Search

The monitoring of arterial blood pressure (ABP) in anesthetized patients is crucial for preventing hypotension, which can lead to adverse clinical outcomes. Several efforts have been devoted to develop artificial intelligence-based hypotension prediction indices. However, the use of such indices is limited because they may not provide a compelling interpretation of the association between the predictors and hypotension. Herein, an interpretable deep learning model is developed that forecasts hypotension occurrence 10 min before a given 90-s ABP record. Internal and external validations of the model performance show the area under the receiver operating characteristic curves of 0.9145 and 0.9035, respectively. Furthermore, the hypotension prediction mechanism can be physiologically interpreted using the predictors automatically generated from the proposed model for representing ABP trends. Finally, the applicability of a deep learning model with high accuracy is demonstrated, thus providing an interpretation of the association between ABP trends and hypotension in clinical practice.

Background: Proper maintenance of hypnosis is crucial for ensuring the safety of patients undergoing surgery. Accordingly, indicators, such as the Bispectral index (BIS), have been developed to monitor hypnotic levels. However, the black-box nature of the algorithm coupled with the hardware makes it challenging to understand the underlying mechanisms of the algorithms and integrate them with other monitoring systems, thereby limiting their use., Objective: We propose an interpretable deep learning model that forecasts BIS values 25 s in advance using 30 s electroencephalogram (EEG) data., Material and Methods: The proposed model utilized EEG data as a predictor, which is then decomposed into amplitude and phase components using fast Fourier Transform. An attention mechanism was applied to interpret the importance of these components in predicting BIS. The predictability of the model was evaluated on both regression and binary classification tasks, where the former involved predicting a continuous BIS value, and the latter involved classifying a dichotomous status at a BIS value of 60. To evaluate the interpretability of the model, we analyzed the attention values expressed in the amplitude and phase components according to five ranges of BIS values. The proposed model was trained and evaluated using datasets collected from two separate medical institutions., Results and Conclusion: The proposed model achieved excellent performance on both the internal and external validation datasets. The model achieved a root-mean-square error of 6.614 for the regression task, and an area under the receiver operating characteristic curve of 0.937 for the binary classification task. Interpretability analysis provided insight into the relationship between EEG frequency components and BIS values. Specifically, the attention mechanism revealed that higher BIS values were associated with increased amplitude attention values in high-frequency bands and increased phase attention values in various frequency bands. This finding is expected to facilitate a more profound understanding of the BIS prediction mechanism, thereby contributing to the advancement of anesthesia technologies., Competing Interests: Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Sung-Hoon Kim reports financial support was provided by Ministry of Health & Welfare, Republic of Korea. Sung-Hoon Kim reports financial support was provided by Asan Institute for Life Sciences. Junetae Kim reports financial support was provided by National Cancer Center of Korea. Sung-Hoon Kim has patent Method for predicting depth of anesthesia based on EEG signal and anesthesia depth prediction device performing method pending to Asan Medical Center. Hee-Sun Park has patent Method for predicting depth of anesthesia based on EEG signal and anesthesia depth prediction device performing method pending to Asan Medical Center. Junetae Kim has patent Method for predicting depth of anesthesia based on EEG signal and anesthesia depth prediction device performing method pending to National Cancer Center of Korea. Eugene Hwang has patent Method for predicting depth of anesthesia based on EEG signal and anesthesia depth prediction device performing method pending to KAIST., (Copyright © 2023 Elsevier B.V. All rights reserved.)

Accurate segregation of retinal blood vessels network plays a crucial role in clinical assessments, treatments, and rehabilitation process. Owing to the presence of acquisition and instrumentation anomalies, precise tracking of vessels network is challenging. For this, a new fundus image segmentation framework is proposed by combining deep neural networks, and hidden Markov model. It has three main modules: the Atrous spatial pyramid pooling-based encoder, the decoder, and hidden Markov model vessel tracker. The encoder utilized modified ResNet18 deep neural networks model for low-and-high-levels features extraction. These features are concatenated in module-II by the decoder to perform convolution operations to obtain the initial segmentation. Previous modules detected the main vessel structure and overlooked some small capillaries. For improved segmentation, hidden Markov model vessel tracker is integrated with module-I and-II to detect overlooked small capillaries of the vessels network. In last module, final segmentation is obtained by combining multi-oriented sub-images using logical OR operation. This novel framework is validated experimentally using two standard DRIVE and STARE datasets. The developed model offers high average values of accuracy, area under the curve, and sensitivity of 99.8, 99.0, and 98.2%, respectively. Analysis of the results revealed that the developed approach offered enhanced performance in terms of sensitivity 18%, accuracy 3%, and specificity 1% over the state-of-the-art approaches. Owing to better learning and generalization capability, the developed approach tracked blood vessels network efficiently and automatically compared to other approaches. The proposed approach can be helpful for human eye assessment, disease diagnosis, and rehabilitation process., (© 2022. International Association of Scientists in the Interdisciplinary Areas.)

Background: An artificial intelligence (AI) model using chest radiography (CXR) may provide good performance in making prognoses for COVID-19., Objective: We aimed to develop and validate a prediction model using CXR based on an AI model and clinical variables to predict clinical outcomes in patients with COVID-19., Methods: This retrospective longitudinal study included patients hospitalized for COVID-19 at multiple COVID-19 medical centers between February 2020 and October 2020. Patients at Boramae Medical Center were randomly classified into training, validation, and internal testing sets (at a ratio of 8:1:1, respectively). An AI model using initial CXR images as input, a logistic regression model using clinical information, and a combined model using the output of the AI model (as CXR score) and clinical information were developed and trained to predict hospital length of stay (LOS) ≤2 weeks, need for oxygen supplementation, and acute respiratory distress syndrome (ARDS). The models were externally validated in the Korean Imaging Cohort of COVID-19 data set for discrimination and calibration., Results: The AI model using CXR and the logistic regression model using clinical variables were suboptimal to predict hospital LOS ≤2 weeks or the need for oxygen supplementation but performed acceptably in the prediction of ARDS (AI model area under the curve [AUC] 0.782, 95% CI 0.720-0.845; logistic regression model AUC 0.878, 95% CI 0.838-0.919). The combined model performed better in predicting the need for oxygen supplementation (AUC 0.704, 95% CI 0.646-0.762) and ARDS (AUC 0.890, 95% CI 0.853-0.928) compared to the CXR score alone. Both the AI and combined models showed good calibration for predicting ARDS (P=.079 and P=.859)., Conclusions: The combined prediction model, comprising the CXR score and clinical information, was externally validated as having acceptable performance in predicting severe illness and excellent performance in predicting ARDS in patients with COVID-19., (©Hyun Woo Lee, Hyun Jun Yang, Hyungjin Kim, Ue-Hwan Kim, Dong Hyun Kim, Soon Ho Yoon, Soo-Youn Ham, Bo Da Nam, Kum Ju Chae, Dabee Lee, Jin Young Yoo, So Hyeon Bak, Jin Young Kim, Jin Hwan Kim, Ki Beom Kim, Jung Im Jung, Jae-Kwang Lim, Jong Eun Lee, Myung Jin Chung, Young Kyung Lee, Young Seon Kim, Sang Min Lee, Woocheol Kwon, Chang Min Park, Yun-Hyeon Kim, Yeon Joo Jeong, Kwang Nam Jin, Jin Mo Goo. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 16.02.2023.)

Importance: Dual-energy chest radiography exhibits better sensitivity than single-energy chest radiography, partly due to its ability to remove overlying anatomical structures., Objectives: To develop and validate a deep learning-based synthetic bone-suppressed (DLBS) nodule-detection algorithm for pulmonary nodule detection on chest radiographs., Design, Setting, and Participants: This decision analytical modeling study used data from 3 centers between November 2015 and July 2019 from 1449 patients. The DLBS nodule-detection algorithm was trained using single-center data (institute 1) of 998 chest radiographs. The DLBS algorithm was validated using 2 external data sets (institute 2, 246 patients; and institute 3, 205 patients). Statistical analysis was performed from March to December 2021., Exposures: DLBS nodule-detection algorithm., Main Outcomes and Measures: The nodule-detection performance of DLBS model was compared with the convolution neural network nodule-detection algorithm (original model). Reader performance testing was conducted by 3 thoracic radiologists assisted by the DLBS algorithm or not. Sensitivity and false-positive markings per image (FPPI) were compared., Results: Training data consisted of 998 patients (539 men [54.0%]; mean [SD] age, 54.2 [9.82] years), and 2 external validation data sets consisted of 246 patients (133 men [54.1%]; mean [SD] age, 55.3 [8.7] years) and 205 patients (105 men [51.2%]; mean [SD] age, 51.8 [9.1] years). Using the external validation data set of institute 2, the bone-suppressed model showed higher sensitivity compared with that of the original model for nodule detection (91.5% [109 of 119] vs 79.8% [95 of 119]; P < .001). The overall mean of FPPI with the bone-suppressed model was reduced compared with the original model (0.07 [17 of 246] vs 0.09 [23 of 246]; P < .001). For the observer performance testing with the data of institute 3, the mean sensitivity of 3 radiologists was 77.5% (95% [CI], 69.9%-85.2%), whereas that of radiologists assisted by DLBS modeling was 92.1% (95% CI, 86.3%-97.3%; P < .001). The 3 radiologists had a reduced number of FPPI when assisted by the DLBS model (0.071 [95% CI, 0.041-0.111] vs 0.151 [95% CI, 0.111-0.210]; P < .001)., Conclusions and Relevance: This decision analytical modeling study found that the DLBS model was more sensitive to detecting pulmonary nodules on chest radiographs compared with the original model. These findings suggest that the DLBS model could be beneficial to radiologists in the detection of lung nodules in chest radiographs without need of the specialized equipment or increase of radiation dose.

Background and Objectives: Dementia is one of the brain diseases with serious symptoms such as memory loss, and thinking problems. According to the World Alzheimer Report 2016, in the world, there are 47 million people having dementia and it can be 131 million by 2050. There is no standard method to diagnose dementia, and consequently unable to access the treatment effectively. Hence, the computational diagnosis of the disease from brain Magnetic Resonance Image (MRI) scans plays an important role in supporting the early diagnosis. Alzheimer's Disease (AD), a common type of Dementia, includes problems related to disorientation, mood swings, not managing self-care, and behavioral issues. In this article, we present a new computational method to diagnosis Alzheimer's disease from 3D brain MR images., Methods: An efficient approach to diagnosis Alzheimer's disease from brain MRI scans is proposed comprising two phases: I) segmentation and II) classification, both based on deep learning. After the brain tissues are segmented by a model that combines Gaussian Mixture Model (GMM) and Convolutional Neural Network (CNN), a new model combining Extreme Gradient Boosting (XGBoost) and Support Vector Machine (SVM) is used to classify Alzheimer's disease based on the segmented tissues., Results: We present two evaluations for segmentation and classification. For comparison, the new method was evaluated using the AD-86 and AD-126 datasets leading to Dice 0.96 for segmentation in both datasets and accuracies 0.88, and 0.80 for classification, respectively., Conclusion: Deep learning gives prominent results for segmentation and feature extraction in medical image processing. The combination of XGboost and SVM improves the results obtained.

Chest X-rays (CXRs) can help triage for Coronavirus disease (COVID-19) patients in resource-constrained environments, and a computer-aided detection system (CAD) that can identify pneumonia on CXR may help the triage of patients in those environment where expert radiologists are not available. However, the performance of existing CAD for identifying COVID-19 and associated pneumonia on CXRs has been scarcely investigated. In this study, CXRs of patients with and without COVID-19 confirmed by reverse transcriptase polymerase chain reaction (RT-PCR) were retrospectively collected from four and one institution, respectively, and a commercialized, regulatory-approved CAD that can identify various abnormalities including pneumonia was used to analyze each CXR. Performance of the CAD was evaluated using area under the receiver operating characteristic curves (AUCs), with reference standards of the RT-PCR results and the presence of findings of pneumonia on chest CTs obtained within 24 hours from the CXR. For comparison, 5 thoracic radiologists and 5 non-radiologist physicians independently interpreted the CXRs. Afterward, they re-interpreted the CXRs with corresponding CAD results. The performance of CAD (AUCs, 0.714 and 0.790 against RT-PCR and chest CT, respectively hereinafter) were similar with those of thoracic radiologists (AUCs, 0.701 and 0.784), and higher than those of non-radiologist physicians (AUCs, 0.584 and 0.650). Non-radiologist physicians showed significantly improved performance when assisted with the CAD (AUCs, 0.584 to 0.664 and 0.650 to 0.738). In addition, inter-reader agreement among physicians was also improved in the CAD-assisted interpretation (Fleiss' kappa coefficient, 0.209 to 0.322). In conclusion, radiologist-level performance of the CAD in identifying COVID-19 and associated pneumonia on CXR and enhanced performance of non-radiologist physicians with the CAD assistance suggest that the CAD can support physicians in interpreting CXRs and helping image-based triage of COVID-19 patients in resource-constrained environment., Competing Interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: Chang Min Park holds stock of Promedius and stock options of Lunit Inc. and Coreline Soft. Eui Jin Hwang, Hyungjin Kim, and Chang Min Park received research grants from Lunit Inc, outside the present paper. We checked all authors’ competing interests and confirmed that our statement does not alter the adherence to PLOS ONE policy.

Purpose: We aimed to develop a deep learning (DL)-based algorithm for automated quantification of aortic valve calcium (AVC) from non-enhanced electrocardiogram-gated cardiac CT scans and compare performance of DL-measured AVC volume and Agatston score with those of visual gradings by radiologist readers for classification of AVC severity., Method: A total of 589 CT examinations performed at a single center between March 2010 and August 2017 were retrospectively included. The DL algorithm was designed to segment AVC and to quantify AVC volume, and Agatston score was calculated using attenuation values. Manually measured AVC volume and Agatston score were used as ground truth. To validate AVC segmentation performance, the Dice coefficient was calculated. For observer performance testing, four radiologists determined AVC grade in two reading rounds. The diagnostic performance of DL-measured AVC volume and Agaston score for classifying severe AVC was compared with that of each reader's assessment., Results: After applying the DL algorithm, the Dice coefficient score was 0.807. In patients with AVC, accuracy of DL-measured AVC volume for AVC grading was 97.0 % with area under the curve (AUC) of 0.964 (95 % confidence interval [CI] 0.923-1) in the test set, which was better than the radiologist readers (accuracy 69.7 %-91.9 %, AUC 0.762-0.923) with manually measured AVC volume as ground truth. When manually measured AVC Agatston score was used as ground truth, accuracy of DL-measured AVC Agatston score for AVC grading was 92.9 % with AUC of 0.933 (95 % CI 0.885-0.981) in the test set, which was also better than the radiologist readers (accuracy 77.8-89.9 %, AUC 0.791-0.903)., Conclusions: DL-based automated AVC quantification may be comparable with manual measurements. The diagnostic performance of the DL-measured AVC volume and Agatston score for classification of severe AVC outperforms radiologist readers., (Copyright © 2021 Elsevier B.V. All rights reserved.)

In the face of increasing irregular temperature patterns and climate shifts, the need for accurate power consumption prediction is becoming increasingly important to ensure a steady supply of electricity. Existing deep learning models have sought to improve prediction accuracy but commonly require greater computational demands. In this research, on the other hand, we introduce DelayNet, a lightweight deep learning model that maintains model efficiency while accommodating extended time sequences. Our DelayNet is designed based on the observation that electronic series data exhibit recurring irregular patterns over time. Furthermore, we present two substantial datasets of electricity consumption records from South Korean buildings spanning nearly two years. Empirical findings demonstrate the model's performance, achieving 21.23%, 43.60%, 17.05% and 21.71% improvement compared to recurrent neural networks, gated-recurrent units, temporal convolutional neural networks and ARIMA models, as well as greatly reducing model complexity and computational requirements. These findings indicate the potential for micro-level power consumption planning, as lightweight models can be implemented on edge devices. [ABSTRACT FROM AUTHOR]

Continuous efforts were made in detecting cracks in images. Varied CNN models were developed and tested for detecting or segmenting crack regions. However, most datasets used in previous works contained clearly distinctive crack images. No previous methods were validated on blurry cracks captured in low definitions. Therefore, this paper presented a framework of detecting the regions of blurred, indistinct concrete cracks. The framework divides an image into small square patches which are classified into crack or non-crack. Well-known CNN models were employed for the classification and compared with each other with experimental tests. This paper also elaborated on critical factors—the patch size and the way of labeling patches—which had considerable influences on the training performance. Furthermore, a series of post-processes for measuring crack lengths were introduced. The proposed framework was tested on the images of bridge decks containing blurred thin cracks and showed reliable performance comparable to practitioners. [ABSTRACT FROM AUTHOR]

Recently, deep learning models proliferate in the prediction of power demand for efficient planning of power consumption. However, the "black-box" characteristics of deep learning hinders from establishing a specific plan because it cannot explain the cause of the prediction. Recently, there are several attempts to explain the result of deep learning through the analysis of the input attributes that influence the prediction, but they lack of appropriate explanation because of ignoring the time-series property of the input data. In this paper, we propose a deep learning model to explain the impact of the input attributes on the prediction by taking account of the long-term and short-term properties of the time-series forecasting. The model consists of (i) two encoders to represent the power information for prediction and explanation, (ii) a decoder to predict the power demand from the concatenated outputs of encoders, and (iii) an explainer to identify the most significant attributes for predicting the energy consumption. Kullback–Leibler divergence in the loss function induces the long-term and short-term dependencies in latent space constructed by the second encoder. Several experiments on the benchmark dataset of household electric energy demand show that the proposed method explains the prediction appropriately with the most influential input attributes in the long-term and short-term dependencies. We can trade off the gain of the time-series explanation of the result against a slight degradation of the prediction performance. [ABSTRACT FROM AUTHOR]

This work discusses the challenges of multi-label image classification and presents a novel Efficient Shuffle Net (EffShuffNet) based on a convolutional neural network (CNN) architecture to address these challenges. Multi-label classification is difficult as the complexity of prediction increases with the number of labels and classes, and current multi-model approaches require optimized deep learning models which increase computational costs. The EffShuff block divides the input feature map into two parts and processes them differently, with one half undergoing a lightweight convolution and the other half undergoing average pooling. The EffShuff transition component shuffles the feature maps after lightweight convolution, resulting in a 57.9% reduction in computational cost compared to ShuffleNetv2. Furthermore, we propose EffShuff-Dense architecture, which incorporates Dense connection to further emphasize low-level features. In experiments, the EffShuffNet achieved 96.975% accuracy in age and gender classification, which is 5.83% higher than the state-of-the-art, while EffShuffDenseNet was even better with 97.63% accuracy. Additionally, the proposed models were found to have better classification performance with smaller model sizes in fine-grained image classification experiments. [ABSTRACT FROM AUTHOR]

According to the World Alzheimer Report 2015, 46 million people are living with dementia in the world. The diagnosis of diseases helps doctors treating patients better. One of the signs of diseases is related to white matter, grey matter and cerebrospinal fluid. Therefore, the automatic segmentation of three tissues in brain imaging especially from magnetic resonance imaging (MRI) plays an important role in medical analysis. In this research, we proposed an effective approach to segment automatically these tissues in three-dimensional (3D) brain MRI. First, a deep learning model is used to segment the sure and unsure regions. In the unsure region, another deep learning model is used to classify each pixel. In the experiments, an adaptive U-net model is used to segment the sure and unsure regions, and the Local Convolutional Neural Network (CNN) model with multiple inputs is used to classify each pixel only in the unsure region. Our method was evaluated with a real image database, Internet Brain Segmentation Repository database, with 18 persons (IBSR 18) (https://www.nitrc.org/projects/ibsr) and compared with state of art methods being the results very promising. [ABSTRACT FROM AUTHOR]

Liver cancer is a major cause of death, with a high mortality rate worldwide. For better treatment and rehabilitation of liver cancer, assessing and evaluating drug delivery and response prediction is vital. In this study, we propose a novel approach for the prediction of the liver anticancer drug response using modified ResNet101 deep learning with transfer learning (TL) concept for deep features extraction. Dimension reduction algorithms, PCA and t -SNE are applied to capture global and local structures of deep features, respectively. Herein, a new fusion scheme is introduced that uses global and local levels of image information to improve the classification. Reduced deep fused features are used to develop a quadratic discriminant analysis (QDA) prediction model. The proposed approach is evaluated on fluorescent images of human hepatocellular carcinoma (HepG2) treated using an anticancer drug functionalized by cobalt ferrite@barium titanate (CFO@BTO) nanoparticles. A dataset comprising 203 HepG2 microscopic images is used to train and test the model using the split ratio of 75%:25% by employing a 5-fold cross-validation technique. The proposed system achieved a high accuracy of 98.0% compared to other state-of-the-art approaches. The developed pipeline is flexible and can be extended for prostate, lung, and breast cancers. [ABSTRACT FROM AUTHOR]

The brain is more sensitive to stress than other organs and can develop many diseases under excessive stress. In this study, we developed a method to improve the accuracy of emotional stress recognition using multi-channel electroencephalogram (EEG) signals. The method combines a three-dimensional (3D) convolutional neural network with an attention mechanism to build a 3D convolutional gated self-attention neural network. Initially, the EEG signal is decomposed into four frequency bands, and a 3D convolutional block is applied to each frequency band to obtain EEG spatiotemporal information. Subsequently, long-range dependencies and global information are learned by capturing prominent information from each frequency band via a gated self-attention mechanism block. Using frequency band mapping, complementary features are learned by connecting vectors from different frequency bands, which is reflected in the final attentional representation for stress recognition. Experiments conducted on three benchmark datasets for assessing the performance of emotional stress recognition indicate that the proposed method outperforms other conventional methods. The performance analysis of proposed methods confirms that EEG pattern analysis can be used for studying human brain activity and can accurately distinguish the state of stress. [ABSTRACT FROM AUTHOR]

Utilizing pre-trained models involves fully or partially using pre-trained parameters as initialization. In general, configuring a pre-trained model demands practitioners' knowledge about problems or an exhaustive trial–error experiment according to a given task. In this paper, we propose tuning trainable layers using a genetic algorithm on a pre-trained model that is fine-tuned on single-channel image datasets for a classification task. The single-channel dataset comprises images from grayscale and preprocessed audio signals transformed into a log-Mel spectrogram. Four deep-learning models used in the experimental evaluation employed the pre-trained model with the ImageNet dataset. The proposed genetic algorithm was applied to find the highest fitness for every generation to determine the selective layer tuning of the pre-trained models. Compared to the conventional fine-tuning method and random layer search, our proposed selective layer search with a genetic algorithm achieves higher accuracy, on average, by 9.7% and 1.88% (MNIST-Fashion), 1.31% and 1.14% (UrbanSound8k), and 2.2% and 0.29% (HospitalAlarmSound), respectively. In addition, our searching method can naturally be applied to various datasets of the same task without prior knowledge about the dataset of interest. [ABSTRACT FROM AUTHOR]

Forecasting, commonly used in econometrics, meteorology, or energy consumption prediction, is the field of study that deals with time series data to predict future trends. Former studies have revealed that both traditional statistical models and recent deep learning-based approaches have achieved good performance in forecasting. In particular, temporal convolutional networks (TCNs) have proved their effectiveness in several time series benchmarks. However, presented TCN models are too heavy to deploy on resource-constrained systems, such as edge devices. As a resolution, this study proposes a stride–dilation mechanism for TCN that favors a lightweight model yet still achieves on-pair accuracy with the heavy counterparts. We also present the Chonnam National University (CNU) Electric Power Consumption dataset, the dataset of energy consumption measured at CNU by smart meters every hour. The experimental results indicate that our best model reduces the mean squared error by 32.7%, whereas the model size is only 1.6% compared to the baseline TCN. [ABSTRACT FROM AUTHOR]

Algorithmic bias indicates the discrimination caused by algorithms, which occurs with protected features such as gender and race. Many researchers have tried to define the fairness and devise methods to mitigate bias, but it is still premature to reach the unanimous definition and evaluation metrics of fairness according to society, times and cultures. In this paper, we introduce three evaluation metrics such as parity difference, equalized opportunity and equalized odds that can deal with various definitions of algorithmic bias, and concretely divide the three general approaches further into seven methods with some challenges, resulting in relabeling, generation, fair representation (for pre-processing), constraint optimization, regularization (for in-processing), calibration and thresholding (for post-processing). Among them, the pre-processing method is widely used due to its versatility, but it has limitation to deal with the information on data and features related with bias appropriately. In order to preserve the characteristics of the original data while excluding the information about the features causing bias, we propose a preprocessing approach based on information theory that avoids collision in the dual optimization, where the latent space is divided into two subspaces. Experiments are conducted with the well-known benchmark datasets of Census and COMPAS, and two real-world tasks: facial emotion recognition and text sentiment analysis. The information theoretic approach is promising to achieve fair machine learning by reducing the bias caused by several features such as age, race and gender. [ABSTRACT FROM AUTHOR]

Crop leaf disease management and control pose significant impact on enhancement in yield and quality to fulfill consumer needs. For smart agriculture, an intelligent leaf disease identification system is inevitable for efficient crop health monitoring. In this view, a novel approach is proposed for crop disease identification using feature fusion and PCA-LDA classification (FF-PCA-LDA). Handcrafted hybrid and deep features are extracted from RGB images. TL-ResNet50 is used to extract the deep features. Fused feature vector is obtained by combining handcrafted hybrid and deep features. After fusing the image features, PCA is employed to select most discriminant features for LDA model development. Potato crop leaf disease identification is used as a case study for the validation of the approach. The developed system is experimentally validated on a potato crop leaf benchmark dataset. It offers high accuracy of 98.20% on an unseen dataset which was not used during the model training process. Performance comparison of the proposed technique with other approaches shows its superiority. Owing to the better discrimination and learning ability, the proposed approach overcomes the leaf segmentation step. The developed approach may be used as an automated tool for crop monitoring, management control, and can be extended for other crop types. [ABSTRACT FROM AUTHOR]

The prognostics and health management (PHM) aims to provide a comprehensive solution for equipment health care. Classifying the operation mode of excavator, one of the tasks in the PHM, is important to evaluate the remaining useful lifetime. Several studies have been conducted to classify the operations with either video or sensor data, but they have several limitations to use only one type of data. A model trained with sensor data cannot classify the similar operations such as "digging" and "ditch digging" , whereas a model with video data is vulnerable to surrounding condition like weather. In this paper, to overcome these shortcomings, we propose a deep neural network ensemble called FusionNet that classifies the operations of excavator. Two models are trained with sensor data and video frames respectively, where the feature extractors are transferred to the FusionNet. The proposed network ensemble performs a flexible and well-optimized classification by automatically calculating weights according to the extracted feature vectors and combining them. To verify the proposed model, several experiments are conducted with the real-world data. The proposed model achieves the accuracy of 99.17% which outperforms the conventional methods. We also confirm that the proposed model can address the shortcomings of using only one type of data and maximize the benefits through the automatic weighting of extracted features. [ABSTRACT FROM AUTHOR]

Objective: This study aimed to stratify the early pneumonia trajectory on chest radiographs and compare patient characteristics in dyspneic patients with coronavirus disease 2019 (COVID-19). Materials and methods: We retrospectively included 139 COVID-19 patients with dyspnea (87 men, 62.7±16.3 years) and serial chest radiographs from January to September 2020. Radiographic pneumonia extent was quantified as a percentage using a previously-developed deep learning algorithm. A group-based trajectory model was used to categorize the pneumonia trajectory after symptom onset during hospitalization. Clinical findings, and outcomes were compared, and Cox regression was performed for survival analysis. Results: Radiographic pneumonia trajectories were categorized into four groups. Group 1 (n = 83, 59.7%) had negligible pneumonia, and group 2 (n = 29, 20.9%) had mild pneumonia. Group 3 (n = 13, 9.4%) and group 4 (n = 14, 10.1%) showed similar considerable pneumonia extents at baseline, but group 3 had decreasing pneumonia extent at 1–2 weeks, while group 4 had increasing pneumonia extent. Intensive care unit admission and mortality were significantly more frequent in groups 3 and 4 than in groups 1 and 2 (P <.05). Groups 3 and 4 shared similar clinical and laboratory findings, but thrombocytopenia (<150×103/μL) was exclusively observed in group 4 (P =.016). When compared to groups 1 and 2, group 4 (hazard ratio, 63.3; 95% confidence interval, 7.9–504.9) had a two-fold higher risk for mortality than group 3 (hazard ratio, 31.2; 95% confidence interval, 3.5–280.2), and this elevated risk was maintained after adjusting confounders. Conclusion: Monitoring the early radiologic trajectory beyond baseline further prognosticated at-risk COVID-19 patients, who potentially had thrombo-inflammatory responses. [ABSTRACT FROM AUTHOR]

Noise from automobiles, such as buzzing, squeaking, and rattling (BSR) noises, is a key factor in automobile quality assessment. It is necessary to classify these noises for appropriate handling and prevention. Although many researchers have conducted studies to classify noise, they suffer from several problems: difficulty in extracting appropriate features, insufficient data to train a classifier, and weak robustness to surrounding noise. This paper proposes a method called latent semantic controlling generative adversarial networks (LSC-GAN) to solve these problems. To capture the features of data, a variational autoencoder (VAE), an autoencoder with approximate inference in a latent Gaussian model, learns the data representation by projecting them into the latent space according to their features and reconstructing the projected data. Because the generator and the discriminator of the LSC-GAN are trained simultaneously, the capacity to extract the characteristics of the data is improved and a knowledge space of classifiable data is also expanded with insufficient data. While data are generated by the generator, the encoder projects them back to the latent space according to their characteristics to advance the ability to extract features. Finally, the encoder is trained to the classifier, which is trained to classify BSR noises. The proposed classifier outperforms other models and achieves an accuracy of 96.68%. We confirm using a confusion matrix that the proposed model classifies the types of insufficient class better than other models. Our proposed model classifies data with accuracy of 94.68%, even if the data contains surrounding noise, which means it is more robust to BSR with surrounding noise than other models. [ABSTRACT FROM AUTHOR]

Recent empirical works reveal that visual representation learned by deep neural networks can be successfully used as descriptors for image retrieval. A common technique is to leverage pre-trained models to learn visual descriptors by ranking losses and fine-tuning with labeled data. However, retrieval systems' performance significantly decreases when querying images of lower resolution than the training images. This study considered a contrastive learning framework fine-tuned on features extracted from a pre-trained neural network encoder equipped with an attention mechanism to address the image retrieval task for low-resolution image retrieval. Our method is simple yet effective since the contrastive learning framework drives similar samples close to each other in feature space by manipulating variants of their augmentations. To benchmark the proposed framework, we conducted quantitative and qualitative analyses of CARS196 (mAP = 0.8804), CUB200-2011 (mAP = 0.9379), and Stanford Online Products datasets (mAP = 0.9141) and analyzed their performances. [ABSTRACT FROM AUTHOR]

Urban mobility is a vital aspect of any city and often influences its physical shape as well as its level of economic and social development. A thorough analysis of mobility patterns in urban areas can provide various benefits, such as the prediction of traffic flow and public transportation usage. In particular, based on its exceptional ability to extract patterns from complex large-scale data, embedding based on deep learning is a promising method for analyzing the mobility patterns of urban residents. However, as urban mobility becomes increasingly complex, it becomes difficult to embed patterns into a single vector because of its limited capacity. In this paper, we propose a novel method for analyzing urban mobility based on deep learning. The proposed method involves clustering mobility patterns and embedding them to capture their implicit meaning. Clustering groups mobility patterns based on their spatiotemporal characteristics, and embedding provides meaningful information regarding both individual residents (i.e., personalized mobility) and all residents as a whole, enabling a more effective analysis of mobility patterns. Experiments were performed to predict the successive points of interest (POIs) based on transportation data collected from 1.5 million citizens in a large metropolitan city; the results demonstrate that the proposed method achieves top-1, 3, and 5 accuracies of 73.64%, 88.65%, and 91.54%, respectively, which are much higher than those of the conventional method (59.48%, 75.85%, and 80.1%, respectively). We also demonstrate that the proposed method facilitates the analysis of urban mobility through arithmetic operations between POI vectors. [ABSTRACT FROM AUTHOR]

In this letter, to improve data rate over wireless communication channels, we propose a deep Q network (DQN)-based adaptive modulation scheme by using Markov decision process (MDP) model. The proposed algorithm makes the reinforcement learning agent to select rate region boundaries as the states, which divide signal-to-noise ratio (SNR) range into rate regions. The simulation results show that spectral efficiency can be improved on the average by 0.5395 bps/Hz in wide SNR range. [ABSTRACT FROM AUTHOR]

In the conventional approaches using reinforcement learning (RL) for resource allocation in wireless networks, the structure of the policy depends on network circumstances such as the number of users and quality-of-service requirements. Due to this dependence, the policy is hard to be used in a practical system where the network circumstance is dynamically changing. To resolve this issue, we propose a circumstance-independent policy that can effectively address the different network circumstances even with a single policy. Thus, contrary to the conventional RL approaches, the proposed policy can be easily applied in the practical system. We then develop a deep RL algorithm to learn it. Through simulation results, we show that a single proposed policy can be used over different circumstances, and it achieves a close performance to the circumstance-dependent policy for each circumstance, which learns the optimal policy for the corresponding circumstance. [ABSTRACT FROM AUTHOR]

Music listening is a very personal and situational behavior for modern people who always carry smartphones in everyday life. Therefore, contextual information, such as user’s current activity and mood state could be used to greatly improve music recommendations. In this paper, we develop a smartphone-based mobile system that includes two core modules for recognizing human activities and then accordingly recommending music. In the proposed method, a deep residual bidirectional gated recurrent neural network is applied to obtain high activity recognition accuracy from accelerometer signals on the smartphone. In order to improve the performance of tempo-oriented music classification, an ensemble of dynamic classification using a long-term modulation spectrum and sequence classification using a short-term spectrogram is used. Music recommendation is performed using the relationship between the recognized human activities and the music files indexed by tempo-oriented music classification that reflects user preference models in order to achieve high user satisfaction. The results of comprehensive experiments on real data confirm the accuracy of the proposed activity-aware music recommendation framework. [ABSTRACT FROM AUTHOR]

Social cybermedia has greatly changed the world's perspective on information sharing and propagation. These platforms build virtual user communities in the form of followings or followers that can spread positive or negative content. These virtual communities play a vital role in the spread of hate content, which may create a severe law-and-order situation and social unrest in any country. Hate content can include offensive, anti-state, anti-religion, sexist, and racist posts. Classifying such content and identifying influential individuals who participate in these communities are challenging tasks for law enforcement agencies. This study proposes novel deep learning and graph-based approaches to identifying hate content, followed by approaches to detecting communities and exploring social media to detect hate content. Twitter is used as a case study, and tweets are extracted and annotated by linguistic experts to develop a dataset for experimentation and validation. The proposed customized LSTM-GRU model is used to classify hate content into six categories. The developed model offers an accuracy of 98.14% on the obtained dataset. The Girvan–Newman algorithm is employed for Twitter community detection and successfully identifies the most influential individuals. Moreover, the proposed approach precisely detects intraclass communities and highly influential persons. The developed model effectively identifies hate tweets and communities and can be used to monitor social media to identify any potential threat of unrest. • Development of an annotated dataset comprising six classes. • Development of various customized LSTM -and RNN-based classification models. • Combination of customized LSTM and GRU algorithms for better text recognition. • The proposed LSTM-GRU model shows promising performance for text classification. • Hate-spreading communities are detected using Girvan-Newman graph algorithm. [ABSTRACT FROM AUTHOR]

• We propose a novel deep learning model to stably predict electric energy consumption. • The model can explain the prediction results by exploiting the latent space. • We analyze the model with the large data collected in an actual residential house. • We achieve the highest performance in high resolution compared with the previous works. • We explain the variables of appliances that influence the prediction performance. Recently, many studies have exploited the potential of deep learning to forecast energy demand, but they cannot explain the results. They only analyze the simple correlations between the input and output to discover the most important input features, or they depend on the manual investigation of the latent space embedded with power demand patterns. In this paper, to overcome these shortcomings, we propose a deep autoencoder that can explain the prediction results by manipulating the latent space. It consists of 1) a power encoder that embeds power information, 2) an auxiliary encoder that embeds auxiliary information for an interpretable latent space in two dimensions, 3) a predictor that predicts power demand by using concatenated values of the latent variables extracted from the two encoders, and 4) an explainer that provides the most important input features in predicting the future demand by utilizing the interpretable latent variables. Several experiments on a dataset of household electric energy demand show that the proposed model not only performs better than conventional models, with a mean squared error of 0.376 in predicting electricity demand for 60 min, but also provides the capacity to explain the results by analyzing the correlation of inputs, latent variables, and energy demand predicted. [ABSTRACT FROM AUTHOR]

The optimization for hardware processor and system for performing deep learning operations such as Convolutional Neural Networks (CNN) in resource limited embedded devices are recent active research area. In order to perform an optimized deep neural network model using the limited computational unit and memory of an embedded device, it is necessary to quickly apply various configurations of hardware modules to various deep neural network models and find the optimal combination. The Electronic System Level (ESL) Simulator based on SystemC is very useful for rapid hardware modeling and verification. In this paper, we designed and implemented a Deep Learning Accelerator (DLA) that performs Deep Neural Network (DNN) operation based on the RISC-V Virtual Platform implemented in SystemC in order to enable rapid and diverse analysis of deep learning operations in an embedded device based on the RISC-V processor, which is a recently emerging embedded processor. The developed RISC-V based DLA prototype can analyze the hardware requirements according to the CNN data set through the configuration of the CNN DLA architecture, and it is possible to run RISC-V compiled software on the platform, can perform a real neural network model like Darknet. We performed the Darknet CNN model on the developed DLA prototype, and confirmed that computational overhead and inference errors can be analyzed with the DLA prototype developed by analyzing the DLA architecture for various data sets. [ABSTRACT FROM AUTHOR]

Smart factories merge various technologies in a manufacturing environment in order to improve factory performance and product quality. In recent years, these smart factories have received a lot of attention from researchers. In this paper, we introduce a defective product classification system based on deep learning for application in smart factories. The key component of the proposed system is a programmable logic controller (PLC) artificial intelligence (AI) embedded board; we call this an AI Edge-PLC module. A pre-trained defective product classification model is uploaded to a cloud service from where the AI Edge-PLC can access and download it for use on a certain product, in this case, electrical wiring. Next, we setup the system to collect electrical wiring data in a real-world factory environment. Then, we applied preprocessing to the collected data in order to extract a region of interest (ROI) from the images. Due to limitations on the availability of appropriate labeled data, we used the transfer learning method to re-train a classification model for our purposes. The pre-trained models were then optimized for applications on AI Edge-PLC boards. After carrying out classification tasks, on our electrical wire dataset and on a previously published casting dataset, using various deep neural networks including VGGNet, ResNet, DenseNet, and GoogLeNet, we analyzed the results achieved by our system. The experimental results show that our system is able to classify defective products quickly with high accuracy in a real-world manufacturing environment. [ABSTRACT FROM AUTHOR]

In this paper, a deep learning-based successive interference cancellation (SIC) scheme for use in nonorthogonal multiple access (NOMA) communication systems is investigated. NOMA has become a notable technique in the field of mobile wireless communication because of its capacity to overcome orthogonality, unlike a conventional orthogonal frequency division multiple access (OFDMA) communication system. In NOMA communication systems, SIC is one of the decoding schemes applied at receivers for downlink NOMA transmissions. In this paper, a convolutional neural network (CNN)-based SIC scheme is proposed to improve performance of the single base station and multiuser NOMA scheme. In contrast to existing SIC schemes, the proposed CNN-based SIC scheme can effectively mitigate losses resulting from imperfections of the SIC. The simulation results indicate that the CNN-based SIC method can successfully relieve conventional SIC impairments and achieve good detection performance. Consequently, a CNN-based SIC scheme can be considered as a potential technique for use in NOMA detection schemes. [ABSTRACT FROM AUTHOR]

Early identification of pneumonia is essential in patients with acute febrile respiratory illness (FRI). We evaluated the performance and added value of a commercial deep learning (DL) algorithm in detecting pneumonia on chest radiographs (CRs) of patients visiting the emergency department (ED) with acute FRI. This single-centre, retrospective study included 377 consecutive patients who visited the ED and the resulting 387 CRs in August 2018–January 2019. The performance of a DL algorithm in detection of pneumonia on CRs was evaluated based on area under the receiver operating characteristics (AUROC) curves, sensitivity, specificity, negative predictive values (NPVs), and positive predictive values (PPVs). Three ED physicians independently reviewed CRs with observer performance test to detect pneumonia, which was re-evaluated with the algorithm eight weeks later. AUROC, sensitivity, and specificity measurements were compared between "DL algorithm" vs. "physicians-only" and between "physicians-only" vs. "physicians aided with the algorithm". Among 377 patients, 83 (22.0%) had pneumonia. AUROC, sensitivity, specificity, PPV, and NPV of the algorithm for detection of pneumonia on CRs were 0.861, 58.3%, 94.4%, 74.2%, and 89.1%, respectively. For the detection of 'visible pneumonia on CR' (60 CRs from 59 patients), AUROC, sensitivity, specificity, PPV, and NPV were 0.940, 81.7%, 94.4%, 74.2%, and 96.3%, respectively. In the observer performance test, the algorithm performed better than the physicians for pneumonia (AUROC, 0.861 vs. 0.788, p = 0.017; specificity, 94.4% vs. 88.7%, p < 0.0001) and visible pneumonia (AUROC, 0.940 vs. 0.871, p = 0.007; sensitivity, 81.7% vs. 73.9%, p = 0.034; specificity, 94.4% vs. 88.7%, p < 0.0001). Detection of pneumonia (sensitivity, 82.2% vs. 53.2%, p = 0.008; specificity, 98.1% vs. 88.7%; p < 0.0001) and 'visible pneumonia' (sensitivity, 82.2% vs. 73.9%, p = 0.014; specificity, 98.1% vs. 88.7%, p < 0.0001) significantly improved when the algorithm was used by the physicians. Mean reading time for the physicians decreased from 165 to 101 min with the assistance of the algorithm. Thus, the DL algorithm showed a better diagnosis of pneumonia, particularly visible pneumonia on CR, and improved diagnosis by ED physicians in patients with acute FRI. [ABSTRACT FROM AUTHOR]

As energy demand grows globally, the energy management system (EMS) is becoming increasingly important. Energy prediction is an essential component in the first step to create a management plan in EMS. Conventional energy prediction models focus on prediction performance, but in order to build an efficient system, it is necessary to predict energy demand according to various conditions. In this paper, we propose a method to predict energy demand in various situations using a deep learning model based on an autoencoder. This model consists of a projector that defines an appropriate state for a given situation and a predictor that forecasts energy demand from the defined state. The proposed model produces consumption predictions for 15, 30, 45, and 60 min with 60-min demand to date. In the experiments with household electric power consumption data for five years, this model not only has a better performance with a mean squared error of 0.384 than the conventional models, but also improves the capacity to explain the results of prediction by visualizing the state with t-SNE algorithm. Despite unsupervised representation learning, we confirm that the proposed model defines the state well and predicts the energy demand accordingly. [ABSTRACT FROM AUTHOR]