A High-Precision Feature Extraction Network of Fatigue Speech from Air Traffic Controller Radiotelephony Based on Improved Deep Learning

Zhiyuan Shen,Yitao Wei 한국통신학회 2021 ICT Express Vol.7 No.4

Air traffic controller (ATC) fatigue is receiving considerable attention in recent studies because it represents a major cause of air traffic incidences. Research has revealed that the presence of fatigue can be detected by analysing speech utterances. However, constructing a complete labelled fatigue data set is very time-consuming. Moreover, a manually constructed speech collection will often contain only little key information to be used effectively in fatigue recognition, while multilevel deep models based on such speech materials often have overfitting problems due to an explosive increase of model parameters. To address these problems, a novel deep learning framework is proposed in this study to integrate active learning (AL) into complex speech features selected from a large set of unlabelled speech data in order to overcome the loss of information. A shallow feature set is first extracted using stacked sparse autoencoder networks, in which fatigue state challenge features from a manually selected speaker set of are exploited as the input vector. A densely connected convolutional autoencoder (DCAE) is then proposed to learn advanced features automatically from spectrograms of the selected data to supplement the fatigue features. The network can be effectively trained using a relatively small number of labelled samples with the help of AL sampling strategies, and the addition of a dense block to the convolutional automatic encoder can decrease the number of parameters and make the model easier to fit. Finally, the two above-mentioned features are combined using multiple kernel learning with a support-vector-machine classifier. A series of comparative experiments using the Civil Aviation Administration of China radiotelephony corpus demonstrates that the proposed method provides a significant improvement in the detection precision compared to current state-of-the-art approaches.

Characteristics of La_2O_3―Y_2O_3―Mo cermet cathode with RE_2O_3 nano particles

Jinshu Wang,Wei Liu,Zhiyuan Ren,Fan Yang,Yiman Wang,Yucheng Du,Meiling Zhou 한국물리학회 2011 Current Applied Physics Vol.11 No.3

The secondary electron emission property and microstructure of La_2O_3―Y_2O_3―Mo cermet cathode have been studied. It shows that the cathode prepared by Sol―Gel doping method has fine microstructure containing rare earth oxide (RE_2O_3) nanoparticles and uniform distribution of different substances. Auger Electron Spectroscopy (AES) depth profile result shows that an RE_2O_3 layer about 28 nm in thickness could be formed on the cathode surface, which plays an important role in the emission. Pre-activation of the cathode is favorable for the improvement of emission property. La_2O_3―Y_2O_3―Mo cathode preactivated exhibits good secondary emission property. The secondary emission yield could reach 5.24,about 1.8 times higher than that of original cathode. The enhancement on the conductivity of materials resulting from the pre-activation is favorable for the replenishment and transport of electrons, thus the secondary emission property of the cathode could be improved.

Bone Grafting Can Promote the Prognosis of Displaced Femoral Neck Fractures: A Follow-up of the Clinical Significance of Bone Defects

Xiaozhong Zhu,Wei Wang,Zhiyuan Wang,Yi Zhu,Guangyi Li,Jiong Mei 대한정형외과학회 2023 Clinics in Orthopedic Surgery Vol.15 No.4

Background: Femoral neck fractures (FNFs) comprise a large proportion of osteoporotic fractures in Asia. However, the full range of prognostic variables that affect prognosis remains unclear. Here, we aimed to determine whether the severity of bone defects at the fracture site and other variables impact the prognosis of displaced FNFs. Methods: We evaluated the incidence of FNF internal fixation failures at regular intervals after surgery in data collected retrospectively. Digital Imaging and Communications in Medicine (DICOM) magnetic resonance imaging data of the displaced FNFs of 204 patients (> 20 years old; mean age, 52.3 years; men, 55.4%) who underwent internal fixation were used to construct threedimensional (3D) virtual models of the femoral neck region. We calculated the position and volume of bone defect (VBD) using our independently developed algorithm and Mimics software. Each participant was followed up for at least 24 months; complications were noted and correlated with VBD and demographic and clinical variables. Results: On the basis of VBD values calculated from virtual reduction models, 57 patients were categorized as having a mild defect, 100 as having a moderate defect, and 47 as having a severe defect. Age (p = 0.046) and VBD (p < 0.001) were significantly correlated with internal fixation failure. Multivariate analysis revealed that severe bone defects were associated with internal fixation failure (adjusted odds ratio [aOR], 23.073; 95% confidence interval [CI], 2.791–190.732) and complications (aOR, 8.945; 95% CI, 1.829–43.749). In patients with a severe defect, bone grafting was inversely associated with internal fixation failure (aOR, 0.022; 95% CI, 0.002–0.268) and complications (aOR, 0.023; 95% CI, 0.002–0.299). Conclusions: Bone defect severity was associated with internal fixation failure and other complications. For young adults with large VBDs, bone grafting of the defect can reduce the risk of internal fixation failure. These results provide useful new quantitative information for precisely classifying displaced FNFs and guiding subsequent optimal treatments.

Preparation and Photoelectric Properties of Silver Nanowire/ZnO Thin Film Ultraviolet Detector

Zhenfeng Li,Wei Xiao,Hongzhi Zhou,Zhiyuan Shi,Rongqing Li,Jia Zhang,Yang Li,Peng He,Shuye Y. Zhang 대한금속·재료학회 2023 ELECTRONIC MATERIALS LETTERS Vol.19 No.5

Ultraviolet (UV) detectors have important applications in many fi elds. ZnO is an excellent semiconductor material for the preparation of UV detectors because of its large direct gap in forbidden bandwidth, its intrinsic response band in the UV region, and its high exciton binding energy. In this paper, high-performance ZnO thin fi lms with the optically advantageous nonpolar structure were prepared by using an atomic layer deposition, and the dominant crystal plane gradually changes from the amorphous phase to the (100) crystal plane. The conventional photoconductor structure ZnO UV detector was enhanced by the surface plasmon exciton eff ect of Ag nanostructure. When the operating voltage is 5 V and the response light is 350 nm, there is a maximum optical responsiveness of up to 131 A/W. The UV/visible rejection ratio can reach 1824 times. When the ZnO thin fi lm deposition thickness is 400 deposition cycles and about 72 nm, the ZnO thin fi lm UV detector obtains the highest responsiveness (5 V, 365 nm) of 365 A/W. Comparing the photovoltaic performance of the ZnO thin-fi lm detector with the enhanced ZnO thin-fi lm detector and its optimal response wavelength, it is found that the enhanced ZnO thin-fi lm detector increased the photoresponse value by about 100 times. The optimal response wavelength in the UV region is blueshifted, and the UV-visible rejection ratio and optical response rate are signifi cantly improved.

Edge Computing Model based on Federated Learning for COVID-19 Clinical Outcome Prediction in the 5G Era

Ruochen Huang,Zhiyuan Wei,Wei Feng,Yong Li,Changwei Zhang,Chen Qiu,Mingkai Chen 한국인터넷정보학회 2024 KSII Transactions on Internet and Information Syst Vol.18 No.4

As 5G and AI continue to develop, there has been a significant surge in the healthcare industry. The COVID-19 pandemic has posed immense challenges to the global health system. This study proposes an FL-supported edge computing model based on federated learning (FL) for predicting clinical outcomes of COVID-19 patients during hospitalization. The model aims to address the challenges posed by the pandemic, such as the need for sophisticated predictive models, privacy concerns, and the non-IID nature of COVID-19 data. The model utilizes the FATE framework, known for its privacy-preserving technologies, to enhance predictive precision while ensuring data privacy and effectively managing data heterogeneity. The model's ability to generalize across diverse datasets and its adaptability in real-world clinical settings are highlighted by the use of SHAP values, which streamline the training process by identifying influential features, thus reducing computational overhead without compromising predictive precision. The study demonstrates that the proposed model achieves comparable precision to specific machine learning models when dataset sizes are identical and surpasses traditional models when larger training data volumes are employed. The model's performance is further improved when trained on datasets from diverse nodes, leading to superior generalization and overall performance, especially in scenarios with insufficient node features. The integration of FL with edge computing contributes significantly to the reliable prediction of COVID-19 patient outcomes with greater privacy. The research contributes to healthcare technology by providing a practical solution for early intervention and personalized treatment plans, leading to improved patient outcomes and efficient resource allocation during public health crises.

A dual-experience pool deep reinforcement learning method and its application in fault diagnosis of rolling bearing with unbalanced data

Yuxiang Kang,Guo Chen,Wenping Pan,Xunkai Wei,Hao Wang,Zhiyuan He 대한기계학회 2023 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.37 No.6

A dual-experience pool deep reinforcement learning (DEPDRL) model is proposed for rolling bearing fault diagnosis with unbalanced data. In this method, a dualexperience pool structure is designed to store the sample data of majority and minority classes. A parallel double residual network model is established to extract deep features of the majority and minority input samples, respectively. In the process of training, the proposed balanced cross-sampling technique is used to randomly select samples from dual-experience pool in a certain proportion to realize the training of a double residual network model. We show the effectiveness of our method on three standard data sets, and compared with Resnet18, DCNN, DQN and DQNimb methods, the results show that DEPDRL has the best performance. Finally, with wavelet time-frequency graph as input, DEPDRL is applied to rolling bearing fault diagnosis with unbalanced test data. The results show that on a variety of unbalanced data sets, both the diagnostic accuracy and the G-means value of the DEPDRL are more than 5 % higher than other algorithms, which fully indicates that the DEPDRL has a very high fault diagnosis ability of rolling bearing with unbalanced data.

Kinetic study on the reaction of palmitic acid with ethanol catalyzed by deep eutectic solvent based on dodecyl trimethyl ammonium chloride

Shan Jiang,Zuoxiang Zeng,Weilan Xue,Wei Zhang,Zhiyuan Zhou 한국화학공학회 2020 Korean Journal of Chemical Engineering Vol.37 No.9

This study explored the direct esterification of palmitic acid and ethanol using a deep eutectic solvent (DES) as catalyst to produce biodiesel. Three novel deep eutectic solvents (DTAC-PTSA, DTAC-2PTSA, DTAC3PTSA) were successfully prepared by mixing dodecyl trimethyl ammonium chloride (DTAC) and p-toluenesulfonic acid monohydrate (PTSA) in a molar ratio of 1: z (z=1, 2, 3). After testing, DTAC-3PTSA was found to have the best catalytic performance among the three types of DESs and was therefore selected as the catalyst for subsequent experiments. The effects of agitation speed, ethanol to palmitic acid molar ratio (), temperature and catalyst dosage were studied by investigating the change of palmitic acid conversion rate with time under different conditions, respectively. Then, the pseudo-homogeneous (PH) model was utilized to describe the kinetic behavior of this reaction between 328.15-348.15 K and it was found to work well for the experimental data obtained. Moreover, the catalytic performance of DTAC-3PTSA was detected to have no significant change in the cycle test. Therefore, DTAC-3PTSA can be considered as a substitute for traditional catalysts to produce biodiesel and the kinetic data obtained here can be used for further up-scaling study