscape design, this paper studies rural landscape design strategies based on deep learning models. The extreme learning machine algorithm, DBN-RBM algorithm model and the improved DBN-DELM algorithm are the main technical means to obtain research data and parameter calibration results for tourism planning and development work, and the rural planning direction and planning theme is determined through the rural landscape design pre-analysis work. The data show that the main motives of tourists’ rural experience tourism are close to nature 85.90% and leisure vacation 75%, followed by understanding culture 45.30%, novelty 30.70%, parent-child education 29.20%, health retreat 30.40%, and business meeting 5.90%. In this paper, the study of rural landscape planning and design can effectively alleviate the contradiction between people’s production and living and ecological environment and coordinate the benign development of rural and tourism elements in their respective spaces.

To verify the feasibility of robust speech recognition based on deep learning in sports game review. In this paper, a robust speech recognition model is built based on the generative adversarial network GAN algorithm according to the deep learning model. And the loss function, optimization function and noise reduction front-end are introduced in the model to achieve the optimization of speech extraction features through denoising process to ensure that accurate speech review data can be derived even in the game scene under noisy environment. Finally, the experiments are conducted to verify the four directions of the model algorithm by comparing the speech features MFCC, FBANK and WAVE. The experimental results show that the speech recognition model trained by the GSDNet model algorithm can reach 89% accuracy, 56.24% reduction of auxiliary speech recognition word error rate, 92.61% accuracy of speech feature extraction, about 62.19% reduction of training sample data volume, and 94.75% improvement of speech recognition performance in the speech recognition task under noisy environment. It shows that the robust speech recognition based on deep learning can be applied to sports game reviews, and also can provide accurate voice review information from the noisy sports game scene, and also broaden the application area for deep learning models.

In order to have a better product display and thus attract consumers’ purchases and increase the economic benefits of the enterprise, in this paper, we propose a deep learning model for brand 3D image design. A feedforward neural network that estimates the error of previous layers based on the error of the output layer assigns the convolutional kernel weight parameters of the network in the interval and stops when the error reaches a preset accuracy or reaches a preset maximum learning count. The locally-aware convolutional neural network acquires local features that are finer than the global features and outputs the feature maps of the convolutional layers after passing the activation function to calculate the sensitivity of the sampled layer units. Given the sensitivity information of the feature map, the gradient of the kernel function weights is obtained, and the updated parameters are trained to achieve feature map recursion and solve the image boundary problem. A 3D recurrent neural network is constructed using data-driven multiple or single images, transformed into a low-dimensional feature matrix, processed with 3D pixel data, extracted perceptual features, and generated high-resolution images. The analysis of the results shows that the CD value of the used model is 0.477 and the EMD value is 0.579, which makes the constructed 3D images with more obvious detail levels and more accurate structural design, while the model of Pixel2Mesh focuses more on surface information, so the generated model is more realistic and closer to the real image.

To smoothly realize the information conversion from the original language information to the target language, this paper constructs a deep learning-based fuzzy translation model for news reports so that the translated text can faithfully convey the meaning of the original language text information and achieve natural semantic equivalence. A neural probabilistic language model is used to construct objective functions in speech recognition and lexical annotation so that the translated text can provide a more appropriate linguistic representation of the polysemantic words in the original language text according to the differences in contextual morphology. A deep learning occurrence mechanism model is constructed through fuzzy semantic reasoning and fuzzy translation logic, and learning state indicators such as emotional interaction are designed to evaluate the occurrence status of fuzzy translation accurately. The simulation results show that the natural language processing (GLUE) test score of the deep learning-based fuzzy translation model for news reports is 89.8, 9.2, and 6.9 points higher than 80.6 and 82.9 for the other two models, respectively. The average error discrimination ability of the model designed in this paper is 93.57, and the average training set, development set, and test set values are 98.425, 10.16, and 45.95, respectively. Thus, it can be seen that the deep learning-based fuzzy translation model for news reports can more naturally and accurately respond to the dynamic changes in language, which promotes the rapid development of translation theory and practice.

Foreign language learning anxiety is prevalent in students’ foreign language learning process. To better alleviate students’ anxiety situation, this paper constructs a self-coding training model for anxiety regulation strategies based on recurrent neural networks in deep learning models. Anxiety-related factors affecting English learning are input into the recurrent neural network as the input layer. The data are corrected by balancing the data in the input layer through encoding and decoding in the implicit layer. The corrected data is reconstructed and transformed as the input layer of the next level of the recurrent neural network. The above steps were repeated continuously for layer-by-layer training until the same output layer parameters as the pre-trained model were reached. This resulted in a learning anxiety regulation strategy of changing the student learning environment and self-regulation. To verify that the above strategies can reduce the anxiety value, a simulation was conducted, and the results showed that the number of superior students in the multimedia environment was 2% higher than that in the traditional teaching model. Students’ anxiety was reduced from 18% to 7% after active and effective self-regulation. From these results, it is clear that the anxiety regulation strategy derived from the deep learning model is feasible and ensures the healthy development of students.

The development of art education in colleges and universities is a direct reflection of the soft power of culture. In this paper, based on two models, BP neural network and long and short-term memory neural network in deep learning model, we study the factors affecting the development of college art education. Firstly, the article introduces the basic models of the BP neural network and LSTM neural network and selects the indicators under three dimensions of college art education scale, college art education input, and output to visually analyze the development status of college art education and find the possible problems. Secondly, five dimensions, namely, population scale, policy support, economic strength, industrial structure, and faculty level, were selected as the influential indicators to study the level of art education in colleges and universities. The population size passed the 1% significance test, GDP per capita and faculty strength passed the 5% significance test, the influence coefficient of art education expenses was 0.349, and only the industrial structure failed the test. According to the analysis, we conclude that the factors affecting the development of art education in colleges and universities are mainly four aspects: population size, policy support, economic strength, and faculty level.

The development of international Chinese language education greatly impacts the building of a global Chinese cultural environment. This paper constructs a model structure based on a convolutional neural network with a deep learning model and explains it in detail, and uses it to analyze the current development situation of international Chinese language education. The paper also presents the performance evaluation of the convolutional neural network model and discusses the problems in the development of international Chinese language education and the innovative development direction of online teaching. From the characteristics of innovative online teaching materials, rich media, mobility, interactivity, personalization, timeliness, and open sharing become the main melodies of the development of international Chinese language education, and the five-year average values of each characteristic are 53.59%, 51.17%, 49.77%, 47.84%, 45.94%, and 42.19%, respectively. The convolutional neural network model based on deep learning can effectively analyze the problems of international Chinese language education and the direction of innovation, providing an effective technology to help the development of Chinese culture in China.

Under the digitalization of big data information, how to use big data technology to combine with visual communication design teaching has become a key topic of concern in the current education sector. Firstly, this paper proposes using a deep learning model based on big data technology for teaching practice research of visual communication design majors in colleges and universities and constructs a deep learning model by combining perceptron and bp neural algorithm for data mining property and accurate prediction. Secondly, the deep learning model is optimized by using the activation function and knowledge distillation function in response to the limitation problem of the deep learning model. Then the teaching evaluation data of visual communication design majors in colleges and universities are obtained through questionnaire survey work, and evaluation indexes are determined according to the evaluation data. Finally, students of visual design majors at Guizhou Vocational and Technical College were selected as research samples, and the practice performance of communication design majors was analyzed based on the deep learning model. The results showed that the deep learning model: from week 1 to week 20, the score improved from 52 to 88, which improved the score by 69.23%. The traditional method: from week 1 to week 20, the score improved from 52 to 80, improving the score by 53.84%. This study is conducive to promoting the healthy development of the visual communication profession and cultivating design talents that meet the needs of the information age and is of great significance to the development of the visual communication profession in China.

Exploring the influence of good Chinese storytelling on the foreign communication of local culture in Shaanxi is to develop local culture better and expand the influence of the local cultural industry in Shaanxi. In this paper, starting from deep learning neural networks, the attention mechanism is introduced by using convolutional neural networks, and the attention mechanism-based detection model for cultural information dissemination is constructed by vectorizing the data of the pre-trained model. The performance of the CNN-ATT model constructed in this paper is experimentally analyzed on two data sets. The accuracy of this paper’s model on the Ma_Dataset and CED_Dataset datasets is 96.23% and 95.21%, respectively. The best results are obtained when the convolution size is 3,4,5, and the accuracy is improved by about 4.03% compared with the convolution size of 4,5. This shows that the model of this paper can effectively analyze the influence of cultural information dissemination and also provides a research basis for the detection of the influence of telling Chinese stories on expanding the foreign communication of local culture in Shaanxi.

In this paper, we apply a deep learning model to discriminate sentiment in an interactive model of online college English education and propose a fusion model that splices convolutional neural networks and bidirectional long- and short-term memory neural networks horizontally. Convolutional neural networks are good at capturing the sentiment feature vectors using multi-channel convolutional kernels but are unable to extract the sentiment information above and below the sentiment sequence. The short and long-term memory neural network is able to extract the sentiment feature vectors by using recurrent neural networks, which can better compensate for the shortcomings of the convolutional neural networks. The online teaching of college English is selected as the object of analysis, and the improved Flanders interaction analysis system is used to study the online interaction process of college English so as to propose suggestions for the interaction of online teaching of college English. Then the performance of the model is analyzed through simulation experiments. Compared with the traditional TextCNN and BiLSTM, the CNN −BiLSTMATT sentiment analysis model has an accuracy of 0.8611, precision of 0.8471, recall of 0.8691, and F1 of 0.8562, so the CNN − BiLSTMATT sentiment analysis model is more suitable for college English online interaction. This study overcomes the disadvantages of online interaction and continuously improves the efficiency of online teaching interaction.

This paper analyzes the factors that affect the mental health of college students, and the focus of this analysis is on emotional-emotional factors. The extroverted presentation of emotional affects is used as visual information to study the mental health status of college students. Based on the advantages of long and short memory neural networks based on deep learning models in processing two-dimensional images, a computer vision task is used to perform visual recognition, target detection, and expression image classification of college students’ facial expressions. The use of video facial expression recognition with multi-feature fusion is utilized to effectively identify the facial expression machine of college students in both laboratory-controlled and outdoor environments. The mental health status of college students was analyzed in terms of facial expression recognition and feature extraction. The recognition rate for general features was 80.3%, 89.3% for six specific facial emotions, and 84.4% for LBP features.

This paper aims to analyze and design the system’s functions by student work management requirements, focusing on the business and user modules. The security testing algorithm based on the deep learning model verifies the system’s operation, analyzes the adversarial sample attack in image and text scenarios, and uses a black-box-white-box attack algorithm and defense algorithm for system security testing. Perform environment testing, compatibility testing, data response performance testing, and instance model evaluation for implementing the student work management system. Less than 500ms is the response time, over 400 requests are processed per second, and the timeout response rate is below 5%. The security evaluation coefficient was 66.982 for Example Model I and 74.628 for Example Model II, showing the system has good loadability and security.

In this paper, we propose a deep learning restoration method for fine art painting images based on a contextual encoder and select a convolutional self-encoder-based image restoration model for workflow analysis, including forward inference and backward network propagation. We design and develop a software system for predicting the grades of college art and painting students and apply the neural network prediction model to this system to realize the grade prediction of college art students. The accuracy of the GDPN model is 0.922, the precision is 0.904, and the recall is 0.966, which can consider both the temporal and overall information in the click behavior and achieve a better prediction effect.

AbstractIn the precision machining industry, machine tools are usually affected by various factors during machining, and various machining errors generated accordingly. Where thermal error is one of the most common and difficult to control factors for machine tools. Therefore, in this study, six temperature sensors and an eddy current displacement meter are provided in a machine tool with 4-axis for dataset collection required for the model training, then data are organized and normalized. Next, data are introduced into a variety of learning models and validated by [Formula: see text]-Fold cross-validation for predicting those nonlinear factors that affect the errors. At the end, predicted results are summarized and compared to find out the best two model with better predictive performance for pre-trained model of transfer learning. It observes better predicted results from a retraining conducted through applying Multilayer Perceptron (MLP) on these two pre-trained models, wherein MAE value as 0.40, RMSE as 0.52625 and [Formula: see text] score as 0.99696 respectively.

The Hajj is a religious event that attracts a significant number of Muslims from various countries who perform rituals in Mecca, Saudi Arabia. Despite the high volume of pilgrims that typically participate in the event, the number has been reduced in recent years due to the COVID-19 pandemic. The satisfaction of Hajj pilgrims with the quality of hospitality services provided during the event is a crucial factor that must be studied and understood. To achieve this goal, various psychological theories have been employed to explain the phenomenon. The advancement of big data and artificial intelligence has enabled the development of new analytical methodologies for evaluating psychological theories in the hospitality industry. In this study, we present a novel deep learning model that leverages the expectation-confirmation theory to examine the satisfaction of Hajj pilgrims with hospitality services. The model was trained and tested on data obtained from hotel review posts related to the Hajj. Based on our results, the proposed model achieved a high accuracy of 97 % in predicting the satisfaction of Hajj pilgrims. In addition, the results can be used to improve the quality of services provided to pilgrims and enhance their overall experience during the Hajj.

This paper presents a transfer and deep learning based approach to the classification of Sickle Cell Disease (SCD). Five transfer learning models such as ResNet-50, AlexNet, MobileNet, VGG-16 and VGG-19, and a sequential convolutional neural network (CNN) have been implemented for SCD classification. ErythrocytesIDB dataset has been used for training and testing the models. In order to make up for the data insufficiency of the erythrocytesIDB dataset, advanced image augmentation techniques are employed to ensure the robustness of the dataset, enhance dataset diversity and improve the accuracy of the models. An ablation experiment using Random Forest and Support Vector Machine (SVM) classifiers along with various hyperparameter tweaking was carried out to determine the contribution of different model elements on their predicted accuracy. A rigorous statistical analysis was carried out for evaluation and to further evaluate the model's robustness, an adversarial attack test was conducted. The experimental results demonstrate compelling performance across all models. After performing the statistical tests, it was observed that MobileNet showed a significant improvement (p = 0.0229), while other models (ResNet-50, AlexNet, VGG-16, VGG-19) did not (p > 0.05). Notably, the ResNet-50 model achieves remarkable precision, recall, and F1-score values of 100 % for circular, elongated, and other cell shapes when experimented with a smaller dataset. The AlexNet model achieves a balanced precision (98 %) and recall (99 %) for circular and elongated shapes. Meanwhile, the other models showcase competitive performance.

Background: Most of superficial soft-tissue masses are benign tumors, and very few are malignant tumors. However, persistent growth, of both benign and malignant tumors, can be painful and even life-threatening. It is necessary to improve the differential diagnosis performance for superficial soft-tissue masses by using deep learning models. This study aimed to propose a new ultrasonic deep learning model (DLM) system for the differential diagnosis of superficial soft-tissue masses. Methods: Between January 2015 and December 2022, data for 1615 patients with superficial soft-tissue masses were retrospectively collected. Two experienced radiologists (radiologists 1 and 2 with 8 and 30 years' experience, respectively) analyzed the ultrasound images of each superficial soft-tissue mass and made a diagnosis of malignant mass or one of the five most common benign masses. After referring to the DLM results, they re-evaluated the diagnoses. The diagnostic performance and concerns of the radiologists were analyzed before and after referring to the results of the DLM results. Results: In the validation cohort, DLM-1 was trained to distinguish between benign and malignant masses, with an AUC of 0.992 (95% CI: 0.980, 1.0) and an ACC of 0.987 (95% CI: 0.968, 1.0). DLM-2 was trained to classify the five most common benign masses (lipomyoma, hemangioma, neurinoma, epidermal cyst, and calcifying epithelioma) with AUCs of 0.986, 0.993, 0.944, 0.973, and 0.903, respectively. In addition, under the condition of the DLM-assisted diagnosis, the radiologists greatly improved their accuracy of differential diagnosis between benign and malignant tumors. Conclusions: The proposed DLM system has high clinical application value in the differential diagnosis of superficial soft-tissue masses. [ABSTRACT FROM AUTHOR]

Intelligent transportation systems need to realize accurate traffic congestion prediction. The spatio-temporal features of traffic flow are essential to analyze and predict congestion. Our study proposes a data-driven model to predict the traffic congested flow. Firstly, the traffic zone/grid method is used to store the local area roads' average speed of the vehicles. Second, the discrete snapshot set is proposed to characterize traffic flow's spatial and temporal features over a continuous period. Third, the evolution of traffic congested flow in various time dimensions (weekly days, weekend days, and one week) is examined by transforming the global urban transportation network into traffic zones. Finally, the data-driven model is constructed to predict urban road traffic congestion by using the extracted spatio-temporal characteristics of traffic zones' traffic flow, the snapshot set of which serves as inputs for this model. The model adopts the convolutional LSTM network to learn the temporal and local spatial features of traffic flow, while utilizing a convolutional neural network to effectively capture the global spatial features inherent in traffic flow. The numerical experiments are conducted on two cities' transportation networks, and the results demonstrate that the performance of the proposed model outperforms traditional traffic flow prediction models. [ABSTRACT FROM AUTHOR]

Object classification based on tactile perception plays an essential role in robot manipulation process, as it serves for decision-making for the the downstream manipulation tasks. The demand for precise execution by industrial robots in smart factories has increased, and like humans, robots can infer tactile properties and identify object categories through brief motions. However, traditional practices only consider grasping as an instant state, resulting in the absence of time-series information. To address this issue, we propose a spatio-temporal attention-based Long Short-Term Memory (LSTM) network to solve the time-series problem for object classification. The proposed model utilizes a temporal attention mechanism that can dynamically trace the time-related features of the tactile data. Moreover, a spatial attention mechanism coordinates the integration of tactile information from various input features. The model classifies objects based on the entire temporal process of robot-object contact rather than data from a particular moment. To further enhance the model's performance, we also incorporate PCA and Kalman filter. Our extensive experiments demonstrate the proposed model's accuracy and efficiency, validating its ability to perform object classification based on tactile perception. [ABSTRACT FROM AUTHOR]

The accurate analysis of multi-scale flame development plays a crucial role in improving firefighting decisions and facilitating smart city establishment. However, flames' non-rigid nature and blurred edges present challenges in achieving accurate segmentation. Consequently, little attention is paid to extracting further flame situation information through fire segmentation. To address this issue, we propose Flame-SeaFormer, a multi-scale flame situation detection model based on the pixel-level segmentation of visual images. Flame-SeaFormer comprises three key steps. Firstly, in the context branch, squeeze-enhanced axial attention (SEA attention) is applied to squeeze fire feature maps, capturing dependencies among flame pixels while reducing the computational complexity. Secondly, the fusion block in the spatial branch integrates high-level semantic information from the contextual branch with low-level spatial details, ensuring a global representation of flame features. Lastly, the light segmentation head conducts pixel-level segmentation on the flame features. Based on the flame segmentation results, static flame parameters (flame height, width, and area) and dynamic flame parameters (change rates of flame height, width, and area) are gained, thereby enabling the real-time perception of flame evolution behavior. Experimental results on two datasets demonstrate that Flame-SeaFormer achieves the best trade-off between segmentation accuracy and speed, surpassing existing fire segmentation methods. Flame-SeaFormer enables precise flame state acquisition and evolution exploration, supporting intelligent fire protection systems in urban environments. [ABSTRACT FROM AUTHOR]

Background: Due to the increased interest in cryptocurrencies, opinions on cryptocurrency-related topics are shared on news and social media. The enormous amount of sentiment data that is frequently released makes data processing and analytics on such important issues more challenging. In addition, the present sentiment models in the cryptocurrency domain are primarily focused on English with minimal work on Malay language, further complicating problems. Objective: The performance of the sentiment regression model to forecast sentiment scores for Malay news and tweets is examined in this study. Methods: Malay news headlines and tweets on Bitcoin and Ethereum are used as the input. A hybrid Generalized Autoregressive Pretraining for Language Understanding (XLNet) language model in combination with Bidirectional-Gated Recurrent Unit (Bi-GRU) deep learning model is applied in the proposed sentiment regression implementation. The effectiveness of the proposed sentiment regression model is also investigated using the multi-head self-attention mechanism. Then, a comparison analysis using Bidirectional Encoder Representations from Transformers (BERT) is carried out. Results: The experimental results demonstrate that the number of attention heads is vital in improving the XLNet-GRU sentiment model performance. There are slight improvements of 0.03 in the adjusted R2 values with an average MAE of 0.163 (Malay news) and 0.174 (Malay tweets). In addition, an average RMSE of 0.267 and 0.255 were obtained respectively for Malay news and tweets, which show that the proposed XLNet-GRU sentiment model outperforms the BERT sentiment model with lesser prediction errors. Conclusion: The proposed model contributes to predicting sentiment on cryptocurrency. Moreover, this study also introduced two carefully curated Malay corpora, CryptoSentiNews-Malay and CryptoSentiTweets-Malay, which are extracted from news and tweets, respectively. Further works to enhance Malay news and tweets corpora on cryptocurrency-related issues will be expended with implementing the proposed XLNet Bi-GRU deep learning model for greater financial insight. [ABSTRACT FROM AUTHOR]

Objective: To develop a deep learning (DL) model for predicting axillary lymph node (ALN) metastasis using dynamic ultrasound (US) videos in breast cancer patients. Methods: A total of 271 US videos from 271 early breast cancer patients collected from Xiang'an Hospital of Xiamen University andShantou Central Hospitabetween September 2019 and June 2021 were used as the training, validation, and internal testing set (testing set A). Additionally, an independent dataset of 49 US videos from 49 patients with breast cancer, collected from Shanghai 10th Hospital of Tongji University from July 2021 to May 2022, was used as an external testing set (testing set B). All ALN metastases were confirmed using pathological examination. Three different convolutional neural networks (CNNs) with R2 + 1D, TIN, and ResNet-3D architectures were used to build the models. The performance of the US video DL models was compared with that of US static image DL models and axillary US examination performed by ultrasonographers. The performances of the DL models and ultra-sonographers were evaluated based on accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). Additionally, gradient class activation mapping (Grad-CAM) technology was also used to enhance the interpretability of the models. Results: Among the three US video DL models, TIN showed the best performance, achieving an AUC of 0.914 (95% CI: 0.843-0.985) in predicting ALN metastasis in testing set A. The model achieved an accuracy of 85.25% (52/61), with a sensitivity of 76.19% (16/21) and a specificity of 90.00% (36/40). The AUC of the US video DL model was superior to that of the US static image DL model (0.856, 95% CI: 0.753-0.959, P<0.05). The Grad-CAM technology confirmed the heatmap of the model, which highlighted important subregions of the keyframe for ultra-sonographers' review. Conclusion: A feasible and improved DL model to predict ALN metastasis from breast cancer US video images was developed. The DL model in this study with reliable interpretability would provide an early diagnostic strategy for the appropriate management of axillary in the early breast cancer patients. [ABSTRACT FROM AUTHOR]

Passenger flow anomaly detection in urban rail transit networks (URTNs) is critical in managing surging demand and informing effective operations planning and controls in the network. Existing studies have primarily focused on identifying the source of anomalies at a single station by analysing the time-series characteristics of passenger flow. However, they ignored the high-dimensional and complex spatial features of passenger flow and the dynamic behaviours of passengers in URTNs during anomaly detection. This article proposes a novel anomaly detection methodology based on a deep learning framework consisting of a graph convolution network (GCN)–informer model and a Gaussian naive Bayes model. The GCN–informer model is used to capture the spatial and temporal features of inbound and outbound passenger flows, and it is trained on normal datasets. The Gaussian naive Bayes model is used to construct a binary classifier for anomaly detection, and its parameters are estimated by feeding the normal and abnormal test data into the trained GCN–informer model. Experiments are conducted on a real-world URTN passenger flow dataset from Beijing. The results show that the proposed framework has superior performance compared to existing anomaly detection algorithms in detecting network-level passenger flow anomalies. This article is part of the theme issue 'Artificial intelligence in failure analysis of transportation infrastructure and materials'. [ABSTRACT FROM AUTHOR]