Table of contents

The 2021 6th International Conference on Electronic Technology and Information Science (ICETIS 2021) was held on January 22-24, 2021 as a virtual conference due to the growing concerns over the coronavirus outbreak (COVID-19), and in order to protect the well-being of our attendees, partners, and staff as our number one priority. ICETIS 2021 is to bring together innovative academics and industrial experts in the field of electronic technology and information science to a common forum. The primary goal of the conference is to promote research and developmental activities in electronic technology and information science and another goal is to promote scientific information interchange between researchers, developers, engineers, students, and practitioners working all around the world.

This scientific event brings together more than 160 national and international researchers in electronic technology and information science. On top of the local participants coming from different national universities, international participants are also registered from different countries, namely UK, USA, Malaysia. During the conference, the conference model was divided into three sessions, including oral presentations, keynote speeches, and online Q&A discussion. In the first part, some scholars, whose submissions were selected as the excellent papers, were given about 5-10 minutes to perform their oral presentations one by one. Then in the second part, keynote speakers were each allocated 30-45 minutes to hold their speeches.

In the second part, we invited three professors as our keynote speakers. Assoc. Prof. Mahmoud AlShawabkeh, from Guangxi Normal University for Nationalities, China. His research area focuses on Computer Network Security and Information Systems. And then we had Assoc. Prof. Lei Zhang from Henan University, China. He devoted in Machine learning, Big data, Information security and other related research area. Assoc. Prof. Zhihan Lv, from Qingdao University, China. His research mainly focuses on Augmented Reality, Virtual Reality, Geographic Information Systems, Computer Vision, Multimedia, 3D Visualization, 3D Graphics, Serious Game, HCI, Bigdata, Internet of Things. Their insightful speeches had triggered heated discussion in the third session of the conference. Every participant praised this conference for disseminating useful and insightful knowledge.

The proceedings are a compilation of the accepted papers and represent an interesting outcome of the conference. Topics include but are not limited to the following areas: Electronics Engineering, Mechanical Engineering, Computer Engineering and more related topics. All the papers have been through rigorous review and process to meet the requirements of International publication standard.

We would like to acknowledge all of those who supported ICETIS 2021. The help and contribution of each individual and institution was instrumental in the success of the conference. In particular, we would like to thank the organizing committee for its valuable inputs in shaping the conference program and reviewing the submitted papers.

The Committee of ICETIS 2021

List of Committee members are available in this pdf.

All papers published in this volume of Journal of Physics: Conference Series have been peer reviewed through processes administered by the Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.

• Type of peer review: Double-blind

• Conference submission management system: AI Scholar Submission System

• Number of submissions received: 356

• Number of submissions sent for review: 331

• Number of submissions accepted: 230

• Acceptance Rate (Number of Submissions Accepted/Number of Submissions Received X 100): 64.6%

• Average number of reviews per paper: 2

• Total number of reviewers involved: 150

• Any additional info on review process:

• Step 1. Each of the selected papers should be reviewed by two or three experts who are professional in the related area. A report of the reviewing result will come out by each expert.

• Step 2. After collecting the reviewing reports from the experts, one of the editors will review the reports and decide whether the reports are approved.

• Step 3. After that, the editors inform the authors of the reviewing results along with the three reviewing reports.

• Step 4. Authors are required to revise their paper according to the requirements, then hand in the revised version to the editors. The editors send the revised version to reviewers and will decide whether the paper is approved to be published according to the reviewers' final reports. And at last the editors inform the decision with every details of the publication requirements if the paper is accepted.

• Contact person for queries:

• Yongkang Xing

• Chongqing University

• ykxing@cqu.edu.cn

In recent years, with the rapid development of high-bandwidth applications and the rapid growth of Internet data services, the capacity of communication networks is growing rapidly. Optical fiber has become an ideal transmission medium for communication networks due to its advantages such as large bandwidth and low loss. Large-capacity and long-distance optical fiber communication networks have become the key research directions of optical communication systems. In this paper, the main technical means to achieve large-capacity and long-distance optical transmission are investigated. First, the application of orthogonal frequency division multiplexing (OFDM) technology in large-capacity transmission is briefly introduced, including its principle, main structure, advantages and disadvantages. Subsequently, wavelength division multiplexing (WDM) technology based on optical frequency comb is introduced. Finally, the methods to realize long-distance transmission are discussed, including Digital Backward Propagation (DBP), Phase-Conjugate Twin Wave (PCTW) and advanced modulation format.

In a two-degree-of-freedom unilateral collision system with a gap, the two oscillators will produce nonlinear motion. A dynamic model of the two-degree-of-freedom collision system is established, and the two oscillators are analyzed separately by numerical simulation. On this basis, an electronic circuit composed of TL074CN operational amplifier, capacitance and resistance is designed, which is equivalent to the mathematical model. The electronic circuit is simulated by multisim12.0, and the result is consistent with that of numerical simulation. The equivalent electronic circuit model not only has dramatically shortened the calculation time, but also has the advantage of adjustable dynamic parameters, which makes up for the lack of numerical simulation and provides a reference tool for the simulation analysis of multi-degree-of-freedom nonlinear mechanical vibration systems.

Since the performance of laser diodes is sensitive to ambient temperature, thermoelectric cooler temperature control circuits based on standard PID temperature compensation networks are usually designed to achieve diode constant temperature control. The thermoelectric cooler time constant and thermistor thermal time constant introduced in the temperature control system will reduce the stability margin of the system. In view of this, this paper designed a high-precision laser diode temperature control circuit based on the type III compensation network, and established the transfer function model, optimized the compensation network parameter tuning method, made the phase margin of the temperature control system reach π/8, and the long-term temperature control accuracy was better than 3 mK in the ambient temperature range of 5 to 40 °C.

The performance of finite impulse response (FIR) filtering is an important index of the real-time processing ability of digital signal processing system. In this paper, various FIR filtering algorithms are analyzed, and based on fast fir algorithm (FFA), block FFA (BFFA) is proposed. A 128 order configurable FIR filter circuit with 16 order as a block is designed. The analysis of experimental data shows that the 16 bit integer FIR filter performance reaches 881.78Gop/s.

Partial blind signature scheme (BLD-SS) can protect the specific content of the signed message by adding factor to blind the signed message and can also prevent the misuse of the signature by adding common information negotiated in advance. Revocable certificateless signature scheme (RCSS) can prevent the user's privacy from leaking by adding time key in the signature process and signature verification process. BLD-SS and RCSS both have wide applications in electronic-voting, electronic-cash system and other fields. In this paper, it presents a security analysis on a partial blind signature scheme and a revocable signature scheme proposed recently. It points out that the BLD-SS does not satisfy the unforgeability. A malicious user can forge a partial blind signature on any message and any common information without known from PKG (Private Key Generation). It also shows that the RCSS does not satisfy the unforgeability. A malicious enemy can forge a signature on any message and on any user. The PKG also cannot update the time key immediately when a user's key is leaked or his identity is expired. In order to overcome these problems, it proposes some simple improvement methods which are almost the same efficiency with the original signature scheme.

Avoiding the harm caused by nuclear radiation is an important topic in the field of nuclear physics. For nuclear radiation protection, it is necessary to show the distribution of radiation dose in the power plants. In this paper, we propose the methods combining ray-casting volume rendering and geometry model fusion to visualize the 3D nuclear radiation dose field. We implement a nuclear radiation dose field virtual simulation system based on Unity3D, combined with virtual reality to help workers conduct nuclear radiation operations and training. Moreover, we plan the maintenance path of the workers based on the NEAT algorithm, thus providing necessary radiation protection measures. The experimental results and evaluation show the efficiency of our virtual radiation dose visualization and simulation system.

Aiming at the characteristics of the memristor with memory function in the circuit, the establishment of the memristor model and the simulation of the recognition circuit based on the memristor were carried out through the PSPICE simulation software. The NiO memristor was successfully fabricated on Si substrate by magnetron sputtering method, and the typical V-I hysteresis characteristic curve was obtained. Connecting the NiO memristor to the identification hardware circuit, the physical test results show that it is the same as the simulation diagram of the recognition circuit based on the memristor, which further confirms that the identification circuit based on the memristor has the learning function.

The openness and complexity of wireless channels make collaborative spectrum sensing vulnerable to malicious users[1]. Therefore, it is very important to identify the attributes of malicious users before collaborative spectrum awareness networks make data fusion decisions. In this paper, a method combining reinforcement learning and cognitive user credit model is proposed, in which the maximum and minimum eigenvalues of signals are used as the initial information for exchange, and the whole sensing network tends to diffuse and fuse nodes with high credit. Finally, the convergence value of the whole network is compared with the decision threshold to complete collaborative spectrum sensing. By combining with consensus fusion network and traditional collaborative sensing algorithm, the proposed method can effectively improve the convergence speed of fusion network and shorten the sensing time on the premise of effectively identifying malicious users[2], so as to improve the spectrum sensing performance and make the collaborative sensing network more adaptive and stable.

In the field of natural language processing, language-based question answering has been widely studied and made great achievements. As an important form of search engine, question answering system can further improve the search quality compared with traditional search engines. Traditional tree-based indexing methods are inefficient in high-dimensional space. This paper focuses on the application of k nearest neighbor search in large-scale image databases, and researches on high-dimensional indexing technology based on vector approximation method. Fast-Hessian detector is used to detect feature points and generate SURF feature description vector; Then, the initial matching point pair is obtained by fast approximate nearest neighbor search algorithm, and then the unidirectional matching results are matched bidirectionally; at the same time, it makes a detailed analysis of the situation that the events in the sentences are verbs of "abstract change", so as to realize the automatic question and answer after the change of abstract relations between entities.

The construction of a virtual interactive model based on VR technology breaks through the single mode of traditional virtual reality helmets, through optical attitude sensing, immersive cave virtual reality, multi-person network collaboration, 9D interactive media applications, and mobile terminal interaction. The organic integration and application of technologies have improved the user's interactive experience, and the inheritance and development of local folk culture and the continuous research on intangible cultural heritage can be a useful exploration. However, the development cost of the entire design is relatively high, and there is no unified construction standard and policy support for such projects in China, especially the lack of effective support from the network application platform, which makes it difficult to implement online, and the reproducibility is low.

With the advent of information and intelligence era, electronic information technology has been rapidly developed and widely used. In the era of the Internet of Things, electronic information technology will continue to play an important role. This article will verify the integration of electronic information technology and the Internet of Things by exploring the application of electronic information technology in the Internet of Things. The integration of electronic information technology and the Internet of Things realizes the interconnection of people and people, things and things, people and things, satisfies people's needs for information transmission of goods, and provides great convenience for people's daily work.

Aiming at the problem of massive historical telemetry data storage in flight test, a lossless compression method based on adaptive interval run length encoding is proposed. Aiming at the problem of low compression efficiency of traditional run length encoding algorithm for word data, by studying the storage characteristics of telemetry data, this algorithm automatically identifies the frame format of telemetry data, and carries out longitudinal run length adaptive interval encoding for inter frame differential data to improve the compression efficiency. The test results show that the compression ratio of the improved algorithm is improved by 58.1% and 1.5% compared with the traditional run length encoding algorithm and the inter frame differential lateral run length encoding algorithm.

Since the 1990s, speech keyword recognition technology has gradually separated from speech recognition technology and officially become an important branch of speech recognition technology. In this paper, several speech keyword recognition technologies are studied and reviewed, including sample recognition methods, filler model methods, basic speech recognition systems based methods, neural network classifiers and end-to-end methods. Finally, the development prospect and advanced technology of speech keyword recognition will be discussed.

Improving the magnetic stealth capability of ships [7]is of great importance to the safety of ships. The comprehensive degaussing method is an important method of ship degaussing. This paper mainly introduces the principle, coil classification, key and difficult factors of the comprehensive degaussing method and its application prospect in China.

In this work, logarithmic mean method, which is based on the assumption that the diffusion coefficients profile is linear between nodes, is developed to calculate diffusion coefficients at cell interfaces. The new method is evaluated by solving one dimension steady diffusion equation and convection diffusion equation with strongly non-linear diffusion coefficients. Results indicate that the logarithmic mean method can achieve high precision as the Kirchhoff integral mean method even if the number of nodes is small, while the harmonic mean method needs more nodes to achieve the same accuracy. Furthermore, the logarithmic mean method only needs the diffusion coefficients of nodes, the disadvantage of Kirchhoff method which needs concrete expression of diffusion coefficients and numerous integral operation thus can be overcome.

Retinal prosthesis is an effective treatment to restore partial functional vision for degeneration diseases such as retinitis pigmentosa (RP). The persistent degeneration of the retina can influence the effect of retina prosthesis. However, how the different stages of retinal degeneration influence the retinal ganglion cell (RGC) responses to external electric stimulation remains unclear. In our present study, we established the multilayer retinal model of normal mouse and retinal degeneration (rd1) mouse in different degeneration stages (early, middle and late stage). Meanwhile, the morphology-realistic RGC models with membrane dynamics in different degeneration stages were also established. The effects of retinal structure, ganglion cell morphology and Na⁺ conductance on RGC threshold in response to electrical stimulation were explored. The simulation results showed that the RGC threshold of early degeneration stage was basically the same as that of the normal retina. While the thresholds of middle and late degeneration stages were both higher than that of the early stage, and the late stage showed a higher RGC threshold than the middle stage. This study would provide theoretical support for personalizing design of visual prostheses to treat different degeneration stages of RP patients.

With the progress of science and technology and the development of the times, people's living standards are gradually improving, and the use of travel tools is becoming more and more frequent. However, due to the lack of fuel resources in China and the great environmental pollution caused by the use of fuel resources, the country is vigorously promoting the development of new energy to alleviate the domestic oil crisis, among which electric energy is the most widely used energy. Enterprises can use power resources to realize wireless charging technology of electric vehicles. As far as the wireless charging technology currently studied in China is concerned, many technologies are not perfect, which can not meet the needs of people and enterprises for the performance of the wireless charging automobile industry. Therefore, according to the problems existing in the development of the current wireless charging technology, this paper makes relevant analysis.

Aiming at the measurement method of electrical capacitance tomography, a set of micro-capacitance acquisition system for planar array capacitance sensor is designed. Using the circuit principle design of high sensitivity and anti-interference, a dual-channel signal generator with adjustable amplitude and phase was designed to excite the capacitance sensor. The core capacitance voltage conversion circuit of the acquisition system adopts an AC excitation type conversion circuit with the ability to resist stray capacitance interference, and analyzes the error influencing factors in combination with the actual circuit components. The array control switch group is designed for multi-channel switching acquisition, and optimized in conjunction with the capacitor voltage conversion circuit. For the phase demodulation method, an orthogonal optimization method based on hardware multiplier is proposed. A micro-capacitor experiment is designed to calibrate the acquisition circuit. The experimental results show that the capacitance acquisition resolution of the micro-capacitor acquisition system within the acquisition range is 2.84fF, the linearity is 0.9998, and the acquisition error is less than 0.12fF.

The art design of interactive device is based on Arduino, an open-source technology and Processing, a programming language, whose application enriches the interactivity and expansibility of the device, and improves its visual effect and experience. With the rapid development of social economy, urban noise pollution has become increasingly severe, which has been yet to arouse enough public attention. Therefore, the research enables people to more intuitively experience the impact of urban noise through the interactive device, and realizes the interaction between the device and people through the communication between Arduino and Processing, so that people can deeply understand the severity and danger of urban noise pollution, thus calling for public's higher awareness of noise prevention.

Using neural networks to simulate and predict human visual attention mechanism is a hot topic in the field of computer vision. In this paper, we propose an end-to-end encoder-decoder network architecture to predict the fixation mechanism of human eyes, which consists of modules composed of multiple convolutional layers with different expansion rates to capture multi-scale features in parallel. In addition, the attention module is added on the basis of the encoder network structure, and a self-attention mechanism is introduced to capture the visual feature dependency in the channel size, and the semantic interdependence in the channel dimension is modeled to predict the visual salience more accurately. In this paper, five data sets and selected examples are used to demonstrate the effectiveness of the proposed method. Our method achieves competitive and consistent results on multiple evaluation indicators on MIT1003 and CAT2000 datasets.

In view of the low accuracy of single disturbances under the problem of noise interference, a new method of power quality disturbance classification based on deep belief network was proposed. A smooth wavelet multiscale transform is performed on the power quality disturbance signal, and then the soft threshold function is used to process the estimated wavelet coefficients for reconstructing the original signal. Further, it is proposed to use deep confidence network to classify and recognize the reconstructed single disturbance signal. The simulation results demonstrate that the recognition rate of this method for seven typical single disturbances is high. Even under 20dB noise interference, the classification accuracy rate is as high as 93% or more, which proves that the method has a strong ability to resist noise interference.

Multi-code combined spread spectrum (MCSS) is based on M-ary spread spectrum to improve information payload by combining multiple spread spectrum sequences, which solved the contradictions between spread spectrum gain and transmission efficiency under limited channel bandwidth conditions. This paper focuses on two mapping algorithms of multi-code combination spread spectrum and their decoding algorithms for MCSS, which transmits the information by combining multiple spreading sequences. The R-combined mapping algorithm is provided to keep the sequence orthogonal and the design complexity increases with the increase of the number of correlators at the receiver while the SVC algorithm is adopted and MMP algorithm is used to realize multi-code combination spread spectrum with increasing sequences. The BER performance of the two algorithms is compared through simulation analysis

To solve the problem of displacement survey fleet aircraft pylon and launcher, universal testing equipment is designed. The system is based on industrial computer architecture, the system through the software to control signal analog circuit with accuracy, emission, such as excitation signal, transmission device by signal feedback to complete various suspension test of the internal circuit. It has been proved by use that the test based on IPC is versatile, the operation interface is clear and easy to be used. The use of the test improved the test efficiency and the maintenance reliability, meet the requirement testing of the pylon/launcher.

Current reference is a very important module in analog circuits. The main purpose of this design is to provide a stable reference voltage for a switch magnetism-sensing circuit. Because of the inherent characteristics of the circuit, the reference voltage is set to 3.3V. In this paper, the reference voltage source is designed to have a wide input range, a stable output and a good temperature characteristic. In this paper, the simulation results of the reference voltage and the complete design layout are described, and a current reference is achieved, which is independent of process, voltage and temperature, with little area increases only. The simulation results based on a 4 μm 25 V bipolar process verify the correctness of the theoretical analysis and the feasibility of the proposed methods.

Under the premise of technological progress, people's thinking is constantly changing, and people have certain demands for new technologies. At present, people particularly favor information technology, which has changed people's lifestyles. Especially after the emergence of the Internet of Things, a novel environment has been created for people's living habits and enterprise production, and the distance between people can be effectively shortened. It can greatly improve the efficiency of life, highlighting the convenience. In the new era, the Internet of Things and electronic information technology are combined, and the two work together to help various industries move forward. It has targeted integration in multiple aspects such as QR code, intelligent transportation and electronic ordering, so that the Internet of Things brings forward action full of the times to the society, demonstrating its unique and effective use.

At present, keyword-based video retrieval is gradually difficult to adapt to the needs of the rapid development of the Internet due to its strong subjectivity and huge workload. As a result, multi-modal video retrieval based on deep learning has appeared. This retrieval method can conduct video retrieval through multiple methods such as text, image, and video, which fully meets the different retrieval needs of different users, and significantly improves the accuracy and effectiveness of video retrieval. Based on this, this article discusses in detail the design of a multi-modal video retrieval system based on deep learning, analyzes and designs each functional module of the system to provide reference for future related work.

The bullet velocity parameter is the base of the next analysis of many impact experiments. A complete and rapid measurement of velocity parameter before and after an impact of the projectile on the target is a problem to be solved. We proposed a method aiming at measuring the bullet velocity parameter from high-speed sequential images, completely and fast. We firstly set up the high-speed camera sensor network for grabbing and storing of high-speed sequential images. It contained two high-speed cameras, a synchronous control system, two master-slave computers to control other systems, and the supplement light system. Subsequently, a surveying control network was constructed to solve the parameters of videogrammetry. 3D reconstruction was applied to the solution of the bullet center 3D coordinates, and the bullet velocity parameter was finally solved by the approach of time difference. The experiment of measurement of the bullet velocity was done successfully. The bullet velocity parameter we measured was provided to the criminal investigation department.

Neural network is widely used as a behavioral model of power amplifier (PA) for its high flexibility and nonlinear fitting ability. This paper presents a novel simplified neural network (SNN) PA behavioral model. By employing an input classifier, the proposed SNN model reduces the number of input data needed for training the neural network, while achieving the same modeling accuracy in comparison to the already published neural network PA models. Simulation results show that the proposed model significantly reduces the computational load of PA modeling.

In this paper, we established a computer vision system for missing tablets detection in pharmaceutical industry during the manufacturing process. The images of blister pack with tablets in RGB color format were acquired by using a CCD industrial color camera on the industrial assembly line in real time. In order to reduce the effect of environmental brightness, the obtained RGB format images were transformed into HSI color format. According to the threshold of the hue and saturation, the image segmentation was carried out. Then, the tablets could be identified and determined. The Canny edge detection and Hough circle detection was used to mark the locations of tablets. Based on the locations and number of circles, the location of the missing tablets was obtained. This computer vision system could boost efficiency, reduce the cost of labor, and guarantee the safety of drug consumers.

Based on the different behaviors of multi-user group, this paper gives a method of demand analysis of broadband wireless mobile application. The analysis of broadband wireless mobile application, based on user behavior samples, can adopt the method of fuzzy cluster analysis to classify the user behavior by setting different clustering level, and build the model of each type of user behavior. The function of broadband wireless mobile application is designed by using the function description template based on class diagram, and the relationship between its function and user group behavior is analyzed by mapping matrix. At last, based on the application function and user behavior correlation analysis, the user behaviors are transformed into the key indicators and requirements of the application function.

Most researches on EEG emotion recognition are based on multi-electrode EEG signal acquisition equipment. There are few studies on single-electrode EEG equipment and the classification accuracy obtained is not ideal. In order to further improve the classification accuracy, the paper proposes an EEG emotion recognition method based on the combination of multiple features in the time-frequency domain. Firstly, the wavelet transform is used to decompose the EEG signals under the three types of emotion labels and extracts the time-frequency information of four frequency bands, including alpha, low beta, high beta, and gamma, and then uses different sizes of sliding time windows to calculate separately their statistical characteristics in the time-frequency domain. Finally, the Long Short-Term Memory (LSTM) network with time-frequency domain characteristics is used to extract the sequence information of the deep features and combine the output results of the softmax classifier. The results of experiments show that the average accuracy of the proposed method is 93.09% and 98.36% on emotion food and emotion state dataset, respectively. Compared with traditional machine learning methods and other deep learning methods, the time-frequency LSTM model named TF-LSTM in this paper has better EEG generalization ability and classification performance, and also provides a new feasible scheme for emotion recognition research based on single-electrode EEG signal.

To solve spectral aliasing of digital pulse width modulation (PWM) signal during digital up-conversion, a novel all-digital transmitter architecture using three-level quadrature differential RF-PWM is presented. By using triangle wave with 90° phase shift as the reference signal of quadrature differential PWM, the pulse coding of complex modulation signal and digital up-conversion will be realized. The feasibility of the transmitter is demonstrated by deriving the output formula of 3-level quadrature differential PWM, the design requirement is also given. Simulation and experiment results with QPSK and 16QAM signals show that the proposed scheme can significantly suppress the transmitter in-band noise, eliminate even harmonics, and effectively reduce the system operating frequency.

Based on the objectives and requirements of the pilotage service portfolio, using the design standards of HGDM organization, the nodes, activities and certification modules of ships entering and leaving the port are analysed, and the S-100 standard model of pilotage data is established. According to the characteristics and actual needs of pilotage, the technical design mode of C/S (client/server) mode of pilotage service portfolio is determined; the application example of pilotage service portfolio is verified by the service mode of tide riding opportunity in Ningbo waters for large container pilotage.

The successful research and development, promotion and application of 5G technology in the new era will have a huge impact on the various industries in society. The development of computer and network technology plays an important role in promoting the practical training. VR technology is one of the important products of information and network technology. It takes multimedia technology as the core and presents knowledge to students more vividly in an audio-visual way, which can help students better understand. The paper analyzes the application of 5G+VR technology in the training and teaching of integrated wiring technology in detail, and aims to provide help for better application of 5G+VR technology.

The aim of the study was to explore the correlation between brain network attributes and skin electrical signals in patients with mild cognitive impairment (MCI) for early diagnosis. Neuropsychological assessment for subjects and the resting-state fMRI data were collected. The brain network was constructed based on the resting-state fMRI data, and its attributes were analyzed by graph theory. Then wavelet transform was used to extract the features of skin electrical signal data. Finally, Pearson correlation method was used to calculate the correlation between brain network attributes and skin electrical signal. The results showed that brain network attributes of aLp, aλ, assortativity and synchronization in MCI were higher than those in normal control, while the hierarchical of MCI was lower than that in normal control. There were significant differences between the two groups of skin electrical signals in mean value, maximum value and minimum value. Moreover, the correlation results showed that neuropsychological assessment were correlated not only with brain network attributes but also with skin electrical signals, and there was a correlation between skin electrical signals and brain network attributes. Using skin electrical signals can provide a new early standard to assist in the diagnosis of MCI, with low cost and easy acquisition.

Time-Interleaved Analog-to-Digital Conversion (TIADC) is one of the significant technique in high speed waveform digitization systems. In our latest work, a 20-Gsps 12-bit TIADC system is designed and a series of tests are conducted on the proposed system and a broadband mismatch error correction is implemented. The test methods and results are specifically presented in this paper. The test results show that the -3dB bandwidth of the proposed system is 6 GHz and the ENOB performance is about 8.7 bits at 647 MHz and 7.6 bits at 6 GHz which is significantly enhanced by the mismatch error correction process.

The formulation of gear hobbing processing parameters relies on extensive professional knowledge and artificial experience, and inappropriate processing parameters will lead to low machining precision and high manufacturing costs. In order to obtain suitable processing parameters in high speed dry hobbing gear process, Gradient boosting regression (GBR) and Generalized Regression Neural Network (GRNN) are adopted to establish a process parameter optimization model targeting gear machining accuracy and machining energy consumption. And then a method based on Differential Evolution (DE) algorithm, parameter adaptive multi-object differential evolution (AMDE), for hobbing processing parameter optimization is proposed. Machining experiment verification shows that the established gear hobbing accuracy prediction model and energy consumption estimation model have high prediction accuracy, and AMDE has strong optimization capabilities and robustness.

Aiming at a large number of end-of-life (EOF) electrical products flowing into individual recycling outlets, it will squeeze the living space of formal dismantling companies and cause secondary pollution problems. The purpose of this paper is to establish a stable cooperative relationship between them to guide EOF products flowing from individual recycling outlets into the formal dismantling process. A game model involving the participation of formal dismantling enterprises and individual recycling outlets is constructed, and the role of government punishment mechanism intervention in the formation of active cooperative relations between participants in different parameter ranges is discussed. Then, the stability of the evolutionary equilibrium point is analysed. Finally, the Matlab software is used to simulate the influence of parameter on the formation of a stable cooperative relationship, and suggestions for promoting cooperation are provided.

This article is based on the application of industrial CT testing in additive manufacturing parts. Based on the existing industrial X-ray CT inspection equipment, it aims to solve the problem of identifying small defects under the uncertainty of test results. Under the premise of removing artifacts, this paper studies the measurement methods of small defects, and analyzes the uncertainty of the test results by designing a comparison test block. This paper combines theoretical analysis and experimental research, and compares the test results with the half-width method to study the accuracy of the test method. In order to improve the accuracy of industrial CT quantitative non-destructive detection of small defects, this paper closely studies the engineering application and combines the practical application of industrial CT to solve the basic problems of key research work.

Strengthening intrinsic safety of maritime traffic is essential for further modernization of the maritime traffic safety governance system and the governance capabilities. This paper studies the connotations of the intrinsic safety of maritime traffic, then analyzes the composition of the intrinsic safety system of maritime traffic and explores how to achieve intrinsic safety of the system factors such as crew, ships, navigation environment and safety management by referring to a typical case, and finally proposes maritime regulatory measures based on the intrinsic safety of maritime traffic and the long-term maritime management model for ensuring the intrinsic safety of maritime traffic.

With the rise and development of the Internet and the artificial intelligence boom, natural language processing has been greatly developed in various fields and industries, including power industry. Intelligent document understanding, as a sub field of natural language understanding, uses artificial intelligence technology to enable machines to have natural language understanding ability. It has always been the focus of researchers and industry, and it is also the core problem of intelligent semantic interaction. As a common method to expand the sample set, data enhancement technology is an important text and image processing technology. Its core purpose is to use limited data and produce value equivalent to that from much more data. This technology has been widely used in various fields of deep learning. This paper reviews the development of data enhancement technology, lists and discusses five kinds of text enhancement technology paths and their corresponding representative technologies, including back translation, random word replacement, non core word replacement, text enhancement based on context information, and text enhancement technology based on generative language model. It also analyzes the effectiveness of text enhancement technology in the aspects of regularization, transfer learning, improving model robustness, manifold and so on, to achieve the goal of improving the accuracy and effectiveness of document understanding in power industry.

In order to improve the fire extinguishing effect on lithium-ion battery fires, the paper experimented with compound solution of sodium dodecylbenzenesulfonate-dodecanol-sodium carbonate at a ratio of 3:2:2, polyfiouroalkyll betaine-alkylglycoside-potassium carbonate at a ratio of 2: 3: 3 and monoalkyl ether phosphate potassium-fatty acid ester polyoxyethylene etherurea at a ratio of 3:2: 3. The fire of the lithium-ion battery was extinguished with water mist containing additive compound solution by building lithium-ion battery fire extinguishing experimental platform. The experiment selected four different concentrations of 0.5%, 1%, 1.5%, and 2%. The results showed that the polyfluoroalkyl betaine-alkyl glycoside-potassium carbonate compound solution had better fire extinguishing effect in the early stage of thermal runaway, and the battery did not reignite. The sodium dodecylbenzene sul-fonate-dodecanol-sodium carbonate compound solution took the least time to reduce the battery temperature to below 200 °C. In the later period, the sodium dodecylbenzenesul-fonate-dodecanol-sodium carbonate compound solution has the greatest advantage in cooling rate. In the gas absorption test, the sodium dodecylbenzenesulfonate-dodecanoll-sodium carbonate compound solution has the best absorption effect on methane, and the polyfluoroalkyl betaine-alkyl glycoside-potassium carbonate compound solution releases CO. The concentration of the compound has a more obvious drop than the other two compound solutions.

There are lots of chemical laboratories in high-rise R&D buildings. Large amounts of hazardous chemicals were stored in them. Once a fire breaks out in high-rise R&D buildings, its severity will be much greater than that of high-rise residential building fire. Inflammable chemicals, explosive chemicals, toxic chemicals are usually get involved in hazardous chemical fire. FDS software is usually used for numerical simulation of hazardous chemical fire. Then, the impact of hazardous chemicals fire in high-rise R&D buildings can be analyzed under different fire conditions. This paper collected the hazardous chemicals stocks of 69 companies in the R&D buildings and screened the storage data of hazardous chemicals. Typical hazardous chemicals were selected as fire simulation objects. The consequences of hazardous chemical fires can be analyzed by simulation results. The difference between hazardous chemical fire in R&D high-rise buildings and residential building fire can be found out. At the same time, relevant performance indexes will be proposed for the evacuation of high-rise R&D hazardous chemical fire.

The increasing passenger throughput makes the large-scale aviation hub being unable to bear the burden gradually. Flight delays caused by cannot fly through the ATC tower in time are becoming more and more common, which makes airlines have to face the problem of passenger loss. Meanwhile, runway intrusion incidents caused by the high load operation of the airport also occur frequently. Once there is a collision, airlines will suffer huge economic losses. In order to avoid runway intrusion, improve air transport efficiency and reduce airline losses, runway collision avoidance system emerges as the times require. Based on the real airport map geographic coordinate data, this paper draws the airport dynamic map in the on-board moving map for the pilot, defines several kinds of runway intrusion scenarios, and introduces 3D real-time visual display into the platform in order to build a complete set of simulation verification platform for airport runway collision avoidance system. The research results of this paper can be used to verify the optimization algorithm of runway collision avoidance system and the test of airport dynamic map in the future, which is of great significance to improve the traffic condition of large airports in the future.

Transient electronics, a new type of electronic device, is made of degradable materials. After a programmed set period of operations, transient electronics can be partially or completely degraded when they are subjected to certain stimulus. Transient electronics can work normally without extreme conditions, and can degrade immediately after stable operation. It could supplement the limitation of traditional electronics, and could be of great interest for applications in electronic skin and sensors.

The main benefits and social welfare of demand response resources participating in the market are mostly based on qualitative evaluation in previous studies, and no more mature evaluation methods have been seen. This paper uses the evaluation model of the entropy-weighted close value method to analyze the advantages and disadvantages of the evaluation units. Advantages, the standard of each evaluation index is used as the sample to participate in the composition of the original environmental matrix, the entropy weight method is used to determine the weight of each evaluation index, the "most advantages" and "the worst points" are determined in a limited sample, and the sum of each sample point is calculated. The Euclidean distance between the above two points is used to determine the close value of the sample point, determine the order of the merits of each evaluation unit according to the close value, and finally determine the optimal solution for the demand side response.

This article reshapes the construction elements of Organizational Project Management (OPM) from the aspects of OPM governance, OPM methodology, knowledge management, talent management, etc. It establishes three main lines of strategy implementation, end-to-end process and revenue orientation, as well as clarifying the hierarchical division of the project and improving the information management mechanism, etc. Finally, it provides thoughts and suggestions for the construction of the OPM system of medium and large enterprises.

The safety management system standards are the basic document for enterprises to carry out safe production activities. This paper uses the methods of literature research and comparative analysis to sort out the safety standardization requirements of China's power companies. The development history of China's power safety management system standards is summarized. Through comparative research on the safety management system standards of three Chinese power enterprises, the concepts and methods of safety management of Chinese power companies was summed up, as well as the element settings and key features. The results show that the safety management system standards of Chinese power enterprises are mostly guided by risk management concepts to develop modular systems, and provides some guidance and suggestions for the preparation of safety management system standards for other Chinese power enterprises.

In today's society, the sudden nature of emergencies, many inducements, and large impact determine the difficulty of emergency management and decision-making. How to minimize losses and ensure the safety of people's lives and property after an emergency occurs is the top priority of the country's emergency management work today. Bayesian decision-making method is a common method for program selection: use Bayesian formula to modify the prior probability, and then use the modified probability to calculate the expected value of each program's utility. The plan with the higher expected value of utility is the best. Based on previous research results, this paper introduces the application background of original emergencies and the Bayesian decision-making method with fuzzy intervals, and uses specific cases to verify the effectiveness and rationality of the decision-making method.

In order to study the fatigue crack growth of directional solidified superalloy with single hole, the fatigue crack growth test of DZ125 was carried out under different stress levels. The fracture morphology of the specimens were analyzed by field emission scanning electron microscope, and the relationship between stress and life was fitted. It is found that the stress concentration around the hole and the defects lead to the crack initiation near the maximum principal stress point. The da/dN - Δ K satisfies Paris formula well.

This study developed the data acquisition module which involves electronics control hardware and mobile phone APP software, the module could control and monitor wastewater treatment equipment by the APP software running on mobile phone for wastewater treatment in the tribe of eastern Taiwan. The APP user of the tribe can keep abreast of the situation of wastewater treatment and discharge to maintain the water resource with a good water quality.

Safety management systems(SMSs) is a systematic safety management method, which is widely used and researched all over the world. This article reviews the current literature of safety management systems, summarizes and analyzes eight important research topics at the theoretical level, practical level, and standard level, including analysis of the connotation and extension, effectiveness analysis, integration of a safety management system and other systems, steps and content of construction, evaluation, comparison and selection of standards. Furthermore, based on the above analysis results and the current status of the application of safety management systems in China, two recommendations on important research directions are proposed. The research indicates that in order to improve the application effect of safety management systems in Chinese enterprises, we should concentrate on the basic theoretical research. Three questions need to be answered urgently at this stage: "What is a safety management system (SMS)?", "Why does it fail?" and "How to improve?".

The recently discovered large-scale two-dimensional (2D) early transition metal carbides and carbonitrides have captured people's attention as hot topics, called MXenes. MXene is a two-dimensional layered solid, linked by strong metals, ions and covalent bonds, e.g. Ti₂AlC, Ti₃AlC₂ and Ta₄AlC₃, usually by selectively etching the A element from the MAX phase to produce MXene. MXene has a an array of applications in all aspects cause its outstanding electrical conductivity, excellent specific surface area, good hydrophilicity and stability. In this article, the experimental and theoretical developments are reviewed in their synthesis, physicochemical properties and potential applications. MXene is expected to be applied in electrochemical energy storage and also provides current applications in batteries and super capacitor.

Strengthening the cooperative production of cigarette brands is to actively promote the reform of the resource allocation method centered on the brand. It is an important measure to promote the rational flow of production factors and the optimal allocation of industry resources. It is also to promote the rapid development of key brands in the industry, and it is truly a leading an important guarantee for the new pattern of brand development. However, due to many reasons, the cooperative production of cigarettes will inevitably produce various risks, which will endanger the brand reputation. How to achieve the common and sustainable development of cooperative production of cigarettes, strive to promote the high-quality development of cooperative production, and carry out risk prevention of cooperative production Research is very necessary. This subject research and the actual situation of cooperative production to carry out specific research. From a qualitative point of view, the possible risks of cooperative production are classified according to work attributes and work cycles, and the cooperative production is analyzed from the aspects of quality (equipment hardware and technology and key processes), material management, and basic management (plan and personnel). Quality risk sources and identification methods; from a quantitative perspective, combined with some possible risk sources proposed by qualitative analysis, a risk assessment model for cooperative production is constructed.

Li Hongxia (1965-), female, from Zunhua, Hebei, professor, doctoral tutor, currently the deputy dean of the Graduate School of Xi'an University of Science and Technology, her main research direction is safety management, corporate organizational behavior and management, safety economy and management, etc.

In order to better understand the current research status and development trends in the field of miner safety, with the help of CiteSpace analysis tools, 450 papers on the safety of miners from core publications and above in the CNKI database from 1992 to 2020 conduct bibliometric analysis. The results found that the main research content of current miner safety can be roughly summarized into six aspects, including coal mine production and management, miner safety quality, physical and psychological research methods. In the future, in the research of miner safety, it is necessary to adhere to the combination of theory and practice, and multi-disciplinary integrated research, to provide deeper and more comprehensive protection for miner safety.

With open-source programming languages, Processing and Arduino, this device realizes man-machine interaction. It has the characteristics of a wide application range and a strong visual expression. Designers and artists can make their ideas real through Processing and Arduino. In recent years, plastic waste can be seen in every corner of the world, which means that raising people's awareness of environmental protection is urgently needed. Childhood is a key period for developing environmental awareness. In this paper, a popular science device related to environmental protection, taking children as the main target group, was designed and produced. The device provides a combination of technology and art to educate children about environmental enlightenment through external hardware configurations (switches, light sensing and pressure sensing) and on-screen images that show polar bears living in large amounts of plastic pollution.

Ceramic 3D printing technology is using computer aid design techniques to model and produce ceramic products. The basic principle is through X, Y, Z axis of the ceramic 3D printer displacement layer to create the computer-designed 3D ceramic shape, and combined with traditional pottery hand-made techniques to finally complete it. Analysis and utilize the advantages of ceramic 3D printing technology, in order to inject new vitality into the traditional ceramic production industry.

The development strategy of the Yangtze River Economic Belt is an important strategy for China to cope with the new normal of economic development. This paper uses provincial panel data from the Information Network of the National Bureau of Statistics of China and the Development Research Centre of the State Council and conduct a panel data fixed effect regression model to empirically study the impact of technological innovation and industrial upgrading on economic development of 11 provinces in the Yangtze River economic belt. The results of full sample estimation show that technological innovation and industrial upgrading can promote economic growth at a significant level of 1%, but the impact of industrial upgrading is not obvious when they are considered separately. According to the results of sub-regional estimation, there is regional heterogeneity in the degree of effect of technological innovation and industrial upgrading on economy. Therefore, the Yangtze River Economic Belt should increase investment in science and technology, adjust industries according to the development orientation of provinces, pay more attention to the rationalization of industrial structure, rationally allocate resources among industries, and pay attention to the complementary development among regions.

Aiming at the problems of excessive work-in-process, unbalanced production, and unreasonable production methods in the assembly process of S company's low-voltage switch gear A cabinet, the value stream mapping technology was used to propose an improvement plan for the assembly process, and finally the Flexsim simulation technology The optimization effect of the scheme was tested, and the simulation results showed that the production balance rate, delivery cycle, number of products in process, and fatigue strength of workers have all been effectively improved, which also reduced production costs for the company.

In order to clarify the construction process and content of Safety Management System (SMS) and deepen the systematic construction of safety management of power enterprises in China, the paper analyses the current problems of SMS construction of domestic electric power enterprises. A construction method of SMS for China's electric enterprises is proposed based on systematic thinking and system modelling method. Taking State Grid Corporation of China (SGCC) as an example, the construction model of SMS is designed based on the Hall three dimensions structure model. According to China's national regulations and power industry safety management requirements, as well as practical experience of SGCC, ten core elements of the system were screened. The SMS operation cycle of Leadership-Responsibility-Plan-Do-Check-Act (LRPDCA) was designed based on the concept of Plan-Do-Check-Act (PDCA) and the characteristics of China's power safety management. The results have some guidance significance to continually promote safety management capacity of electric power enterprise in the new era.

In order to solve the security problems caused by network vulnerabilities, a web application vulnerability detection method based on machine learning is proposed to effectively prevent cross site scripting attacks of web applications and reduce the occurrence of network security incidents. Through the in-depth study of the existing security vulnerability detection technology, combined with the development process of machine learning security vulnerability detection technology, the requirements of security vulnerability detection model are analyzed in detail, and a cross site scripting security vulnerability detection model for web application is designed and implemented. Based on the existing network vulnerability detection technology and tools, the verification code identification function is added, which solves the problem that the data can be submitted to the server only by inputting the verification code. According to the server filtering rules, the network code bypassing the server filtering is constructed. Experimental results show that the model has a low rate of missed detection and false alarm, and the improved model is more efficient.

In this paper, the wireless communication system based on 802.11g protocol is studied, and a set of FH-OFDM system supporting adaptive transmission rate is designed. The maximum transmission rate can reach 6.75Mbps, which is much higher than 2Mbps transmission rate of traditional FHSS system. A new joint receiver and transmitter detection algorithm is proposed to reduce the error rate of FH position detection by 8.8dB∼11.68dB. The frequency hopping codebook is obtained by counting the bit error rate (BER) of each subcarrier. Selecting the subcarriers with lower bit error rate as frequency hopping points can further reduce the detection error rate by 4.27dB∼6.17dB, the bit error rate by 8.0778dB∼17.4121dB (in the case of 54/8Mbps), and 0.0533dB∼21.6286dB (in the case of 6/8Mbps). And in this paper, we choose the most appropriate transmission rate under different mobile speed and different SNR. By adaptive rate, we can get high performance and efficient equalization of the system.

In this paper we proposed an automatic splicing algorithm for regular cross-cut single-sided shreds of pure Chinese document, which was based on the characteristics of character and typesetting. The algorithm began with extracting the characteristics including character width, character height, line height and line spacing, after which the shreds meeting the typesetting requirement were spliced according to the edge similarity. Afterwards, the horizontal projection vectors of the shreds were corrected and the shreds were clustered by row according to horizontal projection similarity. A greedy algorithm with rejection strategy was then used to complete the intra-cluster horizontal splicing of shreds based on edge similarity. The experiment results revealed that the algorithm proposed in this paper had the characteristics of both high splicing speed and high accuracy.

In this work, a deep reinforcement learning (DRL) method is proposed to address the problem of real-time object tracking. The adopted framework in this paper is based on the 'Actor-Critic' tracker (ACT), since ACT only considers the scale change instead of regression object boundary, which cannot adapt the object size variation. To this end, the ACT method is improved by using a more reasonable action space, which contains a left-top and right-bottom corner coordinates. Precise shape estimation is given by regressing the variation of width and height, respectively. Furthermore, to speed up the whole training and tracking process, the Advantage Function (AF) is adopted, and its performance is compared with ACT, ACT with improved action space (IAS), and ACT with IAS and AF. This method is tested on the OTB100 dataset to validate its effectiveness.

Rust is an oxide produced by the reaction of metal and alloy surface with oxygen. It is a defect that greatly affects the quality of the workpiece. In order to solve the problem of rust segmentation more conveniently, a threshold segmentation algorithm based on cross extremum method is proposed, which determines the threshold of rust segmentation adaptively combining the global and local features of the image. Firstly, as the threshold is hard to determine in the small and medium area rust segmentation of a workpiece, the super red method and saturation method are combined to pre-process the image. Secondly, because the gray value of the rust spot area is the maximum value of the image neighborhood, the cross-extremum method is used to extract the characteristic points of the rust spot. Finally, the threshold of rust segmentation is determined according to the proposed feature points. Experiments show that the horizontal and vertical extremum method can complement each other to extract the feature points of rust as fully as possible. And the algorithm can extract some feature points of rust better for different workpieces in varying backgrounds. The detection accuracy of the algorithm reaches 96.3%, which has important signification of application.

Generative Adversarial Networks (GANs) are an innovative class of deep learning generative model that has been popular among academics recently. GANs are able to learn distributions on complex high-dimensional data which made it efficient in images and audio processing. Nevertheless, in the training of GANs, some major challenges exist namely mode collapse, non-convergence, and instability. In recent years, in order to overcome these challenges, researchers have proposed many variants of GANs by redesigning network architecture, changing the form of objective functions, and altering optimization algorithms. In this research, we conducted a comprehensive investigation on the progress of GANs design and optimization solutions. Finally, according to the classification method, we provided a problem-solving structure to solve conquer the GANs training challenges.

In order to effectively segment the visceral adipose tissue and help the doctors to rapidly diagnose the potential risks of metabolic syndrome, here we developed a deep learning-based method which is based on the U-net architecture for segmenting and measuring the visceral adipose tissue(VAT). And even no matter which orientation that the operator takes, the model can segment the visceral fat area and then use the appropriate outputs to compute the max thickness of VAT. One hundred and fourteen healthy volunteers were enrolled in this study. Ultrasound(US) was performed, and then the visceral adipose tissue was segmented and measured by the model that we use. We regard the distance behind the linea alba in the xiphoid process as the thickest visceral adipose tissue(VAT max). The dice score and accuracy are 3.46%, 96.44% respectively. In addition, compared with the manually outlined segmentation, the pearson correlation coefficient and the mean relative error (MRE) are R=0.9231 (P<0.001) and 10.12% in the measurement of the VAT max between original and output images. The auto-segmentation and measurement of visceral adipose tissue on ultrasound method demonstrate the accuracy of deep learning in segmentation and measurement of visceral adipose tissue.

With the development of the times and the progress of science and technology, the computerized accounting system can no longer meet the problem between the increasing human cost of most companies and the backward high-end talent pool. The transformation of the financial work system is imminent. The emergence of financial robots is undoubtedly an opportunity for the transformation of corporate financial work systems. It has attracted the attention of enterprises due to its unique process automation advantages. To mitigate above issue, this paper analyzes the development, application, and impact of financial robots in enterprises based on a brief introduction to the meaning and development of financial robots, and proposes relevant suggestions for introducing financial robots.

With the rapid increase of scientific papers, it is difficult for researchers to obtain appropriate references. Due to the black-box property of neural network, the existing content-based citation recommendation methods are mainly embedding text into a low-dimensional space, and citation links are not interpretable (e.g., paper A cites paper B because of topic X). To solve this problem, we propose a Weighted Citation Network with Citation Links Explanation (WCN-CLE) algorithm. Because citation links mainly depends on the content, we need to explore textual reasons for citation links. Firstly, we extract word fragments through citation network information for each citation link as its explanation. Secondly, we construct a weighted HIN network with two kinds of vertex (paper and author). The weight of each edge is calculated by its word fragments. Then the query document is linked to the HIN network through author information and word fragments similarity with candidate papers. Finally, each vertex's feature representation is obtained by network representation learning method, and recommend papers through vector similarity. Experimental results on two real-world datasets show that WCN-CLE has better performance due to its integration of structural information and semantic information.

As a promising generative modeling method, Generative Adversarial Networks are a deep-learning-based generative model, in which two networks, namely the generative network and the discriminant network, contest with each other in a game during which the generator "learns" and get trained to be able to fool the discriminator of believing a specific image is (superficially) authentic. It is widely used to generate pictures based on existing data. Based on the MNIST data set, this paper studies and analyzes the application of Generative Adversarial Networks in the generation of handwritten digital images. In addition, by studying the Deep Convolutional Generative Adversarial Network and the Conditional Generation Adversarial Network model, we compared the connections and differences between these three models in training the MNIST dataset to generate handwritten digits algorithms, and gave an optimization method for generating handwritten digits.

We consider the problem of resource elements allocation in a network environment with multiple users. Previous studies have done a lot of works using traditional methods in terms of bandwidth allocation, which is sufficient to serve for 4G network. However, it cannot be neglected to provide more efficient and intelligent scheduling policies in haste, due to growing demands on high resolution video and image transmission in 5G network. To fit the condition taking resource elements as scheduling unit in 5G network, we proposed deep Q network (DQN) algorithm based on the requirement of low time latency and high resource utilization rate to solve resource elements (RE) scheduling problem. Ultimately, we give out the optimal allocation scheme of resource elements (RE) for four users in fixed condition of time latency and resource utilization rate.

Researches on humor identification can advocate a better understanding of human language. Many studies focused on the categorical classification problem of humor, which is less sensible to the intermediate level funny content. Previous work captured the incongruity between words but not sentence snippets. In this paper, a novel method is proposed to exploit snippet-level incongruity features from different aspects, combined with the sentence snippets representations to predict funniness scores. The experiment result shows that this model outperforms most of the competitive methods.

This paper proposed an adaptive end-to-end codec based on multi-source fusion. Our codec adjusts the coding scheme according to the channel condition adaptively, selects deep recursion when the channel is good to obtain sufficient and effective information, and shallow recursion to ensure the transmission of key frames when the channel condition is poor. Besides, communication information is also imported into the codec through parallel branches to avoid confusion caused by multiple sources. We test on the DJI M210 UAV equipped with 5G network to verify the usability of the adaptive end-to-end codec both quantitatively and qualitatively. Under the same target detection level, our codec provides higher compression rate, and adjusts the compression rate to adapt to different conditions.

Geometrical accuracy of an X-ray Computed Tomography (CT) system is crucial to achieve high quality tomographic reconstructions. Methods to correct the resulting geometric artifacts have been comprehensively described in the past few years. Deep convolution neural network is increasingly used in CT imaging, which has a great potential in image feature learning and processing tasks. In this work, a geometric artifact correction method exploiting generative adversarial networks (GAN) is developed. The U-Net structure is employed as the generator of the network to extract the CT image features with geometric artifacts. The convolutional neural network (CNN) based on image block perception acts as a discriminator for the network, further constraining the optimization of the generator. The proposed method has been shown experimentally feasible for geometric artifacts correction in circular cone beam CT by performing more accurate feature extraction. The peak signal to noise ratio (PSNR) of the corrected phantom images increased by 11.812 on average, and the root mean square error (RMSE) decreased by 9.982 on average.

The research on the potential purchase behavior of users can help merchants develop better marketing strategies. At present, many research methods of online purchasing behavior are based on simple rule prediction, and the prediction results are not satisfactory. We design a hybrid model of Gradient Boosting Decision Tree and logistic regression to accurately predict the purchase behavior of users, which combines the association characteristics between users and commodities. Firstly, clustering algorithm and association rules are used to solve the problem of data imbalance and mine more potential related features. This scheme not only improves the processing efficiency of large data, but also solves the problem of user cold start. Secondly, we construct a scalable tree enhancement system (XGBoost) to train the initial feature set, which is a strong classifier composed of several weak classifiers. A new training set combines the new features with the original features through feature reconstruction, and a hybrid machine learning system is constructed by logistic regression (LR) model. Finally, the LR model is trained by the new training set. Compared with the existing schemes, the integrated decision tree model can train more sample sets with less resources. The experimental results show that the accuracy of the hybrid model is better than single model, and the F1_score is higher.

There are two ways to find the distribution of two-dimensional random variable function: distribution function method and convolution formula method, the distribution function method is too large; calculation is easy to make mistakes, the convolution formula method has limitations must be independent, Now let's introduce the generalized convolution formula, The generalized convolution formula does not require independent constraints and can simplify the operation.

The beam selection technology based on the discrete lens array (DLA) can greatly reduce the number of RF chains required by mmWave massive multiple-input multiple-out (MIMO) system. However, the performance of the existing beam selection schemes is on the premise of perfect channel state information (CSI). While in practical situation, accurate channel state information must be obtained by the base station before beam selection. In this paper, the channel estimation scheme based on support detection (SD) is used to estimate the state information of large-scale channel by utilizing the structural characteristics of beamspace. Then on the basis of group-iteration (GI) algorithm, we propose the group-iteration beam selection scheme based on estimated CSI. The simulation results show that this scheme can achieve the performance close to beam selection scheme based on perfect CSI, and compared with the traditional beam selection schemes, the proposed scheme also shows performance advantages.

The existing U-Net structure convolutional neural network is widely used in salient object detection, but the problem of target boundary blur is caused by the convolution pooling operation. In order to keep the edge structure of salient objects clear, a two-stream edge attention guided convolutional neural network (SEANet) method is proposed to strengthen the guided learning of edge features, mainly by guiding the edge features to compare the original image features and the depth features are enhanced to improve the accuracy of salient object detection. The loss is minimized by using consistent cross entropy loss and IOU loss to maximize the coincidence rate of the true value map and the predicted map, as well as the actual boundary and the predicted boundary. At the same time, the obtained prediction map is compared with the current mainstream nine models, and good results have been achieved. Four sets of experiments are carried out in the ablation experiment, and the experimental results also confirmed that the model has a great improvement effect on the performance of salient object detection.

Accurately mining the emotional information contained in product reviews is of great significance to product sales. Existing research on sentiment analysis of product reviews often ignores the importance of modelling aspect words and context separately. Therefore, this paper proposes a sentiment analysis model for product reviews based on RoBERTa-IAN. Firstly, the context and aspect words of product reviews are transformed into low dimensional vectors through the RoBERTa pre-training model. Then, the low dimensional vector is used as the input of Bi-GRU model to extract semantic features and get the hidden representation. Finally, the attention matrices of context and aspect words are obtained by using the Interactive Attention Networks (IAN), which are used as the input of sentiment classification layer to analyse and classify the sentiment polarity of product reviews. The experimental results on the real commodity Chinese dataset show that the accuracy of RoBERTa-IAN has reached more than 90%.

An improved sample entropy (SE) algorithm named weighted sample entropy(WSE) is proposed in this paper to optimize the problems existing in SE based on weighting. The principle of WSE algorithm is given, and the performance of it for dynamical change detection is analyzed using synthetic signal. The results show that WSE can accurately capture amplitude information and amplify the detection results of dynamical changes compared with SE and PE in the case of high complexity, but it requires a higher computation cost than the others. From all the analyses in this paper, we find that WSE has a better performance for dynamical change detection compared with the other three algorithms in abnormal amplitude.

At present, the transmission line in the field has a large potential safety hazard, and professionals often encounter difficulties in patrol inspection and distance measurement. In view of the current situation of safe distance between equipment in power grid construction, this paper proposes a long-span distance measurement method based on aircraft vision, which can automatically achieve high-precision distance measurement between equipment, and ensure personal safety, avoid equipment failure and other problems. Then the long-span ranging technology based on aircraft vision is analyzed, and the research and application of the given technology are discussed.

Human bones have different characteristics in different development stages, so the estimation of bone age can reflect the growth and development level of individuals relatively accurately. Bone age estimation aims to predict the biological age of children, which plays an important role on the diagnosis of some pediatric endocrine diseases. Tradition methods are carried out by doctors, and it is not effective in accuracy and speed. To this end, we proposed a deep-learning based method for bone age estimation. Based on the training set of more than 10000 X-ray images of hand bones from Radiological Society of North America (RSNA), this paper studies the processing, segmentation, feature extraction of X-ray hand bone images by using computer image processing and artificial intelligence learning methods, and uses convolution neural network to process samples and analyze them automatically. The main research work and achievements are as follows: (1) Pre-processing of X-ray hand bone image, unifying the size and cutting, reducing the image area without hand bone; (2) The gray-scale image is transformed into a three-channel image, and pre-processing by EfficientNet of ImageNet. Then convolution neural network is used to learn the features of X-ray hand bone image and evaluate it automatically. Finally, the network is evaluated by the minimum mean square error, so that the minimum mean square error is as close as possible to the minimum value. Through the neural network, the bone age from X-ray hand bone image can be quickly judged, and then it can be applied to clinical research.

COVID-19 sweeps the world with high infection and high death rates. It is essential for researchers to find an effective model to predict the trend of epidemic. With the good performances of several traditional models in predicting and analyzing previous epidemics, we compare several popular machine learning methods, including multi polynomial regression, logistic growth model and Long Short-Term Memory (LSTM) in epidemic prediction. We use least squares method for feature selection to determine the most relevant features and we also scale the data according to different experiment environments. We measure the accuracy using Mean Squared Error (MSE) and R². We conclude that the LSTM model is the most effective model among all the competitors with the highest R² (R² = 0.97). We find that LSTM model is the most effective model among all the competitors. Our study gives a good example of feature and model selection for epidemic prediction and attempts to make a significant contribution to the government and hospital to supply the public resources and provide drugs to handle the incoming issues.

A single shallow learning algorithm cannot fit the characteristics of high-voltage reactors well. Aiming at the above problems, this paper uses the error back propagation neural network and the particle swarm algorithm optimized by adaptive inertia weight to optimize the combined prediction model B for data training, verification and testing are carried out to achieve the purpose of effectively reducing the manufacturing cost of high-voltage reactors. Through experimental verification, the maximum error between the predicted value and the true value is 2.9%, and the minimum error is 0.05%. This provides certain technical support and inspiration for future devices such as optimizing high-voltage reactors.

The related regions with convolutional neural networks (R-CNN) models have been widely used in the field of object detection. Faster R-CNN significantly improves the overall performance by adding RPN, especially in terms of detection speed. However, the application of different pre-training models will result in a great difference in the performance of Faster R-CNN. This paper analyzed the performance of Faster R-CNN models based on different pre-training models and conducted a comprehensive evaluation of the performance of Faster R-CNN. The experimental results showed the accuracy and detection speed of R-CNN, fast R-CNN and faster R-CNN based on three different data sets. They can objectively and comprehensively evaluate the performance of R-CNN, fast R-CNN, and faster R-CNN.

Mobile robots can be roughly categorized in wheeled type, tracked-skidded type, legged type, trunked type, etc. Among them, the wheeled mobile robot can move on the flat road, but with poor performance when crossing an obstacle; the legged type mobile robot has a slow motion speed and poor bearing capacity; the tracked-skidded mobile robot has a large weight; while the trunked mobile robot has limited use occasions. In order to overcome these shortcomings, this paper studies the wheel-legged structure and distributes the performance of various commonly used structures. On the other hand, by analyzing the newly emerged structure, the impact of the organization's own shortcomings on performance is hoped to be reduced. By comparing the structure and distribution of different leg configurations, the advantages of the locomotion performance of each type of wheel-legged mobile robot and the development trend of the wheel-legged mobile robot in the future are analyzed.

Complexity of mechanical production environment has brought a certain degree of interference to human target recognition, so it is necessary to preprocess the target video sequence before target detection. The advantages and disadvantages of various detection methods are compared. Research results show that the inter frame difference method has strong adaptability and the target contour is discontinuous. The optical flow method is suitable for the case of lens motion, which is complex in calculation and poor in noise resistance. ViBe algorithm has good effect in obtaining the target contour, clear edge and less holes, and the operation efficiency is not high. The background subtraction method basically retains the features needed by the moving foreground target behavior recognition and detection, and the processing speed is very fast, which can meet the requirements of real-time processing.

Aiming at the safety risk of storage tank in port area, the control measures of tank safety risk are analysed, and the quantitative risk correction method based on safety measures is studied to reduce the risk of tank accident in port area.

In recent years, many advanced computer technologies have appeared on the market, for example, artificial intelligence. With the maturity of artificial intelligence technology, the transformation of intelligent robots are also more perfect. The difficulty that we have always wanted to break through is to install positioning navigation on mobile robots. Through continuous efforts, the navigation and positioning technology on the robot can now be used proficiently in a known environment, but we need further research when using it in an unknown environment. In the navigation and positioning technology, it introduced simultaneous positioning and map construction algorithms (SLAM, simultaneous localization and mapping). The algorithm can learn about the surrounding environment information through the sensors carried by the robot. If this algorithm can be skillfully applied to navigation and positioning technology, I believe it will be helpful for the application of positioning and navigation technology in unknown environments.

At present, the cyber security situation has turned increasingly serious across the globe. The cyberspace has become the fifth domain of various nations following the four domains of land, sea, sky and outer space. It is gradually obvious for the cyber attacks to collectivize and nationalize. Under this circumstance, as the first line of defense in maintaining cyber security defense, the control on the vulnerabilities is extremely significant. In consideration of the problem of low discovering efficiency in current vulnerability fuzzy testing discovery methods, this paper intends to explore the binary program vulnerability feature extraction and vulnerability discovery guiding strategy generation methods. Then, the extraction of the program control flow and data flow information and slicing operation are obtained according to program logic. Depending on a potential assessment model of binary program vulnerabilities based on the main hierarchy analysis method, the paper achieves a reasonable assessment of the potential of vulnerabilities in different slices of the program. Finally, in accordance with the evaluation results, a generation method for the vulnerability discovery path is proposed to improve the efficiency of binary program vulnerability discovery.

The cost of airborne software development is very high, and the functions implemented by airborne software are becoming more and more complex. Using the COTS software has become a trend in case of saving cost. The COTS software brings conveniency, meanwhile raise problems when showing compliance to airworthiness requirements. This paper researches the design assurance method of COTS software, summaries the certification risks and propose a feasibility analysis of using COTS software, thus providing reference for airborne systems designs.

In the process of successful establishment and development of service ecosystem organization, manufacturing companies not only pay attention to their own resources, business models and competitive advantages, but also pay attention to the positioning in the manufacturing service ecosystem and the symbiosis with other members of the system to achieve complementarity and symbiotic development of resources among members. Based on this, we take the manufacturing service ecosystem in the transformation and upgrading of the manufacturing enterprises in China as the research object, and comprehensively uses the theory of evolutionary game and ecosystem to study the symbiotic evolutionary stability of the manufacturing service ecosystem. Through simulation, we find that integrator is more sensitive to the change of rights than supplier. Our conclusions have certain reference for manufacturing companies to improve the collaborative operation efficiency of manufacturing service ecosystem.

In the process of civil aircraft design and development, interface definition is a very important and complex process element. In this paper, a rule element definition method based on civil aircraft interface development process is proposed, and the rule element definition to support the interface development process is given. Rule elements provide useful input on structure and content for civil aircraft interface definition activities, especially for low-level interface definition activities. All levels of an interface definition can be related to rule elements, but lower-level interface definitions have the greatest amount of data, and most of the definitions are done in lower-level interfaces. For the design activities of the stakeholders in the low-level interface development process, these rule elements are equivalent to the formulation of multi-dimensiona and unified standards and requirements, so that the stakeholders of interface definition (main manufacturer and suppliers at all levels) can carry out coordinated interface development activities under such standards and requirements.

Arctic shipping can bring significantly potential business profit, which has become a focal point at present. However, compared to the enormous risk, related research on the safety and security issues are not sufficient. In the essay, the natural factors, such as weather, ice are summarized, and the facility insufficiency and human factors aroused by low temperature and severe ice conditions are analyzed. Besides that, the requirement of international conventions has already proposed the basic technical demand for the ship facilities and lifeboat in Arctic area. In the end, the suggestions for safety and security, providing better equipment and facilities, enhancing crew competence, decent preparations before navigation are proposed.

In order to solve the problems of correctness and timeliness of requirement information transmission, effectiveness of requirement collaboration, management and control under the Main Manufacturer-Supplier working mode for civil aircraft, a distributed collaborative management method of supplier requirements in the whole life cycle is proposed in this paper. The definition of this method of supplier requirement distributed collaboration can carry out multi-version, multi-authority and cross regional requirement data exchange for different suppliers at different development phases, and monitor the status of data exchange based on the key elements of supplier requirement management, so as to ensure the requirement collaboration and supplier monitoring management between main manufacturer and supplier.

The aircraft-level high-lift system architecture of civil aircraft includes functional architecture, logical architecture and physical architecture. The system architecture describes the hierarchical relationship, cross-linking relationship and information interaction between the function, logic and physics of the system. This article is based on the SysML modeling language, using the MagicDraw modeling tool to transform the traditional design work based on experience and documents into model-based architecture design work, using the association and traceability of the model to ensure the integrity of the information in the design process. This article establishes the design relationship between its functional, logical and physical layers of the high-lift system to verify the efficiency of this modeling method.

The heating furnace has dynamic fluctuations of hazard factors in the actual circumstance and is the core equipment of hydrocracking unit, which is the main equipment of petrochemical enterprises. Owing to the limitation of conventional risk analysis approaches, a dynamic risk analysis approach based on dynamic Bow-tie model(short for DBT) and dynamic Bayesian network(short for DBN) was proposed. Taking the failure accident of oil and gas leakage of heating furnace as a case study, the timing sequence of equipment failure of heating furnace is analyzed first, the risk evolution process is obtained, and the dynamic tie model of accident is established. By mapping the dynamic tie model to the dynamic Bayesian network, the risk factors of oil and gas leakage in heating furnace are deduced in two-way reasoning. The dynamic risk trend of heating furnace was predicted eventually, the probability of oil and gas leakage caused by heating furnace is about 1.26E-07 after two year. Through diagnostic reasoning, 6 key factors and 3 accident paths were determined. The results are in accordance with the actual situation, proving that DBT-DBN model can effectively reflect the dynamic characteristics of accident scenarios and provide reference for the risk assessment of petrochemical enterprises.

As a part of software structural coverage, the data coupling and control coupling analysis help to evaluate the implementation of requirements and the code quality to some extent. This paper researches the data and control coupling definitions, analysis methods and the differences of requirements between DO-178B and DO-178C, thus providing reference for airborne software verification.

From the perspective of patients and medical insurance agencies, the paper used the fuzzy rough set method to select the indexes and established the service quality evaluation index system of medical institutions. Using Dempster-Shafer evidence theory to synthesize the subjective weights and the objective weights, the combined weights of the evaluation indexes were obtained. Combining grey relational analysis and Dempster-Shafer evidence theory to establish a regional medical service level evaluation model in an asymmetric environment. The model can accurately identify the quality of medical service in various regions. We can compare and analyze the service quality of medical institutions in various regions from different dimensions, and provide a theoretical basis for each region to formulate policies related to the rational allocation of medical resources.

Architecting commercial aircraft has been always challenging due to the complexity of the product. A Model-Based Systems Engineering (MBSE) method was employed in this paper to handle this challenge, and a widely used general-purpose systems modeling language - SysML - was adopted to support the architecting work. To ensure compatibility between the modeling language and the specific aircraft architecting methodology, we explored construction and application methods of a Domain Specific Language. A new Domain Specific Language was created, in which several extensions to SysML were made, including Function & Components, Port, Exchanged Item and Relation. Benefiting from this new Domain Specific Language, an overall aircraft architecture model was precisely constructed.

In modern economic and social development, rail transit is the lifeblood of new urbanization and high-quality development in urbanized areas. As a core element of economic development, logistics plays an increasingly obvious role in promoting economic transformation and high-quality development. Under the normalization of epidemic prevention and control, expand the comprehensive logistics distribution system and focus on the deep integration of urban rail transit in densely populated cities. It is very necessary to promote the extension of the supply chain and improve the quality and efficiency of urban logistics distribution.

The outer wing and the central wing of the aircraft are butted at the wing root through rib 1. The load transfer in the connection zone of the wing root is complex. In the plug-in butt form, additional bending moment will be generated, which is the most prone to damage. Based on the analysis of the force transmission in the butt joint area of the wing root, combined with the test data, this paper proves that the main failure mode of the composite central wing lower panel is the delamination failure of the stringer web at the end, and suggestions for optimal design are proposed.

Under the background of China's economy gradually entering the new normal, transformation and upgrading has become the only way for manufacturing industry to improve market competitiveness and sustainable development. In order to promote the transformation and upgrading of manufacturing enterprises, this paper takes 96 literatures on the influencing factors of the transformation and upgrading of manufacturing industry as the research object, and conducts a systematic study based on the grounded theory. Through the collection, collation and coding of literature, many categories are formed. After analysis, it is concluded that the influencing factors of the transformation and upgrading of manufacturing industry include five aspects, namely, the willingness of transformation and upgrading, the ability of transformation and upgrading, the environment of transformation and upgrading, the resources of transformation and upgrading and the development trend. Among them, the willingness and the ability of transformation and upgrading belong to the subjective category. Technological innovation ability is the most important factor affecting the transformation and upgrading of manufacturing industry. Accordingly, this paper puts forward relevant strategies and suggestions to promote the transformation and upgrading of manufacturing industry.

"System engineering" is one of the most important academic terms contemporarily. It is to make reasonable use of all resource conditions, optimize the configuration of the structure, and conduct process management to scientifically use resources. Through the practical process of research-based sketching and design transformation, the article uses plant ecological structure as the original experimental sample, and based on system thinking, explains the deconstruction, reorganization and final visual presentation of graphics from multiple dimensions of overall, three-dimensional, structural and dynamic. Explore and put forward a comprehensive hierarchical system model of "objective-deconstruction-reorganization-optimization and integration" to optimize graphic design behavior patterns and improve the effectiveness of design strategies.

With a large number of important and commemorative historical buildings in the maintenance and reinforcement of the increasingly stringent requirements, the role of inorganic–matrix EBR reinforcement system in the existing masonry structures reinforcement is also increasingly concerned. In this paper, the mechanical properties of textile, mortar and composite materials are compared and analyzed, and the influence of different parameters on the test results is discussed. The results exhibit that the strength range of each material is obvious. For textiles, the tensile strength ranges from 1120 MPa to 5800 MPa, and the elastic modulus ranges from 55.6 GPa to 276.6 GPa. For mortar, the compressive strength is from 12.3 MPa to 80 MPa, and the elastic modulus is from 6 GPa to 39 GPa. In addition, the cracking process of composite specimens can be divided into three stages: un-cracked, crack development and cracked. All kinds of test results can not be compared with the same standard and conditions, due to the significant differences in specimen designs and loading methods, which provides the research direction for the next stage of test and theoretical analysis.

In view of the characteristics of the production of precast concrete components and the cutting of rebar, from a production point of view, the introduction of BPI (business process improvement) management ideas has been introduced to optimize the rebar cutting process of precast concrete components. A multi-objective prefabricated concrete component rebar cutting model is established and an improved genetic particle swarm hybrid algorithm is proposed. It is proved by examples that the model reduces the number of raw materials and the number of rebar cuts by rationally arranging the rebar cutting combination plan, and the optimization of the prefabricated concrete components rebar process can rationally use the rebar remaining material and improve the utilization rate of rebar.

In order to obtain the time to failure of adjacent tanks under multi-thermal radiation conditions, the thermal response of gasoline tanks under multi-thermal radiation was studied. The heat radiation values of the adjacent tanks under pool fire were calculated. The temperature distribution and the corresponding stress of the adjacent tank were studied with ANSYS-Workbench software under the influence of single heat radiation and multi-thermal radiation. According to the physics properties of tank steel Q235B, the time to failure (ttf) of the tank was analysed. The results show that the temperature rise rate affected by the coupling effect of multi-thermal radiation is significantly higher than that affected by single heat radiation; when the storage tank is affected by the single heat radiation, the time to failure is about 22min; when the storage tank is affected by the coupling effect of multi-thermal radiation, the time to failure is about 15min.

Polyethylene pipes have been widely used in the urban gas industry. However, there are few researches on the safety performance evaluation of polyethylene pipes in use at home and abroad, which makes it impossible to carry out Fitness-For-Service evaluation of defects found in the inspection, and thus cannot accurately judge the safety status of defective polyethylene pipes. In this paper, two typical polyethylene pipe materials, PE80 and PE100 are selected. Based on the tensile performance test, the constitutive model is established. Using the method of finite element simulation, the ultimate pressure bearing capacity of 9 defect models was analyzed and studied. And the relationship between defect size and pressure bearing capacity was initially determined. Provides a foundation for subsequent engineering applications.

The supply chain performance of prefabricated construction plays an important role in the development of prefabricated construction. In this paper, in order to improve the supply chain performance of prefabricated construction and identify its factors, the indicators system for the supply chain performance of prefabricated construction based on the stakeholders' perspective on the three dimensions of schedule, cost, and quality was built. Based on the indicators system, the system dynamics simulation model of supply chain performance of prefabricated construction was established. This study provides a dynamic performance management tool for prefabricated construction supply chain stakeholders.

In the modern civil aircraft, the number and complexity of airborne software are increasing rapidly. The software may change at any stage of the development process and the change process should be under control. This article analyzes the basic unit of software change control and the configuration identification used for airborne software, then introduces a process and strategies of software change control for civil aircraft.

Although the software tool qualification should be based on the specific airborne software, it is impractical to develop the software tool for one software program. Most software tools, especially for the COTS tools, would be reused in multiple software programs. This paper capture and analyzes the requirements for the previously qualified software tool in the related standards and certification guidance, and proposes a strategy and framework for the reusing the previously qualified software tool to meet the requirements of different standards and certification authority.

In view of the characteristics, status and present situation of small and micro-sized enterprises and small and micro-sized mines, this paper investigates the common safety problems or hidden dangers of small and micro-sized mining enterprises on the basis of practice, analyzes their causes, discusses the current safety management system and operation mechanism of small and micro-sized mines, and puts forward various effective ways to improve the safety management of small and micro-sized mines. It has strong practical and theoretical significance for mine safety.

The traditional cost management mode of China is used to make the cost of the production, and the cost management of the "commodity" is the cost management of the production process, the cost management of the product development and design of the pre-production, which is increasingly difficult to adapt to the needs of the modern social and economic development, has greatly hindered the improvement of the competition capacity of the enterprise commodity market. Based on the operation cost method, the paper will provide an example for the application of the operating cost method in manufacturing. In addition, the research on the final implementation of cost management is not very good. Combined with the methods of reducing the cost of industrial engineering, this paper continuously improves and controls the problems in the company's operating cost management, and verifies the feasibility through empirical research and a large number of experiments.

There is a growing interest in using additive manufacturing to produce smart structures, which have the capability to respond to thermal and mechanical stimulation. In this report, Selective Laser Melting (SLM) is used to build a thin-walled titanium alloy structure. The part creating a ultra lightweight is 2.15g, while its strenght achieve at 132.4MPa. The stages of compression process was analyzed in detail, meanwhile the deformation and failure mechanism of the thin-walled part was described.

Automatic CAD design is difficult in construction field, since the construction environments vary case by case. Although the steps are similar, each creation is very time-consuming. Therefore, in this research a method that could automatically generate 3D CAD models and 2D drawings from the onsite measured geomorphological features is proposed to solve the complicated and similar process. This proposed method was implemented through the API (Application Programmable Interface) of CAD system and has been verified in practice that labor costs and time costs have been greatly reduced. In addition, mixed reality is applied to verify the assembly results, and an effective and realistic experience is provided to the user and the customer.

In today's world, marine pollution is becoming more and more serious, and the protection of marine ecosystems and endangered marine organisms has become an important international issue. In 2012, the blue whale was officially listed on the IUCN Red List of Threatened Species. As the top predator in the ocean, blue whales are seeing a drastically decreasing population, which will break the ecological balance of the region. The interlocked food chain in the ocean means it would eventually lead to the forming of an "ocean desert". Therefore, it is urgent to start the popularization of science to protect the endangered marine organisms for the public. Based on human-computer interaction and sound visualization, the author designs an interactive installation with the theme of "The Last Whale Sound", aiming to increase the public's attention to endangered marine organisms, to arouse people's self-reflection and environmental awareness, and to make the public consciously participate in protecting marine ecosystems.

Aircraft product development process is quite complex, involving many stakeholders. Therefore, aircraft owner manufacturers need to fully capture the needs of stakeholders at the beginning of design, establish analysis methods of stakeholder's needs, and transform "customer voice" into aircraft development requirements, so as to ensure that the aircraft products can meet the needs of stakeholders. This paper introduces stakeholder's needs analysis methods such as quality function deployment (QFD), analytic hierarchy process (AHP), and analytic network process (ANP). This paper establishes a Stakeholder's needs analysis model, which can be used to carry out stakeholder's needs analysis of civil aircraft.

In recent years, the international aviation organization has implemented new performance-based navigation technology and issued the operational requirements of required navigation performance (RNP). It involves the cross-linking of multiple complex systems, which puts forward higher requirements for aircraft development and management. In this paper, the development process of RNP functional architecture model is defined by using magic grid software. The modeling is completed from three aspects: problem domain, solution domain and implementation domain. The decomposition and implementation of complex functions are completed from top to bottom. This method can effectively solve the complexity problem of civil aircraft system, and lay the foundation for demand-based model forward design.

Based on the review of the existing research, this paper summarizes the application of knowledge services in supply chain. Using CiteSpace software for visualization analysis, it can be found that there is a lack of cooperation between research authors and institutions of agricultural products integrated supply chain, but the research content is relatively relevant. "Internet+" and "fresh agricultural products supply chain" will be the frontier hot spots of agricultural products integrated supply chain in the future. Therefore, enterprises need to combine some big data technology to further integrate and optimize the agricultural product supply chain.

Since the concept of quantum entanglement was proposed, quantum entanglement has been paid more and more attention. Especially in recent years, with the interdisciplinary and interdisciplinary development of quantum informatics, the general definition, physical properties and entanglement measurement of quantum entanglement have been comprehensively and deeply studied. Due to its nonlocality, quantum entanglement has been widely used in quantum information, especially in quantum communication. In view of the application status of quantum entanglement in quantum communication, this paper presents a survey of quantum entanglement and its application in quantum communication. In this paper, 230 staff members from 15 Chinese companies involved in quantum communication are selected as the investigation samples, and the proportion of quantum entanglement in the application is obtained, which can promote the quantum communication. The capacity, speed and security of quantum entanglement and ordinary quantum in the application of quantum communication are compared. Through the analysis, it is concluded that each index of quantum entanglement is conducive to the development of quantum communication. Quantum entanglement provides a technical guarantee for the research of quantum communication.

Online Classes has been widely popularized since corona virus doesn't allow students to have face-to-face classes. However, due to comfortable environment that students have, students may get sleepy in class and cannot help with their fatigue. So students need a supervised environment to alarm them not to fall asleep. So in this article, a real-time method to detect whether the student in front of the camera is sleeping, based on the demo of students in online classes, is introduced. This article focus on detection of states of eyes of students. It uses Adaboost as the training method to detect the face of the student, uses Canny to process the picture with human face, and calculates the area of the exposed eyeball and the radius of the eye. Finally, PERCLOS is used as the criterion to determine whether the student is sleeping in class. The method in this article has high detection rate on faces and eyes, which is respectively 99.7% and 98.5% for image set that made by the author. It can be applied on online classes and shows good results in detecting the fatigue of people in practical.

Because the power CPS network with community structure has the unique structural characteristics of internal nodes and boundary nodes, the boundary nodes are more likely to cause cascading failures when attacked, which plays an extremely important role in network risk fault propagation. In this paper, the Boundary node is regarded as an important immune object, and these key node targets are quickly identified to take immune measures. A Boundary center-point-search (BCS) algorithm is proposed. According to the disadvantage of the original risk propagation model which ignores the fact that normal nodes can get immunity through artificial means, a risk propagation model is proposed. Compared with other local immune algorithms, the proposed BCS immune algorithm not only can accurately identify the boundary center points between communities, but also has a faster computational speed, which can provide reference for risk propagation control and prevention in large-scale complex networks with incomplete structural information.

As a platform for sharing and communication, Weibo can produce a large amount of short text data every day. Because the Weibo text is short and the text features are sparse, it contains less information, serious colloquialization and weak anti-noise ability, so it is difficult to achieve the desired results by using the traditional text classification method to classify Weibo text. This paper proposes a Weibo text classification method based on knowledge graph, which uses knowledge graph to expand text semantics and enrich text features on the basis of the original Weibo text. Then the TextCNN model is used to classify Weibo texts with extended semantics. The experimental results show that the classification accuracy of this method is better than that of Weibo text classification directly.

The research on emotional response generation is one of the directions in dialog system research. However, most of the researches on emotional dialog generation focus on the expression of expected emotion in single round dialogue. In this paper, we propose a method to model the emotion in multi-turn dialogue. Our model bases on a proposed model: Emotional Chatting Machine, and solves the problem of ECM model that which emotion should be used to the response. We compared the effects of various models and finally chose LSTM to model emotions in multi-turn dialogue. Experiments show that the performance of the dialogue system can be effectively improved by modelling emotion in multi-turn dialogue.

Speaker recognition is a technology that recognizes a person's identity through a person's voice. Speaker recognition is one of the important research topics in the field of biometrics. This article first introduces the background and research significance of speaker recognition; secondly, it briefly introduces speaker recognition technology; again introduces the method research used in speaker recognition; and finally analyzes the research difficulties and challenges of speaker recognition.

Since the 21st century, more and more attention has been paid to the education of college students in our country, and the curriculum reform has been actively carried out. It is necessary to cultivate students' core quality and improve their ability to adapt to society and self-learning. Under the new situation, universities should constantly reform, update their teaching concepts, adopt the latest teaching equipment and ideas, readjust their teaching methods and plans, and provide reference for the future. According to relevant research, the administrative system of universities affects the cultivation of students' quality and ability to a certain extent. This paper combines AI and corporate governance to explore the influence of the administrative mechanism on the operation of universities under the new situation, and puts forward some solutions and suggestions.

This study aims to generate more topic-related and coherence essays based on user-defined topic words. Existing research generates essays without considering the semantic information from the corpus. However, the corpus contains statistical relationships of words which can be used to guide the model to generate more coherent and fluent essays. To fill this gap, we propose a corpus-based topic-to-essay generation model (C-TEG). We elaborately devise a background network based on the co-occurrence relationships of words from the corpus. The empirical results demonstrate that our approach has achieved 4.14 average score in subjective evaluation and a better BLEU-2 score, which shows that our model is able to generate more topic-related and coherent text than existing models.

Improving the quality of experience (QoE) of video streaming is a significant task in the wireless network scenario. Buffer starvation in the transmission process will cause playback freeze, and a certain number of packets must be prefetched before the service restarts. Taking into account the shortcomings of buffer in video streaming services, this paper proposes a deep learning-based starvation probability calculation model and a reinforcement learning-based packet prefetching model. The deep learning approach extracts the correlation between different timing inputs through the recurrent neural network module to return an explicit result and the precise distribution of the number of buffer starvation. The reinforcement learning approach leverages a better trade-off between start-up/rebuffering delay and buffer starvation by adjusting the packet prefetching strategy, so that the long-term objective quality of experience (QoE) of the video stream is optimized. Our framework can be applied to actual scenarios including finite video streaming and long video streaming transmission.

The application of distributed system in the Internet of things makes the system control completely networked, realizes the connection of point-to-point and star data transmission control, makes the system highly scalable, convenient for centralized management, and is suitable for multi-functional Internet of things applications. In the distributed architecture, data is distributed on multiple servers. A distributed database is composed of multiple logically related databases distributed on the Internet of things. Each node on the network has independent processing ability, so it is necessary to find a suitable distributed system solution between code and database.

In recent years, facial expression recognition has played an important role in the field of human-computer interaction, and the application of deep learning technology has enabled it to develop more rapidly. In this paper, we create a lightweight network model for real-time emotion classification to recognize student facial expressions. The system includes: using Haar cascade for face detection, combined with the idea of Xception to propose a lightweight CNN network model, using pre-activation in the residual block to optimize the model and reduce the impact of overfitting. The experimental results on the FER2013 database show that our model has a better expression classification effect than other state-of-the-art methods. Besides, our model uses fewer parameters, which reduces the complexity of network training. When real-time emotional recognition of students, the system can help teachers adjust their teaching methods according to the emotional state of students.

Meteorological data are important information data in China, and have made tremendous contributions to disaster prevention and disaster reduction in China. However, the current meteorological system has a long-term dependence on foreign hardware and software platforms, showing obvious security problems. It is an urgent task to complete the domestication and performance improvement of the meteorological system. At the same time, as the black box architecture of flash memory devices greatly hinders the co-optimization of software and hardware, open-channel flash memory has begun to attract attention as a new type of flash memory architecture. In this article, we have adopted new excellent domestic products including Phytium chips, Starblaze open-channel flash memory devices, and Kylin operating systems as components of a new domestic software and hardware platform. The object storage system of meteorological data has been implemented in a real environment by Ceph. We provides a built solution and optimizes the parameters. Based on the comparison with the commercial platform storage system, the problems and challenges faced by the meteorological data distributed object storage system in autonomy are discussed. Finally, this article looks forward to the distributed storage of meteorological data under domestic platforms.

This article uses CiteSpace software to perform a visual analysis of the knowledge graph of the literature about the direction of big data management in the CNKI database using the method of document measurement. The analysis content includes: paper attention, author of the article, hot keywords, and the organization of the article, and big data is found. The hot research of management can be summarized into four categories: education management, information management, enterprise management and government management. The domestic scholars' research status on big data management is obtained, and the development trend of big data management research is revealed, with a view to providing reference for the research and development of big data management in various subject areas.

As an important branch of AI (Artificial Intelligence), speaker recogniti-on technology has developed rapidly in recent years. At present, speaker recognition technologies based on traditional methods and deep learning methods are very mature and have achieved good results. As a key technology in human-machine voice intera-ction, speaker recognition has changed human life in many ways. However, there is almost no research on speaker recognition in Tibetan. This paper uses Gaussian mix-ture model and deep neural network TDNN as the theoretical basis. On the Tibetan language corpus, the speaker identification on the ivector of GMM-UBM and the speaker identification on the xvector of TDNN are performed respectively, and the experimental results are compared. The experimental system is divided into prepro-cessing, feature extraction, model training and decision.

Informational education is a new requirement for education in the information age and it is an inevitable trend in the development of education modernization. In the rapid development of the Internet, the use of cloud computing technology can play a greater role in the construction of smart campuses. It can not only improve the school's digital service management level, but also promote the progress of education. This article briefly introduces the concept and necessity of smart campus construction, analyzes the role of big data technology in the construction of smart campus, and proposes smart campus construction measures based on big data.

In order to solve the problem of sidelobe error caused by probe pattern in current antenna planar near-field and far-field transformation, a new near-field and far-field transform algorithm with probe compensation is studied by combining the advantages of traditional pattern development and edge current approximation method. By measuring the array antenna in a microwave anechoic chamber, the traditional mode expansion method is compared with the hybrid algorithm. The results show that the measurement accuracy of this method is 1.445dB higher than that of the traditional method, and the calculation efficiency is significantly improved, which proves the effectiveness and engineering practicability of this method.

In order to reduce the problem of mismatch between high-performance processors and DRAM speeds, current processors have added a cache structure, but the low cache hit rate also seriously affects the actual performance of the program. Data prefetching technology can alleviate the problems of memory access latency and low hit rate caused by the speed difference between high-performance processors and DRAM. Based on the LLVM open source compiler, this article first implements the data prefetch module on the Shenwei platform. This paper improves the prefetch distance algorithm, proposes a new prefetch scheduling algorithm, introduces a cost model to evaluate the prefetch revenue, and accurately determines the insertion timing of the prefetch instruction to improve the cache hit rate. SPEC2006 performance test results show that after optimization, Shenwei 1621 processor single-core can achieve a maximum performance improvement of 50%, and an average performance improvement of 11%.

The paper firstly explains the importance and research significance of image description generation and makes a brief research summary on the generation of various algorithms for image description; it further introduces the principle and the algorithm process of the LSTM deep learning algorithm in details, specifically describes establishing steps and the model parameter selection process of the LSTM deep learning algorithm of the Microsoft COCO dataset, implements the deep learning models by employing Python, and finally applies the trained LSTM deep learning model to image description generation by using the Microsoft COCO dataset as an instance.

At present, the development of video action recognition is very rapid in many fields, such as video understanding, intelligent monitoring, and human-computer interaction. However, there are some challenges in the development of action recognition, and researchers have tried to put forward some explorations. Convolutional neural network (CNN) is applied to action recognition, which improves the performance of action recognition. It is divided into 3 methods in this paper. In addition, C3D, Two-stream and I3D, three classic CNN algorithms, are reproduced. And their recognition rates are 72%, 78.0% and 97.6% respectively on the UCF101 dataset.

With the development of network technology, the scope of information search and the ability of data mining are also increasing. On this basis, a WebCrawler technology is proposed. WebCrawler technology plays a very important role in improving the ability of data information search. This article mainly analyzes the working principle of WebCrawler technology and the principle of Web information search and data mining, and focuses on the application in information search.

GCN is an excellent single-modal machine learning model for processing non-Euclidean data, such as skeleton data. Conventional GCN only uses single-modal data as its input. When we have multiple modal data, the application of GCN will be limited. To utilize multi-modal data advantages, we propose a method to combine GCN and LSTM to extract skeleton feature and acceleration feature, respectively. To test our method, we did a lot of experiments on Cooking Activity Dataset.

Firstly, this paper expounds the concepts of visual communication design and computer graphic image design, analyze the relationship between visual communication design and computer graphics design in detail, including the similarities and differences between them. Then, it analyzes the requirements of visual communication design and computer graphic image design, including design ideas, skills and concepts. Then, it elaborates the practical applications of computer graphic image design and visual communication design, including multimedia products, artistic creation and design, and commodity appearance design. Finally, it analyzes the combination of visual communication design and computer graphics image design. It is hoped that more people in the industry will pay attention to visual communication design and computer graphic image design, which will help promote the further development of computers.

In the past two decades, the field of quantum computing has developed rapidly, is gradually able to solve some difficult classical problems. With the development of the facilities for Noisy Intermediate-Scale Quantum(NISQ), the NISQ technology addresses the existing Quantum programming languages or frameworks, suggesting that they cannot fully meet the requirements of Quantum computers. In recent years, some quantum programming languages or frameworks have emerged which claim to support the classical computing algorithms of heterogeneous quantum. However, the code is too cumbersome and redundant, and it is difficult to map the code to heterogeneous quantum classical computation while meeting the requirements of timing. Therefore, a modular programming framework of heterogeneous quantum classical computing algorithm with NISQ function is proposed. The framework optimizes the quantum code and retains the quantum classical interactions that satisfy the temporal constraints.

LLVM converts the loop to LCSSA during the loop transformation process. The loop of LCSSA has better locality and can facilitate other optimizations of the loop, but at this stage LLVM does not support some special LCSSA loops in the automatic vectorization process, in the case of external use of non-array and non-inductive variable instructions in the loop, automatic vectorization cannot be performed, which makes the compiler lose some optimization opportunities for automatic vectorization. In response to this problem in LLVM, this paper proposes an algorithm to reconstruct PHI nodes. By adding new basic blocks, reconstructing the value of PHI nodes to eliminate the influence of external use in the loop on automatic vectorization, so that the loop can be automatically vectorized. Improve LLVM's automatic vectorization capabilities. Through the test on the TSVC test set, the vectorized recognition rate has increased by 23%.

In this paper, by studying the laser marking trajectory model, for the two different marking modes of zigzag parallel and contour parallel, with or without contour nesting, different laser marking path generation algorithms are designed. And compare the time-consuming of the two algorithms, discuss and propose methods to optimize the algorithm to improve the efficiency of the algorithm to meet the needs of industrial applications. According to the requirements of the distance between the retracted boundary and the hatch line, the zigzag parallel trajectory generation algorithm and the contour parallel trajectory generation algorithm are designed, and the hatching curve is generated according to the given contour point set. Based on the algorithm above, in the face of contour nesting, the algorithm is improved, and the hatching curve is generated according to the given contour point set. For the zigzag parallel trajectory generation algorithm, considering the intersection of the scan line and the inner contour. For contour parallel trajectories, this paper uses a vertex-based offset algorithm, which takes a long time. This is because the vertex-based offset algorithm is more complicated to solve the problem of self-intersection. This paper proposes that for the zigzag parallel method, the angle of the workpiece can be adjusted to find the optimal angle, so that the marking path is the smallest. And for the contour parallel trajectory generation algorithm, this paper proposes an optimization too.

When the size of the neural network is too large, calculating the bound of neural network is a difficult problem. Therefore, "size-independent" is what needs to look for here. This paper follows the path of "Size-Independent Sample Complexity of Neural Network ", and tries to get a better expression of Rademacher Complexity of neural networks.

Plenty of digital text is generated and shared in nowadays life, which constitutes many unstructured text resources. An automatically information extraction method is highly in demand as a large amount of useful information can be stored and output through it. In this paper, several basic concepts about entity relation extraction are introduced, including word embedding, positional embedding and convolutional neutral networks. Several types of relation extraction are also discussed: supervised relation extraction, relation extraction using distant supervision and relation extraction using few-shot approach. In addition, the existing challenges and problems are discussed, like overlapped triples, wrong labeled data and the difficulties that few-shot learning approaches are now facing.

In the Information Age, the success of companies and individuals increasingly depends on how quickly and efficiently they can turn massive amounts of data into actionable information. This paper first introduces the definition of recommendation algorithm and the process of collaborative filtering recommendation algorithm, then introduces the classification of algorithms in the Apache Mahout Framework, finally, the principles of user-based personalized recommendation algorithm and project-based personalized recommendation algorithm in Mahout framework are described and implemented respectively.

This article makes full use of geographic information and other industry thematic data, builds a township-level database and informational integrated management service platform, develops mobile APP and multiple characteristic applications, and deeply promotes the guarantee service of geographic information at the township level (sub-district office) and below, helping rural revitalization, building beautiful villages, deepening the promotion and application, and forming a township information service system that integrates government decision-making, department management, and public application, in order to promote grassroots information management, improve grassroots human settlements, scientific planning and construction, strengthen ecological and environmental protection services to provide useful references.

Objective: To explore the active ingredients and mechanism of Irpex lacteus against HUA (hyperuricemia) by using network pharmacology and molecular docking technology. Methods: Through literature search and collection, drug-like screening, target prediction, construct a target library of chemical components related to Irpex lacteus; search and construct HUA disease target library through disease database. The PPI network of the crossover gene between the components of Irpex lacteus and HUA was constructed using STRING and Cytoscape; the GO function enrichment analysis of the crossover genes and the enrichment analysis of the KEGG signaling pathway were realized through the Metascape database; the purpose of the "target-signal pathway" network was to analyze its specific mechanism of action; molecular docking to verify the binding ability of the main active ingredients and key targets. Results: 56 active ingredients were screened, the core target of PPI network mitogen-activated protein kinase 3(MAPK3), mitogen-activated protein kinase '(Mitogen-activated protein kinase 1, MAPK1), etc.; involved 1387 GO biological processes, 48 cell components, and 113 molecular functions; signaling pathway involved AGE-RAGE signaling pathway in diabetes complications, PI3K-Akt signaling pathway, IL-17 signaling pathway, etc; molecular docking results show that the core component tremutins A/C/E had good docking activity with each target. Conclusion: The related products of Irpex lacteus can treat HUA through multiple components, multiple targets, and multiple pathways, which could provide reference for further research and development.

In today's rapid economic development, the competitiveness of enterprises is more and more reflected in the talent competition, and the important means to obtain talents is recruitment. With the development of economy, recruitment channels are becoming more and more diversified. Enterprises will choose recruitment channels according to various reasons to recruit suitable candidates for the vacant positions of the enterprise. However, after finding the suitable people, how to keep them or hire them. After the recruitment of employees, whether the performance of employees is related to the recruitment channels has an important reference role for the selection of human resources. We use AI technology to measure the performance of employees through the three dimensions of subjective initiative, job satisfaction and job self dedication. Therefore, it has theoretical and practical value for improving the recruitment effect and contribution.

Aiming at the unsatisfactory performance of the particle swarm optimization algorithm in the optimization process of complex problems and being easy to fall into premature convergence, a power law function to improve the weight value of the particle swarm algorithm is proposed in the paper. It is compared with the linearly decreasing weight particle swarm algorithm, and the test is performed on four typical single-peak and multi-peak functions. The test results show that the particle swarm optimization algorithm in this paper has a better optimization effect.

Motion detection is vital for consumer electronics and the Internet of things (IOT). For a scenario where the motion is slow and gentle, the resolution of the motion sensor is critical for the detection, while the algorithm development is another critical issue to differentiate the motion signal from noise measurement. This paper investigates the feasibility of using higher order statistics kurtosis as a motion indicator. Statistical hypothesis test has been developed to assess the motion presence. Several experiments are conducted to test the feasibility and performance of the approach. The results show the approach is feasible, but with some limitations.

The modular identification of wireless communication signals is widely used in civil and military fields. With the rapid development of modern communication technology in China, signal modulation becomes complex and diverse. How to accurately identify communication signal modulation methods has become a research hotspot. Among them, multi band quadrature amplitude modulation (MQAM) is widely used in microwave communication, satellite communication and network communication. It has the advantages of high frequency band utilization and flexible modulation. Therefore, the recognition of MQAM modulation mode is of great significance. Traditional clustering algorithm has some limitations when it is used to identify MQAM signal modulation, such as too many iterations and long recognition time. For high-order modulated signals, a modulation recognition method based on Constellation Clustering is proposed. The constellation is used as the feature of modulation recognition, and the constellation of received signal is reconstructed in groups. The simulation results show that the collected characteristic parameters have strong interference ability to multi-channel, and the recognition rate reaches 97.6%. It has important engineering application value for signal modulation recognition in low SNR environment.

High quality Data and its derivative machine learning (ML) models are gradually becoming commercial commodity and provided to work effectively in an increasing number of areas. These ML model providers possess a set of trained models along with immense amount of source data stored on their servers. In order to obtain specific models, providers will require consumers to upload their domain-specialized data and conduct domain adaptation at the server side. However, considering the protection of the private information reflected by consumers' training data, as well as maintaining the commercial competitiveness of ML service, it is best that there is no data exchange between servers and consumers. Besides, consumers' data is always lack of supervision, i.e., classification labels, thus we are searching for how to conduct unsupervised domain adaptation (UDA) with no data exchange among domains in this work. We are the first to propose a novel memory cache based adversarial training (AT) strategy for UDA at the target side without the source data (the existence of source data is an essential requirement for regular AT). And our method includes a multiple pseudo labelling operation which is more accurate and robust than single pseudo labelling. The AT and multiple labelling work collaboratively to extract shared features among domains and adapt the learning model more specific to target domain. We carry out extensive evaluation experiments with a number of data sets and baselines, and according to the results, our proposed method can perform very well and exceed the state-of-art performance on all tasks. In the end, we also discuss how to extend our method to partial and open-set domain adaptation.

Lung cancer is the main malignant tumour affecting the health of residents in China. Automatically discriminating benign and malignant pulmonary nodules can facilitate the early detection of lung cancer, which reduces lung cancer mortality. The rising quantity of public available lung CT datasets made it possible to use deep learning approaches for lung nodules malignancy classification. Unlike most of the previous models that focused on 2D convolutional neural nets (CNN), here we explore the use of the DenseNet architecture with 3D filters and pooling kernels. The performance of the proposed nodule classification was evaluated on publicly available LUNA16 dataset, a subset of lung image database consortium and image database resource initiative dataset (LIDC/IDRI). It achieved a 92.4% classification accuracy. The proposed method provides an independent module with encouraging prediction accuracy that can be easily incorporated with a lung cancer computer-aided diagnosis system.

Feature extraction and classifier design are the main problems of acoustic signal recognition algorithms. In this paper, we extract the mel-frequency cepstral coefficients of the acoustic features of vehicles in complex scenes. Collaborative representation is introduced for the design of a classification scheme (Collaborative Representative Classification, CRC), which synthetically considers the relationship among samples. Experiments show that the proposed algorithm produces good performance in vehicle recognition for the case of a complex data set. Compared with other classification algorithms, the method improves the precision of recognition.

Generating a decent colored anime drawing from black-and-white sketches is not a simple task in the anime industry. For this problem, a novel deep learning model is introduced based on Conditional Generative Adversarial Networks(C-GAN), which can automatically colorize the anime sketches and output satisfactory pictures. By testing the structure of this model, a typical structure is found to have the best performance.

Speaker recognition is a technology that uses identity information in the human voice for identity recognition, which owns many advantages in convenient information gathering, low gathering cost and high recognition accuracy. However, the difficulty in gathering messages within short utterance declines the voiceprint recognition function rapidly. We propose a recognition model based on SincNet in the aim of obtaining enough feature information in short utterance. The model used a set of learnable Sinc-based filter banks to extract feature directly from primordial voice in featured extraction layer, which enabled neural networks to discover more valuable voiceprint information; In the pooling layer, we designed the pooling method of dual attention mechanism, which combined multiple self-attention mechanism and self-attention mechanism to enrich the feature information and enhance the differentiation degree of key features so as to solve the defect of short speech with less information; choose ArcFace as the loss function, which can maximize the classification limit in the Angle space, thus improving the classification ability of the model. Experimental results demonstrate that the proposed model performs better than the benchmark model.

The high efficiency of quantum algorithms is caused by the quantum parallelism of the superposition principle and the quantum entangled, but traditional quantum genetic algorithms only use the quantum superposition principle. In order to further improve the performance of the algorithm, this paper proposes a new higher-order quantum genetic algorithm, which adds quantum entanglement properties on the basis of the principle of quantum superposition. Finally, the traditional quantum genetic algorithm and the higher-order quantum genetic algorithm are used to test a set of nonlinear equations many times. The results show that compared with the traditional quantum genetic algorithm, the evolutionary update operation of the higher-order quantum genetic algorithm does not involve multiple judgment conditions of the traditional quantum gate, requires less evolutionary algebra, and can converge quickly.

Aiming at the shortage of research on micro-blog short text fine-grained sentiment classification, a fine-grained sentiment classification method about micro-blog short text based on PLSA model and K-means clustering model was proposed. PLSA is used to calculate the probability matrix between documents and topics, words and topics in the corpus. In terms of the probability distribution of words and topics, K-means algorithm is used to cluster the probability distribution of words on topics and merge the similar topics. Based on the sentiment ontology library, emotion recognition is carried out for the merged topics. Then, according to the merged document and topic probability matrix, the document sentiment category is classified. The experimental results show that the sentiment analysis method integrated with PLSA and K-means can obtain higher classification accuracy than the PLSA model method alone.

Computed radiology system (CR) test technology has been applied in the testing of industrial welding defect after the promulgation and implementation of digital radiographic imaging test standards. The performance of CR system directly affects the testing rate of welding defects. Therefore, testing and evaluating the performance of CR system becomes the premise and guarantee for the application of CR system. In this paper, based on the testing image of CR phantom and the simulation of scanner dither results, an algorithm combining fault-tolerant voting and improved point-based hoof transform is proposed to test the performance index of scanner dither in CR system. Experimental results show that the proposed algorithm can test not only the row and column dithering, but also the subtle arc dithering. Compared with standard Hough transform algorithm, this algorithm has higher accuracy and can avoid the influence of other image indicators around the T-shaped image quality indicator used in CR phantom.

With the rapid development of Internet technology, multimedia network teaching has become a new teaching mode in the teaching work of major universities. In the context of multimedia technology, the use of mind maps to carry out teaching activities in various disciplines can give full play to the advantages of both, which can use mind maps to present the knowledge learned in an intuitive and vivid way. This not only helps students sort out their knowledge, but also enables students to master important and difficult knowledge faster and more efficiently, and enhance their memory and thinking skills. This is also conducive to the cultivation of students' divergent systems, systematic thinking and creative logical thinking. Based on this, this article will discuss and study the application of mind mapping in college multimedia network teaching environment. The author hopes that the combination of multimedia technology and mind mapping can maximize the effectiveness of teaching.

Aiming at the ant colony algorithm is weak to solve the problem of QoS routing, this paper proposes a routing algorithm that combines quantum computing with the max-min ant colony algorithm. The algorithm introduces a quantum selection strategy, uses qubit probability amplitude to encode the path, and dynamically adjusts the quantum rotation gate to update the path quantum coding on the basis of the minimum cost to meet the constraints and other optimal conditions, combined with the maximum and minimum pheromone determination mechanism to find the optimal path, effectively control the evolution speed and direction, and accelerate the algorithm convergence. Experimental results show that the algorithm is effective in controlling the speed and direction of evolution, and has better optimization ability and convergence performance than ant colony algorithm.

In the process of rapid development of computer technology, the application of computer technology in many fields has become more and more common. The application of computer technology in the process of corporate financial and accounting office is beneficial to improve the efficiency of financial and accounting office and ensure the quality of accounting office. Enterprises not only need to strengthen the research work on the application of computer technology, but also put forward effective measures for the application of computer technology in the accounting and office process. Only by improving the application level of computer technology can it truly play its positive role in the accounting and office system.

Recently, the Generative Adversarial Network (GAN) has attracted much attention in a large range of applications, such as image classification, plausible image generation and image completion. Lots of GAN variants have been proposed. This paper focused on two popular GAN variants, including GAN and Auxiliary Classifier Generative Adversarial Network (ACGAN) and made a comparison between them. The experiment on CIFAR-10 and anima dataset shows that ACGAN can perform better than original one with better accuracy. Furthermore, we find ACGAN has a worse inception score (IS), which indicates that ACGAN is not capable of generating realistic pictures.

According to the multi-objective and multi-decision problem of traditional financial Investment, a Financial Fuzzy Investment model based on Combinatorial Entropy Value Method is proposed. The company financial index system and index calculation model are constructed, an improved entropy method is proposed to calculate the index weight value, which can comprehensively consider the influence of historical evaluation data and current evaluation data, and the fuzzy membership function for investment satisfaction is proposed. Finally, an example is used to verify the effectiveness of the Investment model.

In an era of information, people's demand for electronic products is greatly increased. As an important part of electronic components, Printed Circuit Board (PCB) has a huge annual output and a variety of sizes and types. Therefore, the traditional method of manually detecting PCB defects may fail to meet the required production standards due to the high error rate. With the development of deep learning, a batch of PCB defect detection models combined with deep learning have been produced, which improves the detection efficiency. However, there are still some problems in these methods, such as low automation degree, low detection degree, and poor stability. This paper proposes an improved algorithm, based on YOLOV4, which uses PCB defect data set released by the Intelligent Robot Laboratory of Peking University, and has abundant images of different defect types, which greatly increases the reliability of the model. By analyzing the feature distribution of CSPDarkNet53 structure layer and the detection target defect size distribution in the data set used, in the data pre-processing and input stage, the image is automatically subdivided according to the average size of the callout box of the detection image, and the probability of anchor containing detection target is increased. Experimental results show that the improved YOLOV4 algorithm has a mean Average Precision (mAP) of 96.88%.

In recent years, the continuous development of the basic construction industry has promoted the progress of urban social economy, and the safety issues in construction have attracted more and more attention. The integrity of the external protection net is of great significance to the safety of the construction personnel. Most of the traditional methods of detecting damage to the protection net have problems such as poor real-time performance and low detection accuracy. In this paper, we present a novel approach based on SSD for solving the problem of the integrity detection of the external frame protection net in the construction site. The improved network merges adjacent feature layers to enrich the semantics of shallow feature maps, thereby improving the detection accuracy of the network, and simultaneously introduces a shallower feature layer and a more advanced receptive field module to improve the detection ability of small targets. The experimental results show that compared with the SSD algorithm, the improved algorithm has higher detection accuracy in the detection task of the external protective net on the construction site, and can meet the real-time detection requirements.

The rapid development of Internet and the extensive user base have led to a significant increase in the propagation speed of information and public opinion. Quick and accurate sentiment analysis of public opinion can help analyze and manage public opinion. Since BiGRU can simplify network parameters, and the Multi-Head Self-Attention(MHSA) mechanism can learn long-distance features more accurately and quickly. In this study, we propose a BiGRU-MHSA network to perform Chinese sentiment classification. The innovation of the model lies in the combination of BiGRU and MHSA in order to focus on several semantic centers and comprehend the whole text more precisely and efficiently. At the same time, we use the Chinese comment datasets on Weibo to verify the effectiveness and accuracy of the model. From the experimental results, we find that our proposed model improves the accuracy by about 1% on different comments datasets without consuming too much computing resources.

With the development of the big data analysis technology, more and more companies are willing to apply this technology. In order to further promote the big data investigation teaching of Sichuan Police College, we accurately grasp the new situation, new tasks and new requirements faced by the big data investigation, and further deepen the solidification of the implementation path. Through on-the-spot visit, discussion and communication, we will focus on understanding the application technology of big data, extend it to various work of public security organs' attack, prevention and control. Then we want to promote the establishment of practical application systems, and promote the fundamental change of criminal investigation mode through the construction of big data.

How to use artificial intelligence to assist physicians in analyzing complex breast cancer medical data has become an urgent problem to be solved. From the perspective of auxiliary diagnosis, we propose a method to classify breast cancer grades by cases. Based on the Chinese electronic medical records of breast cancer provided by the Department of Ultrasound at West China Hospital, we compared neural network methods, including FASTTEXT, TEXTCNN, TEXTRNN. And proposed the N-LM-ATT model. N represents the n-gram feature, LM represents the Long Short-Term Memory (LSTM) network, and ATT represents the Attention Mechanism (AM). This article focuses on the classification of Chinese breast cancer text data based on natural language processing (NLP) method. By comparing the above methods, the N-LM-ATT model achieved the best performance on the breast cancer ultrasound dataset of West China Hospital.

Recently, a new SARS-CoV-2 coronavirus named COVID 19 sparked panic worldwide. More than 1000 ongoing therapeutics were dealing with it. However, the efficacy of these treatments was still fuzzy. In this paper, we analyzed 4 main types of treatments individually and mutually based on clinical trials data. T-test, models like COX and GEE logistic regression are adopted to evaluate the efficacy.

Aspect-based sentiment analysis is to analyse the sentiment polarity of the aspect in a specific context. The previous methods mostly use the working mode of pipeline model, and internally use Recurrent Neural Network to predict the sentiment polarity of the aspect. This way clearly defines the order of aspect extraction and aspect-level sentiment classification, which will lead to the problem of error transmission. And traditional models have great limitations. In this paper, we propose a multi-task learning network based on attention (MNA), which uses shared word embedding to transmit to downstream tasks, allowing extraction and classification tasks to be processed simultaneously. We design a network architecture based on attention mechanism, and use the improved multi-head attention mechanism to transfer information from extracted tasks to classified tasks. Experimental results on three benchmark datasets show the effectiveness of MNA model.

Teaching evaluation is an important means to evaluate the quality of university education. The traditional teaching evaluation process requires a lot of manpower and material resources to uniformly allocate resources with inconsistent data storage methods. Therefore, it is of great significance to develop a teaching evaluation system for data management and display. The system adopts the PHP development language, the mature oneThink development framework, the data management method is used to process the data, the data is stored in the MySQL database, and the MVC software design architecture and B/S mode are used for the front-end and server interaction. This article describes the key steps of the development system, and proved in practice that the system is stable and compatible with different operating systems and browsers. The results show that this system simplifies the process of teaching evaluation, improves the efficiency of university educational affairs, and has positive significance for the improvement of university teaching quality.

In the traditional education teaching model, the case teaching has been widely adopted by teachers due to its many special advantages. However, case teaching also has certain limitations and scope of application throughout the education stage. Only by continuously optimizing the traditional case teaching model can its limitations be broken through. Throughout the education process, the concept of blended learning can be fully used for reference to construct a blended case teaching model based on the information technology environment.

With the growth of Android apps market share, there are more and more kinds of malicious apps for Android system. At the same time, the method for malware detecting is evolving. In order to deal with the growing threat of malware, many detecting methods are combined with machine learning algorithms. In this paper, we compared the ability to detect malware of five algorithms on a dataset about Android apps networking. The result revealed that the networking data could be used as a reference for the classification of Android apps. Moreover, among the five algorithms used in this paper, the ensemble learning algorithm using the decision tree as base leaner performed the best in the comparison. It reached more than 75% accuracy when predicting malware using the test set.

In order to solve the problems of poor security and low throughput of the current electronic medical records, this paper proposes an electronic medical record storage model based on H-Algorand consensus mechanism of credibility assessment combined with blockchain technology. To ensure the normal operation of blockchain system effectively, the model used a credibility assessment algorithm to obtain the trust value and replace the account balance in the H-Algorand consensus mechanism with the trust value as the node weight. The results show that the storage model adopted here has higher security and throughput efficiency.

In deep learning-based object detection, especially in face detection, small target and small face has always been a practical and common difficult problem due to its low resolution, blurred image, less information and more noise. In some applications, sensing image data is hard to collect, leading to limited object detection performance. In this paper, we investigate using a generative adversarial network model to augment data for object detection in images. We use generative adversarial network to generate the diverse objects based on the current image data. An improved generative adversarial network is added in the network and a new loss funtion is applied during the trianing process to generate diverse and high-quality traing images. Experiments show that images generated by generative adversarial network have higher quality than counterparts.

Deep learning plays the leading role in the field of image recognition, natural language processing, and speech processing. But many researches reveal that the existence of adversarial examples would make models vulnerable. Most of the researches focus on restricting norm distance which result in perturbing images too much. With examples given in this paper, we demonstrate that restricting the number of modified pixels is easier to classify in human vision instead of norm distance, meanwhile, it can get a high accuracy of non-targeted attack on VGG16, ResNet50 and DenseNet201. Our method is adding Bézier curves on original image using differential evolution algorithm (DE). Through continuous evolution, kept the good individuals and eliminated the bad ones, DE can guide the search to the best solution. With DE, we do not need to know the gradient information about models, therefore, our method is robust on different models. When we add more Bézier curves on original image, we can get a higher accuracy which also shows that our method is effective. We find that the distribution of prediction labels focuses on a few categories, which means some labels are easier to attack than the most others. We also think the encode order about labels has an influence on the distribution of misclassified labels.

In this paper, an orthogonal chirp-division multiplexing (OCDM) symbol-based radar signal processing approach used to estimate the distance and velocity of the target echo in the Radar-Communication (RadCom) system is introduced. And to improve the accuracy of radar resolution, two radar signal processing algorithms, respectively based on periodogram and multiple signal classification (MUSIC) are presented. Simulation results show that these algorithms can significantly improve the scatterer of the radar range profile.

Autonomous driving technology has developed rapidly in recent years, and computer vision has shown significant role in automated vehicles. Researchers use the combination of computer vision and deep learning to improve the speed and accuracy of road element detection technology. This article focuses on the semantic segmentation and road detection elements, especially in lane line segmentation and signal light detection. It summarizes the visual inspection technology and a series of visual perception algorithms based on deep learning. Although semantic segmentation and target detection technologies in general scenarios are already very mature, they are not ideal when applied to autonomous driving environments, especially where segmentation and detection are performed simultaneously. Therefore, this paper builds a fusion network of lane line segmentation and signal light detection suitable for autonomous driving scenarios.

From the perspective of marine environmental forecasting and in combination with the marine environmental security service for "Southern maritime silk road", this paper introduces the basic framework of the data-acquisition system for marine-environment forecasting of the "Southern Maritime Silk Road", which is constructed by research, business and service fields of marine environmental forecasting technology in China: including data source engineering remotely sensed by marine satellite, data source engineering by oceanographic survey (inspection), data source engineering through observation at fixed marine stations, and its support system - transmission network of the observed global Marine environmental data and the supercomputing center. The reanalysis system of data remotely sensed by marine satellite, data system by oceanographic survey (inspection) and data system through observation at fixed marine stations are mainly thrown light upon. The data-acquisition system framework for the marine environment forecasting of the "South Maritime Silk Road" lays a solid foundation for the development and sustainable production of big data industry products of the downstream Marine environment forecasting and security guarantee service.

The design of interactive art installation, which integrates interactivity, artistry and education, leads people to think their relationship with nature and society by human-installation interactive movement. Due to the repaid development of science and technology, the simple interaction with settled program cannot satisfy people's demands. Instead, they want to step into the world of installation through personalized emotional catharsis. The current mode of communication will be changed by the in-application program within the interactive art installation. The single installation totally produced by artists' thought will be replaced by the finished installation product with the participation of the audience. In this way, participants will immerse in the story plot and feel various emotional changes psychologically by sound, image and touch, which relies on the interactive art installation. The Listen is taken as the example to spread the idea of sharing the ecological nature of human beings and the world. Arduino, an open source platform, will be studied to show the change of modern interactive communication mode, exploring the impact of programming tool on the development of future interactive installation and providing some new ideas for the integration of artistic creation and science and technology in the future.

With the advent of high-speed railway era, users' demands for high link capacity and high reliability are increasing. In this paper, a resource allocation algorithm based on deep reinforcement learning (DRL) is proposed in the high-speed railway (HSR) scenario. The communication link between high-speed train and high-speed train (T2T) is taken as an agent, and the link capacity and packet transmission success probability (PTSP) are taken as optimization objectives, which are mapped into the reward function. The agents gain experience by constantly interacting with the HSR communication environment to update the neural networks in DRL and eventually learned the optimization policy. The simulation shows that when the trains' speed is 350km/h and the transmission packet size is 2385Bits, the high reliability of T2T link can still be maintained, and the capacity of high-speed train to Infrastructure (T2I) link is improved by 1Mbps compared with the baseline, which proves that the proposed algorithm can mitigate the interference introduced by spectrum reuse and doppler shift.

To solve the problem that existing modern tram track cleaning vehicles cannot automatically and accurately identify grooved rail garbage, this paper proposes a grooved rail garbage identification algorithm based on improved YOLOv3. This algorithm firstly extracts the groove rail region in the image, then adjusts and optimizes the residual block and convolutional layer network in the basic network of YOLOv3, and adopts k-means algorithm for clustering analysis to obtain anchor value of the adaptive dataset, which effectively improved the average identification accuracy of grooved rail garbage. The experimental results showed that the improved YOLOv3 model mAP reached 89.16% on the self-made Grooved Rail Garbage Dataset. For the image of 416×416, the recognition speed reached more than 14 frames per second. Compared with the traditional grooved rail garbage identification algorithm, this algorithm not only has higher recognition rate and accuracy, but also has good robustness to environmental changes.

As a branch of artificial intelligence, machine learning has received a lot of attention and has become a popular research field. It connects the huge database obtained from daily life with data prediction, which is of great help in decision making. In this project, we apply linear regression models and use different gradient descent methods, including general, mini-batch, and stochastic gradient descent, to optimize the error. We used convergence rate and accuracy as evaluation criteria, which will be explained via run time of convergence and mean square error respectively. For the regression models, both single-feature and multi-feature hypothesis models are investigated. In consideration of the risks of multicollinearity and overfitting, only simple linear model and its exponential transformation are adopted. The result indicates that mini-batch gradient descent has the biggest comprehensive advantage.

Informatization has become an inevitable trend in office, and it is also the main way for modern business society to improve work efficiency. From the observation of the reform of file management in practice, its main direction is based on computer technology. This research selects three aspects based on computer technology, including information resource management, technology management and service management. It mainly discusses the integration and utilization of information resources, technology management promotion measures and innovative service management system paths.

For many enterprises, knowledge workers have become an important human resource. As the human capital that can add value, the management of knowledge employees has become an important issue of concern for enterprises. At present, there are still some problems in the incentive research of knowledge workers in China, such as the incomplete research object, the relatively single industry, the lack of pertinence and operability of the research content, the lack of tools and methods to measure the incentive effect of knowledge workers. In the management practice, the incentive of knowledge employees is mainly based on salary, with little attention paid to other incentive measures, the enterprise's understanding of it is not deep enough, the incentive mechanism is not sound, the incentive effect is not significant and other problems need to be solved.

With the rapid development of new media technology, the reading promotion model has also changed. The libraries in university no longer use the traditional promotion model, they pay more attention to the promotion content of reading. The needs of users should be the center in current reading models. To provide more precise services to achieve the results. On this basis, this paper mainly discusses that the reading promotion mode in the new media environment is carried out from the perspective of evaluation.

Planning based text generation from structural data is an emerging approach that manages to plan for "what to say." Existing methods can not dynamically adjust the planning, because they generate complete plans statically without considering the result of surface realization. To address this issue, we propose a model that conducts dynamic text planning and surface realization alternately. In each sentence's generation, our model first plans the records according to the generated text and the plan history, then realizes the sentence conditioned on the corresponding plan. Experimental results demonstrate that our model performs better on both text planning and text generation. Our model with the dynamic planner achieves impressive results on the E2E dataset.

The RTCA/DO-330 "Software Tool Qualification Consideration" is released after the RTCA/DO-178B was widely used for tool qualification in the latest 20 years. In the RTCA/DO-178C collection, the RTCA/DO-330 is used together with RTCA/DO-178C to define the tool qualification requirements for airborne software development. This paper compares and analyses the different requirements in RTCA/DO-178B and RTCA/DO-178C+DO-330 to provide an indication of what are the new requirements and what has not been changed. For those unfamiliar with the tool qualification requirements in RTCA/DO-178C and RTCA/DO-330, this paper serves to provide an entry point to this new certification guidance.

In order to solve the problem of human resource selection of demander for provider in cloud manufacturing environment, this paper proposes a selection decision method from the perspective of provider. Firstly, the evaluation index system of demander is established, and the objective weight of each index is obtained by entropy weight method according to the demander's demand information. Secondly, the subjective weight of each index is determined by improved AHP method and the comprehensive weight of each index is obtained by combining the subjective and objective weight. Then, the traditional TOPSIS method is improved to evaluate the demander and provide reference for the selection of supplier. Finally, an example is given to verify the effectiveness and practicability of the method.

During indoor mobile mapping, localization plays an important role. It is difficult for IMU, odometer, Lidar or SLAM algorithm singly to meet the high efficiency, real-time, accurate and robust performance. So the fusion of the data measured by these different methods is a current research hotspot. However, most of the researches still focus on the loosely coupled fusion based on filtering methods, and the data cannot be fully utilized. In this paper, based on LOM-SAM framework, the Intensity Scan Context is introduced to extract keyframes and detect loop closure which will provide a closed-loop factor, together with IMU pre-integration factor and Lidar odometry factor to construct the factor graph. Then incremental smoothing optimization algorithm is used to get high precision trajectory, realize the tightly-coupled Lidar and IMU positioning. The results show that the number of keyframes is reduced, the elevation error is effectively decreased, and the real-time performance is improved without decreasing the accuracy of LIO-SAM.

With the rapid development of aviation science and technology, the structure and performance of aircraft are more and more complex, and the difficulty of maintenance is correspondingly increased. As a result, quality problems are more likely to occur during the maintenance process. In order to improve the quality of aircraft major overhaul, it is necessary to identify and analyse the risks that affect the quality of aircraft major overhaul. This paper comprehensively uses the improved HHM-RFRM method. At first, the risk factors that may exist in the process of aircraft major overhaul are identified by HHM. And then uses the improved FMEA to carry out triple standard filtration. The multiple standard evaluation is used for further filtering. Finally, 10 important factors are obtained. The purpose of this paper is to put forward a general method to identify the quality-risk of aircraft maintenance projects, which has a certain guiding significance for different levels of aircraft maintenance projects.

In the era of continuous development of e-commerce, the selection of e-commerce affects the development of e-commerce business and management efficiency. Based on the multi-party secure computing technology, this paper designs a set of the electronic voting scheme based on SDGHV multi-party secure data platform for the feedback on customers' opinions after the goods are sold, which provides a useful reference for the selection of e-commerce products.

The sudden and frequent natural disasters have brought a series of damages to both the economy and society of all countries in the world and made people to suffer a double loss of life and property. In this context, it raises a claim for emergency rescue activities in terms of more scientific, more accurate and quicker responses. Emergency rescue is taken as one of the most important ways to save lives and reduce losses when natural disasters attack; while proper emergency supply storage facility location is the key to guarantee the effectiveness and efficiency of emergency rescue operations. Based on the characteristics of emergency rescue activities, this paper analyzes the scientific problems in the field of emergency rescue and summarizes the current progress of emergency supply facility location modeling and algorithms with the purpose of laying a theoretical foundation for the future research.

In order to meet the demand of intelligent city for domestic garbage classification and effective transportation, this paper discusses the overall scheme of the domestic garbage classification management system based on the "four classifications" standard of domestic garbage, using the characteristics of the Internet of Things and the existing intelligent garbage classification equipment. Subsequently, an intelligent domestic waste removal and transportation plan was proposed, and a new model of domestic garbage intelligent removal was established. Through the development of mobile terminals and the design of the background management system, an intelligent domestic garbage removal management platform is built, which improves the efficiency of transportation and realizes the intelligent management of domestic garbage classification and transportation.

As a quantitative tool and implementation means, performance appraisal and motivational mechanisms play an important role in the enterprise safety management system. Taking the negative motivation strength as the research object, this paper established an evaluation method and calculation model of enterprise safety performance motivation. A comparative analysis of the safety performance motivation of four electric power enterprises and three industries was conducted to examine the evaluation method and calculation model. The results illustrated the characteristics of safety performance appraisal strength of those enterprises and industries. This study can further enhance the awareness of safety responsibility of all employees in enterprises, provide a reference for other industries to implement safety performance reward and punishment mechanism and a useful reference for enterprises to improve safety management.

This research aims to solve technical problem of automatic feature recognition in a Computer Aided Process Planning system. This feature recognition method was developed to represent a 3D model as classes of manufacturing features based on graph theory. In addition, "accessibility" was proposed to recognize intersecting features with complicated graph structure which cannot be predefined by users. The experimental result shown in case study proofed that this proposed method is effective in multi-intersecting feature recognition.

The establishment of safety management system is of great significance to improve safety management of state-owned enterprises. In order to provide some help for state-owned enterprises to establish a safety management system, the shortcomings of state-owned enterprises in safety management are summarized and a general behavior paradigm of establishment of a safety management system from the enterprise level is put forward based on literature research. Besides, the key elements of safety management system are discussed by fuzzy Borda number analytical method based on the grille, and the results show that when setting up the key elements of the safety management system, enterprises should pay more attention to objectives and policies, risk control, emergency management, institutions and responsibilities, laws, regulations and standards, occupational health, performance evaluation, interested parties, check and correct, management review, education and training, on-site management, accident management, leadership and commitment, system support and continuous improvement.

Model-based system engineering theory is a new breakthrough in aircraft development methodology, which can solve the problems of aviation system development, high integration, difficult management and high risks. This paper introduces the basic theory of MBSE, and applies MBSE method in the design of communication system to realize audio communication function. Starting from the needs of stakeholders, perform analysis and define system use cases. According to different scenarios, define the function requirements activity flow and determine the functional model. Then map the functional architecture to the physical layer, carry out scheme trade-off analysis, architecture design and generate physical architecture diagram to ensure that the final solution meets the needs of stakeholders. Through model evolution and iteration, product design is realized. Modeling can provide all parties with a universal and unambiguous design information exchange platform, and at the same time facilitate the realization of system virtual integration and improve the existing design process.

In the context of the information age, big data technology and block chain technology are currently rapidly developing technology types. The application of these two technologies in the accounting industry can greatly promote the development of the accounting industry. Big data technology is conducive to increasing the digitization of accounting information. It can promote the digitization of original financial information and non-financial information, thereby expanding the scope of management of financial staff. In the information age, it is essential to ensure the reliability of accounting information. False accounting information will affect the business management strategy of the enterprise to a certain extent, thereby affecting the long-term development of the enterprise. The effective application of block chain technology can improve the efficiency of accounting data analysis. Therefore, in the current development of the accounting industry, companies need to study the specific application of big data technology and block chain technology. Companies also need to analyze specific strategies in the process of integrating financial accounting and management accounting under the background of big data and block chain technology. Only in this way can the reform and innovation of the accounting industry be promoted.

Attitude capture clothing is widely used in virtual anchor, virtual fitting and other fields, and MEMS-IMU is the key component to realize high cost-performance attitude capture clothing. However, a precision turntable is needed in order to estimate gyroscope's static biasing, and measurement errors of MEMS-IMU are difficult to be obtained. Based on the problem, both attitude and gyroscope's static biasing are estimated in field through a Kalman filter, a design method of measurement errors is proposed for improving attitude's accuracy. In the proposed method, a mathematical model of measurement errors is given, and the optimal values of measurement errors are obtained by analysing the relationship between attitude's accuracy and measurement errors. A MEMS-IMU with name of MPU9250 and size of 3 mm*3 mm* 1 mm is used for verifying the feasibility of the proposed method. Experimental results show pitch, roll and yaw deviate from the reference values by 0.008°, 0.006° and 0.6° respectively when the MEMS-IMU keeps movement at a constant speed in arbitrary trajectory, which is significant for promoting the application of MEMS-IMU in attitude capture clothing.

The differences in business models, lending products and risk control of lending products of Internet financial lending platforms have led to the disorderly growth of Internet Financial platforms in China for a long time. With the voice of financial innovation, the number of platforms and the scale of transactions have increased significantly. In this pager, we will enter the network diffusion model to analyze the physical and statistical characteristics of the competitive behavior and fierce competition behavior of the Internet financial lending platform, and understand the overall characteristics of the Internet P2P lending platform under the specific platform classification, and finally put forward the corresponding policy suggestions.

From the end of 2012 to 2019, 93.48 million population were lifted out of poverty in China. In 2020, which was the final year for poverty alleviation, absolute poverty is eliminated. China have realised a great achievement in poverty reduction. According to the World Bank report, 689 million people living on less than $1.90 a day. At higher poverty lines, 24.1 percent of the world lived on less than $3.20 a day and 43.6 percent on less than $5.50 a day in 2017. Hence, this research was designed to synthesise and represent a comprehensive understanding of the existing Chinese scholars' articles in poverty alleviation field during the 13^th Five-year plan by utilising bibliometric techniques. The result demonstrated the state-of-art of researches in this field. The fundamental articles were identified. The prosperous themes were revealed. Sustainable development and multidimensional poverty are the motor themes.

With the development of new energy technology, new energy stations are becoming more intelligent and data-based, and cyber-attacks on new energy stations are increasing year by year. In response to the continuous threats brought by malicious traffic to the network of new energy stations, this paper researches on the traffic anomaly detection technology based on network communication characteristics. An adaptive abnormal traffic detection method for new energy stations based on HHT algorithm is proposed, which improves the efficiency of identifying abnormal network traffic and more accurately identifies network attacks against new energy stations. It is verified through experiments that compared with mainstream classifiers, the method studied in this paper can achieve adaptive detection while adaptively determining the threshold, and the detection accuracy can reach 95%, the false alarm rate is lower than other methods, it can provide more accurate identification results for new energy field station network cyber-attacks detection.

With the rapid development of airborne system and software, the requirement of software's robustness, adaptability and reliability improves continually. Failure of software function may cause severe property loss and personnel casualties. Formal methods are mathematically based techniques for the specification, development, and verification of software aspects of digital systems that have obvious advantages in improving safety and reliability of airborne software. This paper takes Air Data Computer software as an example for formal verification in the whole software development process including requirements process, design process, coding process and integration process to demonstrate a set of formal methods application in the airborne software verification process.

The research on effects of supplier concentration on financial performance of manufacturing enterprises is conducted from the perspective of mediating effect of commercial credit based on financial data of 1090 listed manufacturing companies. MLRM (multiple linear regression model) plus empirical data and the mediating effect test method are used in the research. The results show that the commercial credit plays a partial mediating role in impact of supplier concentration on enterprises' financial performance, that the commercial credit as a financing means can improve the financial performance of enterprises to a certain extent, and that supplier concentration exerts a significant positive effect on financial performance. Thus, this research can provide theoretical reference in supply-chain management for enterprises.

As a new type of artificial intelligence technology, generative adversarial network (GAN) has good data understanding and generation capabilities, and has a wide range of application prospects in the fields of image and speech. However, due to the lack of prior knowledge, its training process is less robust and prone to occur the pattern ignore. Its development is restricted to a certain extent, and its application scope still needs to be expanded. To solve the above problems, this paper introduces a knowledge confidence multimode GAN (KC-MGAN) algorithm, calculates the confidence of the input data through the reasoning method, and then puts the confidence and the input data into the GAN system to generate new sample data. During the training process, the confidence of the input data is continuously calculated, while the generated data samples are continuously evaluated. The training process will end until the GAN system reaches a stable condition. Finally, this paper takes the generation of UAV flight trajectory data as an example to verify the effectiveness of the proposed method. Some explorations have been made for the application of data generation and GAN's training mode with the prior knowledge.

With economic development, ports assume more responsibilities. However, the level of coordination between ports and hinterlands in my country's inland river ports needs to be improved. By constructing a comprehensive evaluation system of port energy level-hinterland economic system, entropy method and coupling coordination Degree model for coordination analysis. The article takes Chongqing Port as an example to study the characteristics of coordinated development between port energy level and hinterland economy from 2010 to 2019. The results found that: 1) The coordination evaluation model effectively reflects the development status of Chongqing Port in the past 10 years; 2) With the increase of time, the port energy level and hinterland economic level of Chongqing Port are in a relatively stable state of improvement; 3) Port energy The degree of coordinated development between the economic system of the Hinterland and the hinterland shows a steady upward trend overall, but there is still room for improvement.

Combine harvester is an important part of agricultural modernization, and the function of the header is more important. The header is mainly composed of a transmission device, cutter, and reel. In the process of practical application, the harvest efficiency of different types of the combined cutting machine is much higher than that of manual harvesting. In this paper, a ground profiling sensing mechanism is designed by building a dynamic model. The results show that: the ground profiling sensor mechanism of combine header based on the symmetric algorithm can truly reflect the agricultural terrain data, and provide data support for the action of digging shovel and improve the profiling accuracy of the combine.

In recent years, with the rise of short video platforms and e-commerce platforms, live streaming with goods has become a hot online sales model. However, the sales volume of live broadcast by big anchors is tens of thousands, which is actually supported by a complete supply chain system. By studying the mode and current situation of the supply chain of live broadcast e-commerce, and analyzing the existing problems, suggestions on the integration and optimization of the supply chain of live broadcast e-commerce are put forward, so as to provide references for relevant personnel in this industry.