The proceedings contain 306 papers. The topics discussed include: CloneBAS: a code clone detection method based on abstract syntax tree and Simhash;research on the a-star algorithm based on path finding;study on vibra...
ISBN:
(纸本)9798350341546
The proceedings contain 306 papers. The topics discussed include: CloneBAS: a code clone detection method based on abstract syntax tree and Simhash;research on the a-star algorithm based on path finding;study on vibration isolation characteristics of UHVDC converter valve saturable reactor;a market transaction model based on blockchain technology for multimicrogrid networks;optimization research on radar signal processing technology based on MTI-MTD;stock prediction using evolutionary attention-based LSTM;the reasons and improvement measures for the long time of replacing cross arm with 10 kV live;research on accurate three-dimensional modeling technology and intelligent mining in coal mines;evaluation on power equipment operation status monitoring system based on datamining;exploring deep learning-based visual localization techniques;and intelligent detection and recognition of pulley paying-off defects based on improved YOLOv5.
This study is based on a series of infrared spectroscopy data of mixed solutions, which recorded the spectral fluctuations by changing the infrared wavelength under known experimental conditions, including different c...
详细信息
ISBN:
(纸本)9798400716959
This study is based on a series of infrared spectroscopy data of mixed solutions, which recorded the spectral fluctuations by changing the infrared wavelength under known experimental conditions, including different concentrations of solutionU(IV), U(VI), and nitric acid. The solution of each concentration was performed in triplicate during the experiment, for a total of 256 experiments to obtain a more comprehensive and reliable data *** techniques based on neural networks are explored after obtaining datasets. We construct an efficient and accurate model of mixed solution concentration detection by fully exploiting the advantages of neural networks in automatic feature extraction and patternlearning. Studies have further explored how to achieve more comprehensive detection effects by mixing multiple models, including neural networks and other state-of-the-art prediction models. By tuning the model hyperparameters, performance analysis, and fine adjustment of weight allocation, we expect to build a hybrid model with good generalization performance and robustness. This study provides an advanced and feasible solution for the field of mixed solution concentration detection, with a wide range of research and application values. Finally, the research results of this paper aim to provide strong support for accurately predicting the concentration of various solutions in mixed solutions. By combining spectroscopy and machinelearning methods, we seek to achieve higher prediction accuracy in practical applications, providing new analysis.
Bridge operation and maintenance is an important solution to how to ensure the safe operation of bridges and extend their service life. In this paper, a dynamic early warning technology of BIM (Building Information Mo...
详细信息
Opinion mining is the study of user opinions as revealed by their text messages. It comprises categorizing user attitudes into several polarities, such as positive, negative,or neutral. For the analysis, a whole other...
详细信息
To alleviate the influence of angle change and dressing style change on gait recognition, a gait recognition network GaitRL based on local gait relationship learning is proposed. First, the pedestrian gait image is ac...
详细信息
Plant leaf diseases have increased in prevalence recently, making it necessary to conduct accurate research. Bean leaf diseases can be prevented from spreading by earlier detection and accurate identification. Disease...
详细信息
Large amounts of historical structured documents are available in archives around the world. Here, we are interested in automatically extracting statistical information of selected attributes contained in these docume...
详细信息
ISBN:
(纸本)9783031705489;9783031705496
Large amounts of historical structured documents are available in archives around the world. Here, we are interested in automatically extracting statistical information of selected attributes contained in these documents. To this end, we propose a pipeline relying on probabilistic indexing and machinelearning to mine and analyze relevant information contained in historical form images. These ideas are assessed on a large collection of images containing almost 295 000 forms with results showing the adequateness of the proposed methods to perform the bigdata analytic task considered.
Face based e-crime identification initiative combines cutting-edge technologies to detect criminals in video footage, allowing for quick and effective online FIR (First Information Report) reporting. Furthermore, it u...
详细信息
Developing a high-level automated security system for identification/authentication has long been an active research subject in practically all fields. Traditional security systems utilize a key or a password to secur...
详细信息
The annotated data set for relational extraction task is usually generated through distant supervision which results in model performance degradation due to noise data introduced. Most of the existing works eliminate ...
详细信息
暂无评论