Search results

The purpose of this study is to investigate and demonstrate the advancements achieved in the field of chest X-ray image captioning through the utilization of dynamic convolutional encoder–decoder networks (DyCNN). Typical convolutional neural networks (CNNs) are unable to capture both local and global contextual information effectively and apply a uniform operation to all pixels in an image. To address this, we propose an innovative approach that integrates a dynamic convolution operation at the encoder stage, improving image encoding quality and disease detection. In addition, a decoder based on the gated recurrent unit (GRU) is used for language modeling, and an attention network is incorporated to enhance consistency. This novel combination allows for improved feature extraction, mimicking the expertise of radiologists by selectively focusing on important areas and producing coherent captions with valuable clinical information.

In this study, we have presented a new report generation approach that utilizes dynamic convolution applied Resnet-101 (DyCNN) as an encoder (Verelst and Tuytelaars, 2019) and GRU as a decoder (Dey and Salemt, 2017; Pan et al., 2020), along with an attention network (see Figure 1). This integration innovatively extends the capabilities of image encoding and sequential caption generation, representing a shift from conventional CNN architectures. With its ability to dynamically adapt receptive fields, the DyCNN excels at capturing features of varying scales within the CXR images. This dynamic adaptability significantly enhances the granularity of feature extraction, enabling precise representation of localized abnormalities and structural intricacies. By incorporating this flexibility into the encoding process, our model can distil meaningful and contextually rich features from the radiographic data. While the attention mechanism enables the model to selectively focus on different regions of the image during caption generation. The attention mechanism enhances the report generation process by allowing the model to assign different importance weights to different regions of the image, mimicking human perception. In parallel, the GRU-based decoder adds a critical dimension to the process by ensuring a smooth, sequential generation of captions.

The findings of this study highlight the significant advancements achieved in chest X-ray image captioning through the utilization of dynamic convolutional encoder–decoder networks (DyCNN). Experiments conducted using the IU-Chest X-ray datasets showed that the proposed model outperformed other state-of-the-art approaches. The model achieved notable scores, including a BLEU_1 score of 0.591, a BLEU_2 score of 0.347, a BLEU_3 score of 0.277 and a BLEU_4 score of 0.155. These results highlight the efficiency and efficacy of the model in producing precise radiology reports, enhancing image interpretation and clinical decision-making.

This work is the first of its kind, which employs DyCNN as an encoder to extract features from CXR images. In addition, GRU as the decoder for language modeling was utilized and the attention mechanisms into the model architecture were incorporated.

As data centres grow in size and complexity, traditional air-cooling methods are becoming less effective and more expensive. Immersion cooling, where servers are submerged in a dielectric fluid, has emerged as a promising alternative. Ensuring reliable operations in data centre applications requires the development of an effective control framework for immersion cooling systems, which necessitates the prediction of server temperature. While deep learning-based temperature prediction models have shown effectiveness, further enhancement is needed to improve their prediction accuracy. This study aims to develop a temperature prediction model using Long Short-Term Memory (LSTM) Networks based on recursive encoder-decoder architecture.

This paper explores the use of deep learning algorithms to predict the temperature of a heater in a two-phase immersion-cooled system using NOVEC 7100. The performance of recursive-long short-term memory-encoder-decoder (R-LSTM-ED), recursive-convolutional neural network-LSTM (R-CNN-LSTM) and R-LSTM approaches are compared using mean absolute error, root mean square error, mean absolute percentage error and coefficient of determination (R²) as performance metrics. The impact of window size, sampling period and noise within training data on the performance of the model is investigated.

The R-LSTM-ED consistently outperforms the R-LSTM model by 6%, 15.8% and 12.5%, and R-CNN-LSTM model by 4%, 11% and 12.3% in all forecast ranges of 10, 30 and 60 s, respectively, averaged across all the workloads considered in the study. The optimum sampling period based on the study is found to be 2 s and the window size to be 60 s. The performance of the model deteriorates significantly as the noise level reaches 10%.

The proposed models are currently trained on data collected from an experimental setup simulating data centre loads. Future research should seek to extend the applicability of the models by incorporating time series data from immersion-cooled servers.

The proposed multivariate-recursive-prediction models are trained and tested by using real Data Centre workload traces applied to the immersion-cooled system developed in the laboratory.

Crack detection of pavement is a critical task in the periodic survey. Efficient, effective and consistent tracking of the road conditions by identifying and locating crack contributes to establishing an appropriate road maintenance and repair strategy from the promptly informed managers but still remaining a significant challenge. This research seeks to propose practical solutions for targeting the automatic crack detection from images with efficient productivity and cost-effectiveness, thereby improving the pavement performance.

This research applies a novel deep learning method named TransUnet for crack detection, which is structured based on Transformer, combined with convolutional neural networks as encoder by leveraging a global self-attention mechanism to better extract features for enhancing automatic identification. Afterward, the detected cracks are used to quantify morphological features from five indicators, such as length, mean width, maximum width, area and ratio. Those analyses can provide valuable information for engineers to assess the pavement condition with efficient productivity.

In the training process, the TransUnet is fed by a crack dataset generated by the data augmentation with a resolution of 224 × 224 pixels. Subsequently, a test set containing 80 new images is used for crack detection task based on the best selected TransUnet with a learning rate of 0.01 and a batch size of 1, achieving an accuracy of 0.8927, a precision of 0.8813, a recall of 0.8904, an F1-measure and dice of 0.8813, and a Mean Intersection over Union of 0.8082, respectively. Comparisons with several state-of-the-art methods indicate that the developed approach in this research outperforms with greater efficiency and higher reliability.

The developed approach combines TransUnet with an integrated quantification algorithm for crack detection and quantification, performing excellently in terms of comparisons and evaluation metrics, which can provide solutions with potentially serving as the basis for an automated, cost-effective pavement condition assessment scheme.

Genre classification of songs based on lyrics is a challenging task even for humans, however, state-of-the-art natural language processing has recently offered advanced solutions to this task. The purpose of this study is to advance the understanding and application of natural language processing and deep learning in the domain of music genre classification, while also contributing to the broader themes of global knowledge and communication, and sustainable preservation of cultural heritage.

The main contribution of this study is the development and evaluation of various machine and deep learning models for song genre classification. Additionally, we investigated the effect of different word embeddings, including Global Vectors for Word Representation (GloVe) and Word2Vec, on the classification performance. The tested models range from benchmarks such as logistic regression, support vector machine and random forest, to more complex neural network architectures and transformer-based models, such as recurrent neural network, long short-term memory, bidirectional long short-term memory and bidirectional encoder representations from transformers (BERT).

The authors conducted experiments on both English and multilingual data sets for genre classification. The results show that the BERT model achieved the best accuracy on the English data set, whereas cross-lingual language model pretraining based on RoBERTa (XLM-RoBERTa) performed the best on the multilingual data set. This study found that songs in the metal genre were the most accurately labeled, as their text style and topics were the most distinct from other genres. On the contrary, songs from the pop and rock genres were more challenging to differentiate. This study also compared the impact of different word embeddings on the classification task and found that models with GloVe word embeddings outperformed Word2Vec and the learning embedding layer.

This study presents the implementation, testing and comparison of various machine and deep learning models for genre classification. The results demonstrate that transformer models, including BERT, robustly optimized BERT pretraining approach, distilled bidirectional encoder representations from transformers, bidirectional and auto-regressive transformers and XLM-RoBERTa, outperformed other models.

To improve the speed and accuracy of turbine blade film cooling design process, the most advanced deep learning models were introduced into this study to investigate the most suitable define for prediction work. This paper aims to create a generative surrogate model that can be applied on multi-objective optimization problems.

The latest backbone in the field of computer vision (Swin-Transformer, 2021) was introduced and improved as the surrogate function for prediction of the multi-physics field distribution (film cooling effectiveness, pressure, density and velocity). The basic samples were generated by Latin hypercube sampling method and the numerical method adopt for the calculation was validated experimentally at first. The training and testing samples were calculated at experimental conditions. At last, the surrogate model predicted results were verified by experiment in a linear cascade.

The results indicated that comparing with the Multi-Scale Pix2Pix Model, the Swin-Transformer U-Net model presented higher accuracy and computing speed on the prediction of contour results. The computation time for each step of the Swin-Transformer U-Net model is one-third of the original model, especially in the case of multi-physics field prediction. The correlation index reached more than 99.2% and the first-order error was lower than 0.3% for multi-physics field. The predictions of the data-driven surrogate model are consistent with the predictions of the computational fluid dynamics results, and both are very close to the experimental results. The application of the Swin-Transformer model on enlarging the different structure samples will reduce the cost of numerical calculations as well as experiments.

The number of U-Net layers and sample scales has a proper relationship according to equation (8). Too many layers of U-Net will lead to unnecessary nonlinear variation, whereas too few layers will lead to insufficient feature extraction. In the case of Swin-Transformer U-Net model, incorrect number of U-Net layer will reduce the prediction accuracy. The multi-scale Pix2Pix model owns higher accuracy in predicting a single physical field, but the calculation speed is too slow. The Swin-Transformer model is fast in prediction and training (nearly three times faster than multi Pix2Pix model), but the predicted contours have more noise. The neural network predicted results and numerical calculations are consistent with the experimental distribution.

This paper creates a generative surrogate model that can be applied on multi-objective optimization problems. The generative adversarial networks using new backbone is chosen to adjust the output from single contour to multi-physics fields, which will generate more results simultaneously than traditional surrogate models and reduce the time-cost. And it is more applicable to multi-objective spatial optimization algorithms. The Swin-Transformer surrogate model is three times faster to computation speed than the Multi Pix2Pix model. In the prediction results of multi-physics fields, the prediction results of the Swin-Transformer model are more accurate.

The research on structure function recognition mainly concentrates on identifying a specific part of academic literature and its applicability in the multidiscipline perspective. A specific part of academic literature, such as sentences, paragraphs and chapter contents are also called a level of academic literature in this paper. There are a few comparative research works on the relationship between models, disciplines and levels in the process of structure function recognition. In view of this, comparative research on structure function recognition based on deep learning has been conducted in this paper.

An experimental corpus, including the academic literature of traditional Chinese medicine, library and information science, computer science, environmental science and phytology, was constructed. Meanwhile, deep learning models such as convolutional neural networks (CNN), long and short-term memory (LSTM) and bidirectional encoder representation from transformers (BERT) were used. The comparative experiments of structure function recognition were conducted with the help of the deep learning models from the multilevel perspective.

The experimental results showed that (1) the BERT model performed best, with F1 values of 78.02, 89.41 and 94.88%, respectively at the level of sentence, paragraph and chapter content. (2) The deep learning models performed better on the academic literature of traditional Chinese medicine than on other disciplines in most cases, e.g. F1 values of CNN, LSTM and BERT, respectively arrived at 71.14, 69.96 and 78.02% at the level of sentence. (3) The deep learning models performed better at the level of chapter content than other levels, the maximum F1 values of CNN, LSTM and BERT at 91.92, 74.90 and 94.88%, respectively. Furthermore, the confusion matrix of recognition results on the academic literature was introduced to find out the reason for misrecognition.

This paper may inspire other research on structure function recognition, and provide a valuable reference for the analysis of influencing factors.

Light field images (LFIs) have gained popularity as a technology to increase the field of view (FoV) of plenoptic cameras since they can capture information about light rays with a large FoV. Wide FoV causes light field (LF) data to increase rapidly, which restricts the use of LF imaging in image processing, visual analysis and user interface. Effective LFI coding methods become of paramount importance. This paper aims to eliminate more redundancy by exploring sparsity and correlation in the angular domain of LFIs, as well as mitigate the loss of perceptual quality of LFIs caused by encoding.

This work proposes a new efficient LF coding framework. On the coding side, a new sampling scheme and a hierarchical prediction structure are used to eliminate redundancy in the LFI's angular and spatial domains. At the decoding side, high-quality dense LF is reconstructed using a view synthesis method based on the residual channel attention network (RCAN).

In three different LF datasets, our proposed coding framework not only reduces the transmitted bit rate but also maintains a higher view quality than the current more advanced methods.

(1) A new sampling scheme is designed to synthesize high-quality LFIs while better ensuring LF angular domain sparsity. (2) To further eliminate redundancy in the spatial domain, new ranking schemes and hierarchical prediction structures are designed. (3) A synthetic network based on RCAN and a novel loss function is designed to mitigate the perceptual quality loss due to the coding process.

Sleep arousal detection is an important factor to monitor the sleep disorder.

Thus, a unique nth layer one-dimensional (1D) convolutional neural network-based U-Net model for automatic sleep arousal identification has been proposed.

The proposed method has achieved area under the precision–recall curve performance score of 0.498 and area under the receiver operating characteristics performance score of 0.946.

No other researchers have suggested U-Net-based detection of sleep arousal.

From the experimental results, it has been found that U-Net performs better accuracy as compared to the state-of-the-art methods.

Sleep arousal detection is an important factor to monitor the sleep disorder. Objective of the work is to detect the sleep arousal using different physiological channels of human body.

It will help in improving mental health by monitoring a person's sleep.

In the process of building the “Belt and Road” and “Bright Road” community of interests between China and Kazakhstan, this paper proposes the construction of an inland nuclear power plant in Kazakhstan. Considering the uncertainty of investment in nuclear power generation, the authors propose the MGT (Monte-Carlo and Gaussian Radial Basis with Tensor factorization) utility evaluation model to evaluate the risk of investment in nuclear power in Kazakhstan and provide a relevant reference for decision making on inland nuclear investment in Kazakhstan.

Based on real options portfolio combined with a weighted utility function, this study takes into account the uncertainties associated with nuclear power investments through a minimum variance Monte Carlo approach, proposes a noise-enhancing process combined with geometric Brownian motion in solving complex conditions, and incorporates a measure of investment flexibility and strategic value in the investment, and then uses a deep noise reduction encoder to learn the initial values for potential features of cost and investment effectiveness. A Gaussian radial basis function used to construct a weighted utility function for each uncertainty, generate a minimization of the objective function for the tensor decomposition, and then optimize the objective loss function for the tensor decomposition, find the corresponding weights, and perform noise reduction to generalize the nonlinear problem to evaluate the effectiveness of nuclear power investment. Finally, the two dimensions of cost and risk (estimation of investment value and measurement of investment risk) are applied and simulated through actual data in Kazakhstan.

The authors assess the core indicators of Kazakhstan's nuclear power plants throughout their construction and operating cycles, based on data relating to a cluster of nuclear power plants of 10 different technologies. The authors compared it with several popular methods for evaluating the benefits of nuclear power generation and conducted subsequent sensitivity analyses of key indicators. Experimental results on the dataset show that the MGT method outperforms the other four methods and that changes in nuclear investment returns are more sensitive to changes in costs while operating cash flows from nuclear power are certainly an effective way to drive investment reform in inland nuclear power generation in Kazakhstan at current levels of investment costs.

Future research could consider exploring other excellent methods to improve the accuracy of the investment prediction further using sparseness and noise interference. Also consider collecting some expert advice and providing more appropriate specific suggestions, which will facilitate the application in practice.

The Novel Coronavirus epidemic has plunged the global economy into a deep recession, the tension between China and the US has made the energy cooperation road unusually tortuous, Kazakhstan in Central Asia has natural geographical and resource advantages, so China–Kazakhstan energy cooperation as a new era of opportunity, providing a strong guarantee for China's political and economic stability. The basic idea of building large-scale nuclear power plants in Balkhash and Aktau is put forward, considering the development strategy of building Kazakhstan into a regional international energy base. This work will be a good inspiration for the investment of nuclear generation.

This study solves the problem of increasing noise by combining Monte Carlo simulation with geometric Brownian motion under complex conditions, adds the measure of investment flexibility and strategic value, constructs the utility function of noise reduction weight based on Gaussian radial basis function and extends the nonlinear problem to the evaluation of nuclear power investment benefit.

Intent detection (ID) and slot filling (SF) are two important tasks in natural language understanding. ID is to identify the main intent of a paragraph of text. The goal of SF is to extract the information that is important to the intent from the input sentence. However, most of the existing methods use sentence-level intention recognition, which has the risk of error propagation, and the relationship between intention recognition and SF is not explicitly modeled. Aiming at this problem, this paper proposes a collaborative model of ID and SF for intelligent spoken language understanding called ID-SF-Fusion.

ID-SF-Fusion uses Bidirectional Encoder Representation from Transformers (BERT) and Bidirectional Long Short-Term Memory (BiLSTM) to extract effective word embedding and context vectors containing the whole sentence information respectively. Fusion layer is used to provide intent–slot fusion information for SF task. In this way, the relationship between ID and SF task is fully explicitly modeled. This layer takes the result of ID and slot context vectors as input to obtain the fusion information which contains both ID result and slot information. Meanwhile, to further reduce error propagation, we use word-level ID for the ID-SF-Fusion model. Finally, two tasks of ID and SF are realized by joint optimization training.

We conducted experiments on two public datasets, Airline Travel Information Systems (ATIS) and Snips. The results show that the Intent ACC score and Slot F1 score of ID-SF-Fusion on ATIS and Snips are 98.0 per cent and 95.8 per cent, respectively, and the two indicators on Snips dataset are 98.6 per cent and 96.7 per cent, respectively. These models are superior to slot-gated, SF-ID NetWork, stack-Prop and other models. In addition, ablation experiments were performed to further analyze and discuss the proposed model.

This paper uses word-level intent recognition and introduces intent information into the SF process, which is a significant improvement on both data sets.